Writers of Pro Football Prospectus 2008

Most Recent FO Features

ManzielJoh14.jpg

» The Week In Quotes: August 15, 2014

A very classy edition of TWIQ brings you observations on obscene gestures, racial slurs, pro wrestling, intrasquad brawling, inappropriate places for urination, and preteen sexual activity.

04 Mar 2008

Four Downs: NFC East

by Daniel Adler and David Lewin

Dallas Cowboys

Coaches, Coaches, Coaches!

As a player, Jason Garrett had one memorable game. Starting in place of an injured Troy Aikman on Thanksgiving Day in 1994, he threw for two touchdowns and more than 300 yards to lead the Cowboys to a second-half comeback over Green Bay and show Jay Fiedler that an Ivy League quarterback could make it in the NFL. That was about the extent of Garrett's impact during his seven years with the Cowboys as a player. However, during that time, owner Jerry Jones supposedly took a shine to him and tabbed him as a future head coach.

When Bill Parcells "retired" after the 2006 season, Jones made Garrett, with just two seasons as a quarterbacks coach in Miami, his first hire of the new staff. Fearing Garrett was not quite ready to be a head coach, Jones hired wily veteran (read: retread) Wade Phillips to be head coach and run the defense. As we all know, Dallas went on to have a remarkable season and Garrett became the "it" coaching candidate of the off-season. Jones gave Garrett about three million reasons and the title of assistant head coach/offensive coordinator to get his protégé to stay. After a pay bump for Phillips to keep his salary slightly above that of his assistant, the Cowboys are faced with a bit of a coaching quandary. Other teams have dealt with high profile assistants before -- well, the Redskins have dealt with high profile assistants, at least: Marvin Lewis handled defense when Steve Spurrier tried to bring his fun-n-gun offense to the NFL in 2002; more recently, Gregg Williams controlled the defense while Joe Gibbs concentrated almost entirely on offense. What is unique and worrisome for the Cowboys is that unlike other instances of highly paid assistants, where the head coach is one of the highest paid in the league (i.e., most secure), Phillips is seen as a temporary solution. Phillips is in an unenviable position: He could win 12 or 13 games and be fired.

As if the Cowboys did not have enough coaching turmoil, Parcells decided to raid the lower rungs of their coaching staff, choosing Tony Sparano to head the Dolphins and promoting others holdovers from his Cowboys staff to higher positions with the Dolphins. Sparano should not be a giant loss; he called plays for the Cowboys in 2006, but was relegated to offensive line coach in 2007 upon Garrett's arrival. While the Cowboys staff loses a former University of New Haven head coach in Sparano, it gains a former Dallas Cowboys head coach in newly reacquired secondary coach, Dave Campo. Why did Phillips hire Campo? Theories abound:

  • He wanted someone who certainly would not usurp him.
  • He wanted someone to give him a fair match in a foot race.
  • He knew Campo had not seen Phillips' daughter's belly-dancing in Charlie Wilson's War and would not taunt him about it.
  • He wanted someone who knows what it is like to be fired as head coach of the Dallas Cowboys.

Who Could Leave?

Believe it or not, the Cowboys do have some non-coaches in their organization. Their 2004 second round pick, Julius Jones, is almost certainly gone as a free agent. As long as the team can keep restricted free agent Marion Barber, this could be a case of addition by subtraction. Barber was extremely good in 2007 with a 31.9 DPAR -- good for third among running backs. Despite not starting on his own team, Barber went to the Pro Bowl (along 12 of his teammates). He only carried the ball 204 times in 2007 and it will be interesting to see how he handles a larger load. Barber received the highest RFA tender ($2.562 million), so if a team wants to sign him to an offer sheet, they will have to give up first- and third-round picks. Despite Barber's great ability, this is too steep a price and it is highly unlikely any team would make that deal. However, the fact that the Cowboys have not yet locked up Barber to a long-term deal has made some speculate they are looking to trade him. In reality, this angle may not have much truth; there is no rush for the Cowboys to commit big money to Barber.

Outside of the running back position, the Cowboys secured the services of a pair of Pro Bowlers in re-signing tackle Flozell Adams and using the franchise tag on safety Ken Hamlin. Cornerback Jacques Reeves was snapped up by the Texans. Dime cornerback Nathan Jones is an unrestricted free agent and expected to leave. With two injury-prone corners left -- Terence Newman on one side and the aging Anthony Henry on the other -- the Dallas secondary looks very thin.

Whom Should they Sign?

While Barber is a great back, the Cowboys need to find a quality running back to platoon with him. The speedy Tatum Bell is an interesting complement to the bruising Barber. Considering this year's strong draft class and the fact that the Cowboys have two first-round picks, it is quite possible the Cowboys draft a runner early. Keep an eye on Felix Jones, Arkansas' other first-round running back. Jerry Jones loves former Razorbacks and if he cannot swing a deal for Darren McFadden, Felix Jones may be the next best thing.

The Cowboys should be considered dark horse candidates for Javon Walker. Even with a premier receiver in Terrell Owens and a new contract for Patrick Crayton, the Cowboys have doubts about their receiving corps due to Terry Glenn's questionable health. The Cowboys would prefer to have two top flight starting wideouts and Crayton as the third receiver.

On the defensive side, the Cowboys look very thin at cornerback, especially if Anthony Henry cannot stay healthy. They have already brought in one former defensive superstar, Zach Thomas, and it is possible they could pursue a veteran like Ty Law; although Law is nowhere near the player he once was, perhaps he could help out in nickel coverage and provide leadership. Another former Patriots cornerback with a monosyllabic last name they may have considered was Randall Gay, but the Saints offered him too much money for a third cornerback.

New York Giants

Everybody Stay Calm

Pretty much everything there is to say about the Giants' remarkable run has been said. It was spectacular, unpredictable, and exactly why people love the NFL so much. However, in the aftermath of the greatest upset of all time, a strange myth has taken root: Eli Manning is a star quarterback.

Now, there are a ton of different ways to define "star," but for the sake of this argument, let's keep it simple and say that a star quarterback is one of the eight to ten best in the game, essentially Pro Bowl level. Using this criterion, Eli Manning is not even close.

Last season Eli Manning was 30th in DPAR (total value) and 35th in DVOA (value per play). By definition, an average DVOA is 0%; Eli was at -10.2%, which is to say significantly below average. Now, in Eli's defense he did have a bit of a down year. His DPAR and DVOA were the worst of his three years as a full-time starter. In fact, he has gotten worse with each passing season (his career-high DVOA is 1.1%). Some have argued that Manning became a different player during the Giants' late-season run. We don't buy it. Manning didn't play nearly as well down the stretch as people think he did -- particularly in the first 58 minutes of Super Bowl XLII -- and based on past returns, we would still label his good but not great performance as fluky.

That's not to say Eli can't take a step forward next season. Fifth-year quarterbacks usually improve significantly, and Eli did seem to make some progress with his decision-making and accuracy. However, it would be a shockingly huge improvement if Eli has as good a year in 2008 as several less-hyped young passers, like Philip Rivers (14.2% DVOA) and Jay Cutler (19.7%), had in 2007.

So, to any ESPN talking head who wants to say that Eli is better at this stage in his career than Peyton, or that he has emerged as an elite quarterback, we can only say, "Dude, what the f**k?"

Who Could Leave?

The biggest personnel issue facing the Giants this off-season seems to be the future of the running back position. Derrick Ward, who was having a very nice season (8.1% DVOA) before getting injured, is an unrestricted free agent. The team seemed quite comfortable with the two-headed monster of Brandon Jacobs and Ahmad Bradshaw in the playoffs, but if Ward can be retained at a reasonable price then that would add further flexibility to a fragile and unproven group.

Already gone are linebackers Kawika Mitchell (Bills) and Reggie Torbor (Dolphins), plus free safety Gibril Wilson (Raiders). All were somewhat overvalued just by virtue of being starters on the Super Bowl champions, and the Giants made the right move to let them go. The Giants did re-sign the kicking duo of Jeff Feagles and Lawrence Tynes, keeping their 20th ranked special teams intact at a moderate cost. Both restricted free agents -- starting strong safety James Butler and backup quarterback Jared Lorenzen -- were tendered and are expected to return.

Of course there is also the omnipresent Michael Strahan retirement saga and the persistent rumors that Jeremy Shockey is on the way out. I sincerely doubt Shockey is going anywhere because there don't seem to be any strong suitors for his services. Strahan, on the other hand, should walk away on top, but probably won't, especially if he is allowed to skip training camp again.

Whom Should they Sign?

After last off-season yielded only one free agent signing and a Super Bowl victory, don't expect the Giants to be particularly active in free agency -- despite the rumors about a potential trade for DeAngelo Hall. It seems unlikely that such a deal will materialize, and the Giants seem fairly intent on maintaining the team chemistry that was integral to their playoff run. Some offensive line depth would be nice, and even better would be a tackle who would allow David Diehl to return to guard, strengthening two positions. However, it doesn't seem likely that they will be able to find that upgrade in free agency.

Philadelphia Eagles

The Unheralded Superstar

The most valuable contribution to the Eagles during 2007 may have come from the media relations department. Miraculously, they made Andy Reid's "drug emporium" story all but disappear once the season got rolling.

However, the most valuable on-field contribution came from Brian Westbrook, who led all running backs in DPAR with 46.5. Like Tiki Barber before him, Westbrook appears to be an undersized back who is getting better with age; he turns 29 just before the start of the 2008 season. Whether Westbrook can remain as productive as Barber through age 31 remains to be seen, but the fact that he is relatively fresh in terms of career carries (1,014 in six seasons) and uses the same trainer as Barber suggests that he should not slow down during 2008. To put it in perspective, Jamal Lewis is four days older than Westbrook, but has more than twice as many pro carries.

With FO binky Kevin Kolb breathing down his neck and leaving nasty notes in his locker, Donovan McNabb went with the full head shave and staved off the grim reaper for the year. McNabb was the 14th rated passer with a DPAR of 52.5 in 14 games, deceptively maintaining his 2006 form (57.0 DPAR in 10 games). However, a closer look reveals a marked drop in DVOA: 25.7% in 2006, 9.9% in 2007. McNabb may hold up another year, and President Joe Banner has adamantly committed to him for the 2008 season, but it is clear their long term plan involves Kevin Kolb. The Kolb pick very much fits with the Eagles style of drafting at positions of strength, preparing players for a couple years, and then inserting them into the lineup when the original players get old. Cornerback Lito Sheppard started nine games in his second year and became a full-time starter in year three; considering Kolb is a quarterback, it's reasonable to push that timetable back a year for him.

Who Could Leave?

The Eagles have shockingly few potential losses this year, with only four unrestricted free agents on the team. Fullback Thomas Tapeh has already signed with Green Bay, and while he did an excellent job clearing space for Westbrook, the Eagles clearly don't think he will be particularly tough to replace. The other three -- William James, Kimo Von Oelhoffen, and Reno Mahe -- were bit players and don't seem to be at the top of the team's priority list.

Some have questioned the decision to put the franchise tag on L.J. Smith, seeing as he is hardly a franchise player (37th in DPAR). However, the tight end franchise number is quite low, only about $4.5 million, and the Eagles are clearly satisfied with L.J.'s performance as a run-blocker. The move may pay dividends by maintaining continuity on a veteran offense.

In other moves, the Eagles released Jevon Kearse, whose continued existence was news to many fans. They decided to keep both of their restricted free agents, agreeing to a one-year deal with safety/kick returner J.R. Reed and tendering a qualifying offer to cornerback Joselio Hanson.

Perhaps the biggest looming loss is the potential trade of Lito Sheppard -- that is, it is the biggest looming loss depending on which Sheppard is being traded. Sheppard is one of the more talented young corners in the league, but he's been horribly inconsistent over the past three years. (His Success Rate, according to FO game charting, has gone from 42 percent in 2005 to 60 percent in 2006 to 41 percent in 2007.) Sheppard's starting job has been given to new arrival Asante Samuel (see below) and the Eagles are said to be open to moving Sheppard if an agreeable contract extension cannot be reached, as Eagles prefer not to allow key players to enter the final years of their contracts unsigned.

Whom Should They Sign?

During his introductory press conference, Asante Samuel commented, "The fans have always been supportive of the team ... that's why I picked the Philadelphia Eagles." Has he ever been to Philadelphia? Considering the fans are notoriously harsh, it sounds like he picked the Eagles for another reason: Money. Not that there's anything wrong with that. The Eagles seem to be feeling quite generous this off-season, shelling out the big dollars for Samuel (6 years, $57 million) and Chris Clemons (5 years, terms undisclosed).

With those two in the books, and relatively few potential losses, the Eagles don't have a lot of needs. Obviously a wide receiver would be nice, and the Eagles made a big last-day run at Randy Moss before he re-signed with New England. It was reported that they had interest in Javon Walker, but that does not seem likely to materialize. Most likely, the Eagles are hoping that the addition of Samuel and slight upgrades in other areas will be enough to get them back to the playoffs in 2008.

Washington Redskins

Quarterback of the Future?

So far the big news of the Redskins off-season has been the hiring of Seattle Seahawks quarterbacks coach Jim Zorn as head coach. Zorn is, by all accounts, a good guy and a good coach, although it remains to be seen whether he is capable of making the leap from position boss to top dog. The move went against Daniel Snyder's traditional philosophy of hiring the biggest name available, but that's no guarantee it's going to work out.

The biggest question raised by this move is how quarterback of the future (and present), Jason Campbell will fit in Zorn's West Coast offense. My guess -- and this is David Lewin writing this section -- is not well. I firmly believe that Campbell has the talent to be a Pro Bowl quarterback. He was 20th in DVOA last year (7.0%), far ahead of one Elisha Manning (-10.2%, 35th), and if you asked me which one I would take I'd say Campbell without hesitating. However, Zorn's hiring means it's not likely that Campbell will reach his potential with the Redskins.

Campbell's strength is his strong arm and ability to throw intermediate and deep passes with accuracy. Criticism of Campbell generally focuses on his slow delivery, and his imprecision on timing and underneath routes. Campbell rarely had to throw short in college because the presence of Ronnie Brown, Cadillac Williams, and Brandon Jacobs made it unnecessary, so it is understandable that he is somewhat underdeveloped in this area. Given these limitations, it seems unlikely that the West Coast offense will be a good fit.

Since his Auburn days, Campbell has been in a different offensive system virtually every year. He seems to have done OK picking them up, but some people have described him as a slow learner. The combination of a new system to which he is poorly suited, a new coach whose public statements about Campbell have been lukewarm at best, and the presence of a proven veteran backup on the bench could spell trouble for Campbell in 2008.

Who Could Leave?

Using creative cap maneuvers fueled by owner Dan Snyder's deep pockets, the 'Skins should have little trouble getting under the cap. Reworking the deals of Chris Samuels, Santana Moss, and others will guarantee them more money, but save on this year's cap. With Todd Collins already re-signed as a free agent, return man Rock Cartwright appears to be the only player the Redskins will have to fight to keep. High-profile bust Brandon Lloyd was cut last week and Mark Brunell, a Joe Gibbs favorite, will likely retire.

Whom Should They Sign?

In chronological order, here are Snyder's highest profile pickups since 2000: Bruce Smith, Deion Sanders, Laveranues Coles, Clinton Portis, Shawn Springs, Antwaan Randle El, Brandon Lloyd, Adam Archuleta, Fred Smoot. The names are becoming less impressive. Recent pickups have not had much greater success, but at least the cost of whiffing is lower. Snyder and Vinny Cerrato, newly promoted Executive VP of Football Operations, are always a threat to throw off the pay scale for the entire NFL when they decide to crack out Snyder's checkbook, but recently they have shown much more restraint. Expect that to continue this year as they have little cap space to work with and seem to finally understand that it is next to impossible to build the core of a team through free agency. The Redskins have publicly declared they will not make any splashy acquisitions, but from the guys who brought you the meteoric rise of Jim Zorn, do not totally rule out a high profile wide receiver signing such as Javon Walker. Their recent experiment in tiny wideouts has failed, and they may look for a big target for Jason Campbell.

Daniel Adler and David Lewin are an Ivy Leaguer and a quarterback, but when combined, they do not become Jason Garrett, Ivy League quarterback. We thank them for pitching in on the final Four Downs.

Posted by: Guest on 04 Mar 2008

205 comments, Last at 12 Nov 2010, 10:44am by Glandos

Comments

1
by Jon (not verified) :: Tue, 03/04/2008 - 12:10pm

Ugh, more stirring the pot with Manning.

When you have Tim Hasselbeck, Phil Simms, Kurt Warner, Boomer Esiason, Jim Miller, et al all in agreement that the Giants' offense is the most QB-unfriendly in the NFL, I can't help but be skeptical of QBs who run paper and scissors offenses like Cutler. It's no coincidence at all that Eli started playing like the QB that we all expected him to be when Coughlin and Gilbride tossed the playbook out the window in week 17. Doesn't hurt that Steve Smith returned from injury, and Toomer finally decided to stop dropping passes. (And no, Shockey being hurt didn't help at all, watch the Buffalo game tape again if you somehow believe that.)

2
by James, London (not verified) :: Tue, 03/04/2008 - 12:18pm

I hope this thread doesn't get ugly, but I'm not optimistic.

3
by Harris (not verified) :: Tue, 03/04/2008 - 12:23pm

Why, if you're Daniel Snyder, do you hire a no-name coach who's system doesn't mesh with your QB's abilities? I could understand it, sort of, if he hired a guy with a well-known pedigree, but why make it harder for a guy you traded two picks to get?

On another note, I'm not quite as convinced as most that Sheppard is definitely out. He's signed until (I think) 2011 at a reasonable number, so they don't HAVE to move him. I think that locker room malcontent stuff is over-rated. Of course, if they swap him and the #19 pick to jump up and take McKelvin I'd be quite pleased.

4
by JasonK (not verified) :: Tue, 03/04/2008 - 12:34pm

I think you've got the wrong backfield targeted as the Giants' biggest personnel issue. The RBs are fine-- some combination of Jacobs, Bradshaw, Droughns, and either Ward, a rookie, or cast-off from another team will be perfectly adequate.

The DBs are the biggest question on the team. At safety, Gibril Wilson walked for huge money, James Butler wasn't that good to begin with (although it's said that he was playing through a hamstring injury for much of 2007), and Michael Johnson is promising but untested. At corner, Aaron Ross looks like the only sure thing. Sam Madison and R.W. McQuarters will be another year older, and, much like Eli Manning, it's tough to say whether Corey Webster's late-season improvements (he caught the last pass Brett Favre threw as a pro!) are a fluke or a herald of things to come. Kevin Dockery is an okay nickel guy, but doesn't have the size you want in a starter.

They signed Sammy Knight yesterday, which will help provide a "floor" at the safety position, in case Butler, Johnson, and/or whoever they draft aren't up to snuff. There may not be any more moves in this area until the draft, but that doesn't mean that their secondary is anywhere close to settled.

5
by Bruce (not verified) :: Tue, 03/04/2008 - 12:37pm

Cudos on Campbell. It's just hard to think he will thrive in a West Coast system. And it was just hard to not be excited about the play of Todd Collins during the playoff run. He hit 3rd and 4th options continually. His strength was his mastery of Saunder's system. You seriously wonder if he could do the same under Zorn. So what am I saying? I'm saying 6 or 7 wins unless Campbell can learn to throw on time in '08. Here's praying Jim Zorn is coach of the year.

6
by BHold (not verified) :: Tue, 03/04/2008 - 12:38pm

We get it, FO as a whole has a strong hatred for Eli and always will. You don't need to keep screaming it. Oh, and stop trying to use DVOA to rate individual players! It is an interesting way to rate teams, but to try and gauge individual players with it is quite misleading.

7
by Fergasun (not verified) :: Tue, 03/04/2008 - 12:41pm

Re: Jason Campbell
Do you know that one of the offenses he ran at Auburn was a WCO-type offense and the season he ran it Campbell completed 70% of his passes? The OC at Auburn was Al Borges at the time.

Anyway, we'll see. I hate the offseason and the media coverage of it. It's really over-rated. No matter who joins or leaves the Redskins, I'll still root for the Redskins in the fall. The most important roster moves are not the 1st round draft picks or the big-money free agents, but 3rd and beyond draft picks and UDFAs. Yet I think 90% of the media focuses on the big names. To be honest, I didn't read many FO articles (I just perused a few of Doug's columns) and they don't look to be too bad... but this is the Dead Sea period for real NFL fans. There's stuff going on but it's not nearly as exciting as when the players put the pads on and play week-in-week-out... must be good for them to get a long break.

8
by Dave (not verified) :: Tue, 03/04/2008 - 12:46pm

No salary cap numbers?

9
by Ray (not verified) :: Tue, 03/04/2008 - 12:53pm

Lets see if we can objectively assess the impact of bias on the story line by FO authors. For the sake of argument, lets assume they are biased. Then according to this analysis, I conclude they are biased....
Whats dissapointing is the editorial choices. An inability to state anything other than the obvious, and pretend its an improvement over the mind-numbing stuff from major media.
If there are no positive objective findings from '07s outcome to apply to '08, then I would suggest there is no value to this statistical method.

10
by S.Bialo (not verified) :: Tue, 03/04/2008 - 1:03pm

Dear FO Authors: Your methods suck.

Sincerely,

Someone who knows very little about said methods but doesn't agree with your conclusions.

11
by JasonK (not verified) :: Tue, 03/04/2008 - 1:03pm

#8:

You can't hope to come up with accurate salary cap numbers until all of the details of the signings and re-negotiations in the last week are public.

12
by Ray (not verified) :: Tue, 03/04/2008 - 1:17pm

The failure of FOs methods bodes well for Eli. Here's a large wager that he significantly outperforms both Rivers and Culter.

13
by Yakuza Rich (not verified) :: Tue, 03/04/2008 - 1:21pm

Terence Newman and injury prone CB? He's missed 1 game in his entire career. If he's injury prone than Dusty Dvorcek is legally dead.

14
by GatoradeDunk (not verified) :: Tue, 03/04/2008 - 1:22pm

The lead-in to this story in my RSS reader says the NFC East is "a division so strong that the worst regular-season team in 2007 goes into 2008 as the defending Super Bowl champion" which is odd, because I'm pretty sure the Eagles, at 8-8 were the worst team in the division, and they're not the defending Super Bowl champions. It's one thing to value DVOA highly; it's another thing to value it above actual standings.

15
by RMGreen (not verified) :: Tue, 03/04/2008 - 1:33pm

Usually I don't have a problem with the 'anti-Eli' bias since said bias is founded in FO's stats. However, it doesn't seem entirely appropriate to me for the author of a supposedly-professional article to refer to a player derisively (eg 'Elisha' Manning).

16
by RMGreen (not verified) :: Tue, 03/04/2008 - 1:36pm

RE: 14 The lead-in to this story in my RSS reader says the NFC East is “a division so strong that the worst regular-season team in 2007 goes into 2008 as the defending Super Bowl champion” which is odd, because I’m pretty sure the Eagles, at 8-8 were the worst team in the division, and they’re not the defending Super Bowl champions. It’s one thing to value DVOA highly; it’s another thing to value it above actual standings.

Actually, the whole point of FO's custom stats like DVOA IS to go beyond the standings in determining a team's quality. If the Giants graded out as the worst team in the league in 2006 in DVOA, there's no reason why they shouldn't reference that number now.

17
by RMGreen (not verified) :: Tue, 03/04/2008 - 1:39pm

Oops! Sorry for the double-post, but I obviously got some years mixed up in my head there. I'll rephrase that "If the Giants graded as the worst team in the NFC East in 2007 according to DVOA, then there's no reason they shouldn't reference that number now."

18
by Temo (not verified) :: Tue, 03/04/2008 - 1:44pm

14. What? DVOA was created in part because of the recognition of the fact that wins and losses don't completely describe the strength of a team. It's completely valid to use that here and say that DVOA rated the Giants as the worst regular season team in the division.

19
by Ray (not verified) :: Tue, 03/04/2008 - 1:47pm

Obviously, they should qualify that intro as "according to our statistics..." so that we know its just meaningless BS.

20
by JasonK (not verified) :: Tue, 03/04/2008 - 1:47pm

#15:

Eli is short for Elisha. It's his name: Elisha Nelson Manning. (see link on my name)

21
by GatoradeDunk (not verified) :: Tue, 03/04/2008 - 1:50pm

RE:16 Actually, the whole point of FO’s custom stats like DVOA IS to go beyond the standings in determining a team’s quality. If the Giants graded out as the worst team in the league in 2006 in DVOA, there’s no reason why they shouldn’t reference that number now.

I guess my point is, if your stats grade out the eventual Super Bowl Champs as the worst team in the division, maybe it's the stats that are flawed rather than the actual games on the field.

22
by Rusty G (not verified) :: Tue, 03/04/2008 - 1:59pm

Jerry Jones loves former Razorbacks? Who, besides Ken Hamlin?

Is FO falling for the "report media rumors and talking points as fact" game too now?

Would be a pity, if so.

23
by Ted Kerwin (not verified) :: Tue, 03/04/2008 - 2:00pm

Giants, Who Should they sign?

Ummm, maybe actually giving us free agents that would fit these needs you describe would be useful. I know you waited until five days had passed to post this article but it seems useless as a point of discussion if you don't give us real names to discuss.

24
by RMGreen (not verified) :: Tue, 03/04/2008 - 2:28pm

RE #20

Wow. Dave Lewin, my sincere apologies. I ought to do more research before writing on the internet. And I'm sorry to Eli to, who's had to go through life with a name that looks like it should have been given to a girl.

25
by GatoradeDunk (not verified) :: Tue, 03/04/2008 - 2:34pm

RE:#24

Elisha is Archie Manning's given name, too. It's a male name, all the way back to the Bible.

26
by John Doe (not verified) :: Tue, 03/04/2008 - 2:39pm

Manning did have a pretty crappy regular season, in conventional stats as well. I believe he had a 56% completion percentage, 20 INTS (compared to 23 TDs), and a QB rating of 73.9. The numbers are what they are, and I'm a Giants fan.

Then again, his playoff run was very nice, I'll trade regular season stats for a post-season performance like that any day. As far as his 'poor' performance in the SuperBowl, if you remove the INT that Steve Smith shoveled from 2" off the ground to 6' in the air (in other words, obviously no way that ball is intercepted due to Eli's end of the deal) then he outplayed Brady.

So bash Eli and the Giants all you want, they still have the rings. I feel sorry for any Pats fans, Brady is going to get hammered next year. I have a feeling he'll be eating through a tube around mid-season after every team sends the house on every play.

27
by Ray (not verified) :: Tue, 03/04/2008 - 2:42pm

#21: That's the point. Comparing a teams play-by-play performance to that of an average teams in the same situations, and then adjusting it for the quality of a teams opponent, creates a metric of no redeeming value. Interesting idea, but it failed.

28
by John Morgan (not verified) :: Tue, 03/04/2008 - 2:45pm

#22, Yeah, I'm only finding one Razorback ever drafted by the Cowboys: Jerry Dorsey in 1970.

29
by young curmudgeon (not verified) :: Tue, 03/04/2008 - 2:46pm

Elisha is the name of an Old Testament prophet, the successor to Elijah. Not too many of them were females.

It's interesting how quickly a name becomes "feminized." Elisha is fairly recognizable as a male name, even sounds like the unmistakably male Elijah, but as soon as one fairly recognizable woman (assuming that Elisha Cuthbert is fairly well known)gets that name, immediately people categorize the name as feminine. Same thing happened to Ashley (remember Ashley Wilkes in Gone With the Wind)and probably to others I could think of if I actually had a functioning memory. There's probably some interesting and meaningful point about sexism, etc. in this, but you guys might prefer to talk about football.

30
by JMNeu (not verified) :: Tue, 03/04/2008 - 2:54pm

24: Elisha = old testament (male) prophet. See First and Second Kings.

31
by fogarty (not verified) :: Tue, 03/04/2008 - 3:03pm

Small nitpick: FB Thomas Tapeh (formerly of the Eagles) has signed with the Vikings, not the Packers

32
by David (not verified) :: Tue, 03/04/2008 - 3:10pm

I guess my point is, if your stats grade out the eventual Super Bowl Champs as the worst team in the division, maybe it’s the stats that are flawed rather than the actual games on the field.

Or maybe they genuinely played better in the postseason than the regular season?

33
by shocker (not verified) :: Tue, 03/04/2008 - 3:21pm

#21, 32
They certainly did play better in the postseason. However, I don't think the DVOA grades were all that great comparative to other postseason teams as well. There are just some parts of a game like football that aren't captured well in stats. Some of this is luck, but also some of it is timing, and being able to play your best at the most important times.

Eli is not a great QB, yet. He may become one, or may not. I for one do not think he is as bad as his DVOA/DPAR suggest. But the stats are what they are. Eli made many mistakes this season, but often the team did not let those mistakes kill them (except vs. Min). He made plays in several games where his play won the game late (Was, Chi, NE). But he's as far from the only reason they won as he is from being a detriment to the team.

34
by 2 Kings 2:23-24 (not verified) :: Tue, 03/04/2008 - 3:50pm

[Elisha] went up from [Jericho] to Bethel, and while he was going up on the way, some small boys came out of the city and jeered at him, saying, "Go up, you baldhead! Go up, you baldhead!" And he turned around, and when he saw them, he cursed them in the name of the LORD. And two she-bears came out of the woods and tore forty-two of the boys.

35
by FullmoonoverTulsa (not verified) :: Tue, 03/04/2008 - 3:57pm

Wade is a retread, Jerry loves former razorbacks. So insightful.

36
by Gozer Gozarian (not verified) :: Tue, 03/04/2008 - 4:18pm

I really think that the Giants are the first case of the reverse FOMBC curse.

37
by mawbrew (not verified) :: Tue, 03/04/2008 - 4:29pm

Re: 34

Proving once again that some people just can not take a joke.

38
by mawbrew (not verified) :: Tue, 03/04/2008 - 4:41pm

Where I think the ELi analysis is weak is the dismissal of apparent improvement from week 17 through the Superbowl. I think that's really debatable and if they're going to suggest his play from then on wasn't as good as perceived they ought to be able to provide some evidence to support that point.

It sure looked to me like Eli did a much better job protecting the ball at the end of the year. Other than the previously mentioned Smith bobble in the SB and one late pick in week 17, he didn't turn the ball over. And he did it without sacrificing their downfield passing game. It's certainly possible that those five games were a small sample size fluke, but it's just as possible they represent a new baseline for performance.

39
by Dean (not verified) :: Tue, 03/04/2008 - 4:59pm

1) Tapeh signed with Minnesota.

2) RE #3 - you are correct. Shepherd is signed through 2011. But is not happy. If they were to trade him to move up, I can think of a lot of guys who can help us more then McKelvin. I'd rather stay put and get a guy like DeSean Jackson, who even if he brings us nothing as a WR, at least he'll be able to return a kick more then 10 yards. If not him, maybe Kenny Phillips falls into our laps?

3) "notoriously harsh" - and the bashing of Philly fans continues. Demanding? Sure. But you'll not find more passionate, more knowledgable, more die-hard fans anywhere.

We now return you to outraged Giants fans.

40
by David Lewin (not verified) :: Tue, 03/04/2008 - 4:59pm

I just find the whole Eli Manning business interesting because in PFP 2007 I took what was, at the time, a pro-Manning stance saying that people should lay off him because he's actually an average to above average QB who you can win with. Now, coming off the worst full season of his career I take the exact same position, and it's perceived as anti-Manning. Personally I don't think 4 playoff games are that much more important than 4 years of mediocre performance, but I understand if people want to disagree.

41
by Dales (not verified) :: Tue, 03/04/2008 - 5:15pm

"I really think that the Giants are the first case of the reverse FOMBC curse."

Ha!

FOMBC-- the sword that cuts both ways!

42
by Mike (not verified) :: Tue, 03/04/2008 - 5:22pm

One tidbit on J. Campbell.

His best year in college is under a WCO in 2004. That year he was SEC offensive player of the year.

43
by BHold (not verified) :: Tue, 03/04/2008 - 5:29pm

#40
This year was not close to Eli's worst year as a starter. Try watching his play instead of blindly following what you can learn through nfl.com play-by-play. The things that matter for a QB improved tremendously over this past offseason for Eli (footwork, accuracy, read progression, sense of pressure).

It is hard to say why his statistical performance departed from his apparent improvement so widely at times this year. Some of it is due to the Giants' offensive system being unneccessarily complex and filled with option routes until seemingly December. Some of it was the Giants having the most drops of any team in the NFL in the regular season. Some of it was having to play a number of bad weather games. It is hard to say what the largest reason was.

44
by DanielA (not verified) :: Tue, 03/04/2008 - 5:59pm

Re: #13

If you're going to attack an article for being wrong, at least look at the guy's stats from this season. Maybe injury prone is a little strong, but it is fair to say Terence Newman battled injuries throughout this season. For the first time in his career, Newman did not start every game he played in (due to the fact that he was not healthy). Also, he only played in 13 games during the regular season (11 starts). Furthermore, his injury struggles were widely reported throughout the season.

45
by Chris (not verified) :: Tue, 03/04/2008 - 6:05pm

The thing that jumps out to me is the fact that D-Lew would rather have an Eli Manning with a ring on his finger than a Jason Campbell who was benched late in the year in favor of career backup Todd Collins. Campbell was healthy enough to play but the skins wanted the "momentum" of keeping Collins in.

Blindly comparing their DVOA is absurd. The Redskins had the most vanilla offense in the league full of WR screens, 2 yard fullback passes, and 5 yard tight end drags. The guy didn't even call audibles, and choaked at the end of numerous games ( Tampa, Dallas, Philly, Giants 1 stick out in my mind).

In DVOA land the Giants are the worst team in the NFC East... back in the real world they are super bowl champs.

46
by Chris (not verified) :: Tue, 03/04/2008 - 6:14pm

Also, the biggest question for the Giants is the RB position? They had 4 backs play a signifigant role last year and with exception of the garbage Reuben D. they all had much success.

That was after they were supposed to collapse last year with Tiki and his huge pct. of that offense leaving. AND offsides Luke moving to Tampa.

I post it almost everytime there is a Giants article, but MANNING MAKES THE RUN GAME BETTER with his audibles and presnap reads. What is so hard to understand about that? Before every snap he points out the Mike and calls audibles accordingly ( as opposed to many QB's who just line up and run the play ( Like Jason Campbell )).

People said the Colts were screwed when they lost Edge, and that the Giants were when they lost the Marquee Tiki. Both teams won the SB and had built up the stock of Addai/Rhodes, and Jacobs/Bradshaw/Ward.

The Giants are a great example of a team with a fungible RB situation. The key NEEDS will be OLB, S. James Butler is a grade D player, and Gerris Wilkinson can only step into one OLB slot and although looked good in limited time, he is inexperienced. They have to draft/sign a stud OLB and a smart Gerry Reese will.

47
by Joe T. (not verified) :: Tue, 03/04/2008 - 6:23pm

Elisha Manning is not disparaging or derogatory. Now, if the authors had referred to him as Elisha Hussein Manning, well, that would be something different altogether.

Watching Campbell play I got the impression that WCO wouldn't be the best fit, but, if what some other posters are saying is true that Campbell played collegiate WCO in '04, then maybe there is yet some hope.

48
by NY expat (not verified) :: Tue, 03/04/2008 - 6:48pm

re: 4

I agree the secondary is still the biggest worry, though there were a couple plays in the Super Bowl where Adalius Thomas went by Diehl in a manner reminiscent of Suggs going by Diehl in the preseason game. I suppose Whimper has been in the "developing" stage, but certainly isn't a sure thing, even if it would be nice to see him "overcome" his name :). Also, I'd think that a linebacker who can cover would be pretty important, especially with covering Cooley and Witten for 4 games next year.

Here's hoping Reese & Co come up with a draft class even half as good as the last one. I think given the level of success they've had, sticking with a few modest deals like Knight's seems perfectly reasonable.

49
by goathead (not verified) :: Tue, 03/04/2008 - 7:18pm

Re: The Giants, Eli, and the Running Game.

I've finally realized that the FO staff are actually writing Trolls for the Giants fans. Quite simply, they seem unable to objectively analyze the "worst team ever to play in the SB" Giants.

To discuss Eli without mentioning the dropped passes is insanity. To forget to mention that his #1 receiver was hurt all year and didn't practice? Right, a little oversight. Coming into 07, Eli's accuracy was horrendous. He DID have some bad moments in the last season (and one god-awful game against Minn), but on the whole his play was improved ALL year, and above average to good in the playoffs.

Between Jacobs, Ward, and Bradshaw, the running game seems to be as good as anyone. Also, the Giants had solid play at FB through the late season and playoffs.

The 2ndary looked to be as bad as any in NFL history through 2.5 weeks, then came together and played well enough through the rest of the season.

8 rookies played in the playoffs.

Honestly, this team seems potentially on the verge of actually putting up some good DVOA #'s in 08, and really that's what its about, isn't it?

OK, so I took the bait. 17-14

50
by JasonK (not verified) :: Tue, 03/04/2008 - 7:18pm

#46:

I don't see OLB as that big a concern for the Giants. Torbor was only starting because of Kiwanuka's injury. With Kiwi back, he'll be the starter on the Strongside (no, he's not moving back to DE), and Zak DeOssie will likely replace Torbor as the #2 at that spot. On the Weakside, Wilkinson looks ready to step in. They need to draft for depth, not for a starter. I'd be surprised if they take a LB before the 3rd round.

#48: I'm happy with the Giants' OL. Yeah, Diehl could be improved on, but you don't find young starting-caliber OLTs available in free agency. He's adequate (and a salary cap bargain), and they've got 3 developmental projects behind him in Whimper (who has played well enough as a fill-in to be the #3 OT), Koets, and Na'Shan Goddard. In the extraordinarily unlikely event that there's a phenomenal OLT prospect available at pick #31, by all means they should take him, but otherwise, they should be fine sitting tight.

The only OL need I see (besides an extension for Snee)is for a young C/G to develop to replace O'Hara in 2-3 seasons.

51
by Brandon (not verified) :: Tue, 03/04/2008 - 7:36pm

44: If I recall correctly, Newman dealt with a foot injury that happened during training camp/preseason for most of the season that he played through simply because the Cowboys needed him to. One year of being injured does not make a player injury prone. It just makes him unlucky for that season.

52
by David Lewin (not verified) :: Tue, 03/04/2008 - 7:45pm

I am aware that Campbell had an excellent year in the WCO under Borges at Auburn. However, he only threw 270 times during that season. He had high yards per completion and completion percentage numbers which makes it tough to tell how much he was throwing long v. short.

It is my recollection that that Auburn team eschewed short passes in favor of runs by Brown and Williams and that Campbell mostly took shots downfield in response to teams stacking the box. He was extremely effective, but I'm not sure that Zorn's system is going to be that similar. I hope so though, I think Campbell is an excellent talent.

Also, I don't think anyone meant to impugn the Giants backfield. I personally love Brandon Jacobs. Still, letting Ward get away would be a mistake.

53
by Matt (not verified) :: Tue, 03/04/2008 - 7:52pm

Irrational. Discussion of Eli Manning is. Just like his tight end, who runs the wrong route a LOT. Fueled by his blonde hair, he never actually has to produce - he's irrational. Just like discussion of his quarterback.

54
by Jon (not verified) :: Tue, 03/04/2008 - 9:09pm

Matt, if you've been watching the Giants the past few years, Shockey is hardly the worst offender in that category. That's why the Gilbride/Hufnagel offense of the past few years has wrecked havoc on my blood pressure and general well being.

I'm pleasantly surprised to not see much Shockey bashing here so far. It's no coincidence that as soon as he went down, Brandon Jacobs became a lot less effective. In fact, according to a recent article in Newsday by Glauber I believe, the Giants think Shockey is the best blocking TE in the league.

55
by Chris (not verified) :: Tue, 03/04/2008 - 9:30pm

46- I was curious about Wilkinson because of the things I heard about him. He played many positions in college and was a team captain at GT. He was a smart player and passing the Giants legendary writing/psyche tests solidify that theory.

I was VERY impressed with Wilki stepping in the Pats game and the other time he got but he is still a question mark. On the other side, I am not sure if Deossie is ready as the only cut I saw was on special teams. I actually like Chase Blackburn as the eventual replacement at MLB, and he is a special teams tackeling machine.

I honestly doubt Kiwi will be playing SAM next year. He is too tall and lacks the awareness/quickness in the passing game. He was abused in the first two games of the year and if Strahan doesn't come back I'd bet dollars to donuts he is starting at DE.

I think the loss of Gibril Wilson will sting as he was probably the most underrated Safety in the league. I know somebody will chime in about Adrian Wilson but most more than casual fans know he's a stud. Wilson will hurt but I understand why they didn't match that huge deal ( 3rd highest paid S in the league).

With Wilson gone, and Butler not even an average player I want the Giants to spend their first draft pick on a stud S or stud OLB. Maybe Zibikowski " don't call me rudy" from Notre Dame ?

Sammy Knight is an aged strong safety that was once known to be a hard hitting tackeler, but weak in coverage. That doesn't bode well with a weak in coverage James Butler. The Giants MUST upgrade the S spot, or else that defense is either going to be giving up big plays or sacking the quarterback. A boom/bust defense if you will.

56
by Harris (not verified) :: Tue, 03/04/2008 - 9:35pm

#39 I like DeSean Jackson too, but the guy is an elf and I can't justify spending a first round pick on a player who's primary, if not only, contribution will be on special teams. McKelvin lead the nation in punt returns last year and he'd be the nickleback before eventually becoming the starter. I wouldn't mind Kenny Phillips either and he's got some experience as a PR (I think he returned two for TDs) but he's the only safety with a first-round grade. If he's gone at #19 where do they go? I doubt Talib or Jenkins will still be around, though Rodgers-Cromartie might. I'm not in love with any of the WR prospects and I don't think OL is an area of great need. Maybe Trevor Laws, the DT from Notre Dame?

57
by Chris (not verified) :: Tue, 03/04/2008 - 9:35pm

Jason Campbell is NOT a good fit for a west coast offense. His strength is a strong arm, height and his size. That made him ideal for that "game manager" role he had last year, or for a more downfield passing offense.

His weaknesses are his slow delivery, lack of experience, the ability to get flustered late in games " or drives", any redskin fan can attest that he was not very good at running 2 minute drills. He threw a lot of interceptions late in games and crunch time, and he failed to manage clock on 2 minute drills ( like Ending the Giants game on the 1 yard line).

His slow delivery doesn't bode well for a quick passing west coast offense. He is built more for what Jacksonville tried to do with Leftwhich, and less for a West Coast offense like Tampa, Seattle, or Green Bay.

58
by Chris (not verified) :: Tue, 03/04/2008 - 9:41pm

Harris- Does a rail thin D. Jack remind you at all of Todd Stinkston? Speedy but built like a human pipe cleaner?

Do you think there is any chance at all that the Iggles land the big prototypical Limas Sweed out of Texas ?

I honestly think that one of the main things that destroyed the Eagles last year was the loss of the tight end in the middle of the field. Especially in that west coast offense. If Smith were healthy I believe the 07' Eagles were a playoff team.

I know you can play that " if only so and so wasn't hurt" game with Mcnabb, Dawkins, Westbrook played through injuries, but in that West coast offense the Smith loss laregely went under the radar.

I clearly rate the Eagles higher than the Redskins for next year, and I think if things played out just a little bit different Philly could have been a 10-11 win team. Just look at GB, Wash 1, Chicago etc. The 07' Eagles were the best last place team I can ever remember.

59
by Chris (not verified) :: Tue, 03/04/2008 - 9:55pm

Let's look at the things against Eli in 07'...

1. Playing in NY
2. His old RB dogged him before the year
3. His top WR didn't practice/played hurt
4. Toomer had the dropsies this year
5. Steve Smith was hurt most of the yr
6. Sinorce Moss was hurt most of the yr
7. Shockey played hurt then went on IR
8. He had 3 starting RB's
9. His receivers dropped more balls than any other WR Corps. in the league
10. He had a Guard that nobody liked playing LT
11. He ran a complex offense

With all of those challenges against him, he still managed to prove his doubters wrong and win the SB. Is he a top 5 QB... NO, but he wasn't as bad as many people around here would lead you to believe, and he is CERTAINLY better than Jason Campbell.

17-14... SB MVP

60
by Matt (not verified) :: Tue, 03/04/2008 - 9:59pm

No, but Shockey is the one who does ridiculously irrational things like spike the ball after picking up a routine first down, making him a better punchline.

Jacobs slowing down did have quite a bit to do with facing tougher defenses at the end. He was able to bull through Buffalo and the second half of New England Part I without Shockey.

61
by PaulH (not verified) :: Tue, 03/04/2008 - 11:03pm

Re: #7

Do you know that one of the offenses he ran at Auburn was a WCO-type offense and the season he ran it Campbell completed 70% of his passes? The OC at Auburn was Al Borges at the time.

I'm not trying to call you out on this one, but I'm afraid you are dead wrong. Borges had a background that was largely dominating by the West Coast offense, but he ran nothing of the sort at Auburn with Campbell.

The Auburn offense in 2004 was just like DL described it in the article. They had a power running game, and forced defenses to load the box in an attempt to stop that. Once that took place, they attacked the defense with a deep passing game.

If you don't believe me or DL, just look at the passing statistics for that year. Campbell averaged 14.3 yards per completion. That was the highest yards per completion in the SEC, and only Minnesota -- another offense with a power running game -- posted a higher yard per completion average on the national level (counting only BCS schools). The #1 receiver, Courtney Taylor, averaged almost 18 yards per completion, and the #2 receiver, Devin Aromashodu over 21 yards per completion. Even the tight ends averaged almost 15 yards per completion.

Obviously, you don't average that many yards per completion by throwing all of the short throws that come with the WCO. DL is right on this one.

62
by Jimmy in Oz (not verified) :: Wed, 03/05/2008 - 12:38am

To all those who forget and claim that Eli only turned the ball over once in the last part of the season and that was because Steve Smith bobbled it:

He threw right to Asante Samuel who then dropped it. That is not good work Eli, that is thank F ing God.

I'm a fan of Eli, but he is streaky. When he's bad, he is very very bad, not like bad meaning good, but like Alex Smith bad.

63
by Dice (not verified) :: Wed, 03/05/2008 - 12:54am

#39: Philly fans in all sports are notorious bad sports. Christ, the Vet was the only NFL stadium with a jail and courtroom. I assume the Linc continues that fine tradition? But what else can be expected from a from a city with a broken symbol and that booed Santa Claus. And the Mummers, let's not even start...I spent the longest two weeks of my life one night in Philly...

64
by Matt (not verified) :: Wed, 03/05/2008 - 2:01am

62. One mistake isn't exactly damning and there aren't a ton of CBs who would have come down with that ball anyway. Not that I'm convinced he's top ten now, but I'm willing to wait and see.

65
by foos (not verified) :: Wed, 03/05/2008 - 3:01am

Re: 55

I honestly doubt Kiwi will be playing SAM next year. He is too tall and lacks the awareness/quickness in the passing game. He was abused in the first two games of the year and if Strahan doesn’t come back I’d bet dollars to donuts he is starting at DE.

I'm pretty sure that Osi and Tuck will be the starting DEs.

66
by Alex (not verified) :: Wed, 03/05/2008 - 3:39am

I love how Lewin said in PFP 07 about Eli: "He won't carry the team, but you can win with him"

And then, the Giants win a SB against a team with the highest scoring offense in NFL history, by a score of 17-14, proving that Eli won't carry the team (or they would've scored more than 17 points), but that you can win with him (which they did). And then Giants fans are all like "See, Lewin, you were wrong all along. You just hate Eli, admit it!" Wow.

The lead-in to this story in my RSS reader says the NFC East is “a division so strong that the worst regular-season team in 2007 goes into 2008 as the defending Super Bowl champion” which is odd, because I’m pretty sure the Eagles, at 8-8 were the worst team in the division, and they’re not the defending Super Bowl champions.

Honestly, it should be pretty obvious that they didn't mean the division's worst team by record, because the team with the worst record doesn't go to the playoffs, and therefore doesn't win the Super Bowl. Unless they've added a third wildcard team to each conference while I wasn't looking.

It’s one thing to value DVOA highly; it’s another thing to value it above actual standings.

Right, because it's crazy talk to say that the Giants were better than the Dallas team that blew them out twice and had a better W-L record. Oh, wait...

Is it possible that the better team doesn't always win a given game, and that a team might even have a worse record than a division rival simply because of a harder schedule and worse luck? Could it be that an 8-8 team that lost a game to a 10-6 team by three points after its kicker hit a FG attempt off the right upright on the final play of the game might possibly have gotten slightly unlucky? Perhaps, had the kick been slightly to the left, they might have won in OT, and then they would have had the same record as this other team. Maybe then, it wouldn't seem so crazy to think that that team was better than the other one. Is it possible that things happened in a slightly flukey way, in that game and/or others throughout the season?

Nope! All game outcomes, W-L records, and playoff results are preordained by God. There is no such thing as luck, and all "non-predictive events" are simply the result of the superior swagger, moxie, grit, and "smart, team oriented football" of the team that they benefit. Any attempt to predict game outcomes with objective statistical analysis is completely and utterly futile.

"I guess my point is, if your stats grade out the eventual Super Bowl Champs as the worst team in the division, maybe it’s the stats that are flawed rather than the actual games on the field.

Or maybe they genuinely played better in the postseason than the regular season?"

Now that's just crazy talk. I mean, sure they won playoff games against three teams that they had gone 0-4 against in the regular season. Sure, they outscored those teams by an average of 3 points in the postseason, while being outscored by an average of over 11 points by those teams in the regular season. But they didn't really play better in the playoffs. The scoreboard just better reflected their dominance when the postseason arrived. Yeah, totally.

67
by kevinNYC (not verified) :: Wed, 03/05/2008 - 4:00am

#66... Not to nitpick, but the Cowboys did not blow out the Giants twice. The first game was extremely tight and the second game was rather close. The closeness of those games was what led some (myself included) to think the Giants had a good chance at beating Dallas.

Just basing opinions on the final score (without really examining the game) is no more intelligent than basing everything solely on W-L record.

68
by Jimmy in Oz (not verified) :: Wed, 03/05/2008 - 4:28am

64: Eli Manning ranks between Dilfer and Grossman in completion % for active QB's.

If Asante Samuel catches the Int, it is a damning mistake because the Giants lose the game. Eli's line becomes 1 TD, 2 Int, under 56% passing.

Every year he loses the Giants a few games, but every year he wins more games for them, which is what separates him from Dilfer and Grossman.

I'm a fan of Eli. He's not the best but he doesn't try to dink & dunk and manage a game which is akin to trying not losing a game.

69
by Maryland Giant (not verified) :: Wed, 03/05/2008 - 4:54am

Re: Manning.

I'll give you guys credit for one thing...you have your story, and you're sticking with it!!!

LOL...I'm kind of bummed though to be honest, I had my heart set on a DVOA Title and all I got was this lousy "Super Bowl Champions" t-shirt. Oh well, maybe next year?!?!?!

70
by PW (not verified) :: Wed, 03/05/2008 - 9:09am

Well no offense... but FO said the Giants in rd 1 were the biggest mismatch in Tampa,then predicted every other playoff game, and Schatz then wrote an article not talking about why NE would win but blow out NY.

Now...the article begins with a teaser of NY being the worst team in the division and then states that Eli's post season run wasn't that good...and then even that overrated run was fluky.

Sorry if I can't help but feel that their is a subjective undercurrent of resentment I detect.

71
by Dean (not verified) :: Wed, 03/05/2008 - 10:33am

RE#56 I look at our corner situation and I don't see a guy on the roster over 5'10". Having watched Plexico maul our CBs, I want a guy with the size to match up with him. The two CBs with size are Cromartie and Talib - and sadly, I'm not sold that either will pan out. That's why I was looking elsewhere.

As for Jackson, we know Reid's not going to play the draft picks anyway, he never does. So a guy who gives us something on special teams is more then we get most years out of a rookie. Yeah, he's small, but I think he's going to be the rare exception - the little guy who pans out. If not him, then I wouldn't mind getting a big target like Limas Sweed. Yeah, he's got some injury concerns, and he's not going to return kicks, but he's 6'5" and he should be on the board when we pick.

Looking elsewhere, you can never have too many DEs. A guy like Calais Campbell would give us the freedom to cut Darren Howard - which would be addition by subtraction.

I do like the other guy you mentioned - Mike Jenkins - a lot. He's not necessarily the highest rated CB, but he's the guy I think will be the best pro of the bunch.

As for DTs, I'd like to see us add some guys there, but between Bunkley and Patterson, I don't think we need to make it a day 1 priority.

At least you're not suggesting linebackers!

72
by Chris (not verified) :: Wed, 03/05/2008 - 10:35am

65. Kiwi is taller, longer arms and is more of a straight away speed guy where as Tuck has that interior quickness. Tuck is the 2nd best DT pass rusher off the top of my head next to Tommy Harris.

Here is what I have penciled in...
WILL = Gerris Wilkinson
MIKE = A. Pierce
SAM = To BE Determined
LDE = Osi O.
DT = Tuck
DT = Robbins
RDE = Strahan/Kiwi

If Strah comes back and I think he will due to the issue with his ex wife, then Kiwi can rotate around like he did his rookie year. He can even be in the 4 aces, or play blitzing LB on passing downs like he did in the past.

The Giants probably have the 2 best starting ends, and the two strong young ends waiting in their wings. I think the comparative advantage Tuck has at DT is too great to not play him there on passing downs. His interior quickness is unmatched and it is amazing how quick he can penetrate right past a slower guard in such limited space.

I like Kiwi a lot as a football player and he is honestly big enough and fast enought to play LB, but he is faster than he is quick and more ideal for DE with the height/long arms/ and quick first step.

I want the G-Men to bring in a solid vet/1st or 2nd round stud OLB to play Sam. Maybe Dan Conners or Keith Rivers.

73
by Gozer Gozarian (not verified) :: Wed, 03/05/2008 - 11:07am

66

No, the Eagles were not better than the Giants last year. In the second game, the last field goal attempt hit the right upright, but it was barely within Akers range. Furthermore, that FG attempt was only set by a Brandon Jacobs fumble that happened on the tail end of a 30 yard run to the Eagles five yard line. On that play, the Eagles benefitted from some good ole' FO fumble luck.

We all know what happened in the first game to the tune of 12 sacks. But if you want to beleive that the Eagles, with four scoring drives (one touchdown) in eight quarters against the Giants, were actually the better team, go ahead. Maybe believing in swagger is stupid, but so is blindly following DVOA and ignoring the fact that the Eagles couldn't keep their quarterback upright and produced only ONE decent drive in two games against the Giants.

Keep the reverse FOMBC coming guys.

74
by goathead (not verified) :: Wed, 03/05/2008 - 11:18am

I love the comments along the lines of 'if Samuel makes the pick' or 'if Smith doesn't dig a ball out of the turf and hand it to the pats'. If you assume only that Eli's bad breaks hadn't happened, he'd have better stats (and the G's likely win by more than 3). If you assume only that his good breaks didn't happen, he'd have worse stats, and they'd lose. This tells you what? Nothing. In any close game if you changed the outcome of a couple of plays (frequently involving some degree of luck), the score would likely be reversed.

What if both Smith and Samuel catch their balls? Most likely a final score of Giants 17 Pats 14.

75
by Crushinator (not verified) :: Wed, 03/05/2008 - 11:53am

68

A Pats Safety dropped a pick on that drive also. I've never seen a player get more credit for great decision making than Eli Manning, who made some awful decisions but just happened to not be punished by them.

If you're playing Russian Roulette and the other guy gets shot down the stretch, it doesn't mean you were somehow skilled or you made a good decision. You just didn't get the chamber with the bullet in it.

76
by Al (not verified) :: Wed, 03/05/2008 - 11:55am

#70 Yes, that resentment is palpable. How would you feel if you spent months working on your science project, only to have it self-destruct during show and tell?

77
by lobolafcadio (not verified) :: Wed, 03/05/2008 - 12:09pm

Not a lot of respect for the FB...
So I'll be the third to mention it : Tapeh is now with the Vikings where he reunites with Brad Childress and will replace Tony Richardson.

How is Tapeh ? Is he a fine receiver ? Could the vikings be the rare team where the FB is a better receiver and a lesser blocker than their TE (Mr Kleinsasser) ?

78
by RoyFlip (not verified) :: Wed, 03/05/2008 - 12:56pm

#62, 64: All I heard after the game was Samuel's "dropped" interception. I didn't remember a drop. Looking at the replay, the ball glances off the fingers of his right handa at 90mph. That's not a drop. Corey Webster in the Dallas game with a ball in his lap all alone...THAT was a drop. As evidenced by Samuel's play, the "overthrow" to Burress, and Romo's "overthrow" to Crayton at the end of that game, sometimes the bad throw is actually a bad play by the receiver who is not where the qb thought he should be.
Maybe I am being defensive here, but look at Roethlesburger's numbers in his super bowl and all he gets is "the youngest qb to ever win a Super Bowl." He won that Super Bowl? No, the Steelerswon it, almost in spite of him. You keep bashing Eli. Unstoppable? No, but give it a rest.

79
by John Doe (not verified) :: Wed, 03/05/2008 - 12:57pm

55: I think your forgetting about Tuck. He had 10 sacks this year and dominated in the SB. He's a DE, even though they lined him up at DT mostly during passing downs. If Strahan retires Tuck should replace him. Tuck is very similar to Strahan, he's plays the run as well as he plays the pass and his bull rush is a thing of beauty.

I see Tuck taking Strahans place and Kiwi staying on as an LB. He did get abused the first two games, but then again that was against the Packers and the Cowboys and he'd never played LB before. He got better as the season went on, he'd definitely make and excellent pass rushing OLB eventually. On obvious passing downs I see them moving Kiwi to DE and Tuck to DT for the 4 Aces package.

Kiwi ended the season with 46 Tackles, 4.5 sacks, and 4 passes defensed in only 9 full games (his first games at OLB ever). For comparison, Kawika Mitchell had 76 tackles, 3.5 sacks, and 4 passes defensed in 16 games. It's likely that Kiwi would have put up better numbers than a Mitchell (a 5 year veteran at OLB) if he had played the entire year. I'd say he deserves a shot at OLB again. Wilkinson impressed me the first Pats games with his ridiculous speed. He was stuck covering Moss TWICE in that game, an OLB verus Randy Moss and he stayed with him the entire way. He got called for PI once in the end zone, but how many OLBs would have been right on Randy Moss in the end zone to defend against a deep pass?

The S crop this year is pretty poor, I'd like to see us take the best LB available or a good LT (this class is deep at LT) if one is available for our first pick. I'd like to see a tall possession receiver at some point also since Toomer is getting old and we seem to be stocking up on short injury prone WRs.

80
by Chris (not verified) :: Wed, 03/05/2008 - 12:59pm

74. I love the " we should count the almost interception" against Eli talk.

So we hate the guy so much we should count the dropped interceptions as picks on his stats too ? Should we count the 5 or 6 dropped competions he had too, including a Kevin Boss ball in the endzone? How about leading the league in dropped balls in the reg season ?

Tom Brady threw more picks in the playoffs than Eli did, and Eli played in one extra game. Oh, but he sucks and is comperable to Trent "game manager" Dilfer and certainly isn't as good as Jason " smoke screen" Campbell.

81
by Chris (not verified) :: Wed, 03/05/2008 - 1:04pm

John- If you like Tuck at DE so much then why not have him at DE on obvious passing downs?

Let's just assume Kiwi and Tuck are equals at end. Tuck is much better at DT and that is a fact. His quickness in space is unmatched by anyone not named Harris.

DE is a more important position than SAM, and if they can have Kiwi as a stud at DE then they should keep him there. Tuck can be an equal DE or the 2nd best pass rushing DT in the league. They should keep it at that.

I was never big on Mitchell in the first place. Sure, Kiwi can be at least an average OLB, or he can be a stud end. Keep him there. David Ricardo's law of comparitvae advantage from the 1800's says so.

82
by nat (not verified) :: Wed, 03/05/2008 - 1:05pm

74:
FO: "Our analysis shows this deck of cards is 25% diamonds."
Doofus: "Diamonds rule! Here I'll draw one. HA! A Diamond! That proves the deck is mostly diamonds! Diamonds outnumber all other suits!"
FO: "Past draws have been about 25% diamonds."
Doofus: "Past, shmast! Diamonds rule!"
FO: "OK, let's look at the cards."
Doofus: "Uhhhh, but the deck is changed since then. This is a new deck. You only get to draw once."
FO: "Sigh. I give up. See ya next year."
Doofus: "Jeez. Those guys are so resentful."

83
by goathead (not verified) :: Wed, 03/05/2008 - 1:39pm

78: Great analogy, but if you're looking at the deck and someone draws 10, 11, 12, diamonds in a row you might decide the composition of the deck isn't as you assumed at the start. Or you might decide they are the luckiest card picker ever.

84
by dmb (not verified) :: Wed, 03/05/2008 - 2:12pm

45, 57: “The thing that jumps out to me is the fact that D-Lew would rather have an Eli Manning with a ring on his finger than a Jason Campbell…” I don’t know where you picked this up. The way I read the article, Mr. Lewin (D-Lew?) was simply asserting that Campbell played better within his system than Manning played in his. He then proceeded to state that because Campbell’s system is changing, Campbell may play more poorly next year.
“…who was benched late in the year in favor of career backup Todd Collins.” Personally, I’d take Steve Young, who was benched late in the year (’91) for career backup Steve Bono, over either Campbell or Elisha. But more to the point, considering the way Campbell, Collins, and Manning all played during the regular season, I think most people would’ve taken Collins. I noted in another thread that Collins’ DVOA and DPAR were near the top of the league (3rd?) during the time he played.
“Blindly comparing their DVOA is absurd. The Redskins had the most vanilla offense in the league full of WR screens, 2 yard fullback passes, and 5 yard tight end drags.” Yep, those 18 catches by fullbacks (17 by Sellars; 1 by Koslowski) were certainly a crucial component of their offense! They actually ran WR screens considerably less frequently than in previous years, probably because the guy they like to run those for (Moss) was hurt for a substantial part of the year.
In the Tampa and @Dal games, Campbell definitely came up short. But I’d say that it was pretty much Campbell, and...Campbell who kept them in the DAL and PHI games in the first place. As for the first Giants game…I have to admit, Campbell really choked by letting the Giants stop Betts on third and fourth down, both from the one. He didn’t block anybody! Actually, it's funny that you cite this game as an example of clock management, because it was actually PERFECT management: If Betts can score from the one on either of his two chances, the 'Skins go ahead with about 30 seconds left. The clock didn't run out on them; the Giants kneeled to end the game.
I have no idea why you enjoy ripping him so much, but I think Campbell is a decent young QB who’s going to start in a new system (again!) that may not play to his strengths.

85
by Kyle (not verified) :: Wed, 03/05/2008 - 3:12pm

What is most comical about the hostility towards, aversion to, and irrational hatred of giving the Giants [b]their due[/b] is that it is from nearly everyone the website. Not limited to one writer, this inability to admit the possibility of subjective editorializing slipping into their works is widespread here.

Fervent defensive posturing when people dare criticize an interesting, useful system of numbers that helps explain part of the picture but fails to explain everything -- which is no slight whatsoever against DVOA, by the way -- or job requirement when applying? You decide.

86
by Ari (not verified) :: Wed, 03/05/2008 - 3:26pm

45

Campbell was not benched. He was injured. Why rush him back from injury when the backup is playing well? What do you think the Redskins are, the Bears?

87
by Chris (not verified) :: Wed, 03/05/2008 - 3:35pm

85. Campbell reiterated that he was healthy enough to play but he didn't play. What do you want to call it? It wasn't that he was " coming back", or " 85%" or any of that. He said he was healthy enough to play but his coach chose to go with the other guy who played better ( and many say Al Saunders liked Collins better anyway) where as Gibbs more favored Campbell and a more ball control offense.

The Redskins more or less opened it up some when Collins played. Camobell was more reserved to the game manager hat.

83. I never said Campbell didn't play well in the Dallas game or even Philly. The Dallas game was probably his best game, except he choked at the end. I was highlighting his end of game non-heroics as a flaw in his game. It didn't just happen once or twice, it happened multiple games in a row and on multiple drives. The mismanagement and late game interceptions happened over and over and over again. Hence why I called it a flaw.

88
by towishimp (not verified) :: Wed, 03/05/2008 - 4:15pm

I love FO for the most part, and really respect them for all the work they've done. But even for an uninterested (when it comes to Giants vs. Pats) party like myself, the anti-Giants/Eli thing is noticable. Okay, the FO states say Eli was bad. Fine, say that. Don't make fun. Continuing to publish these things just makes the site as a whole look like it has a serious case of sour grapes.

89
by Matt (not verified) :: Wed, 03/05/2008 - 4:53pm

I can't figure out why there is this desire to suddenly punish Eli for the mistakes he got away with (and again, nobody is catching the Samuel ball, nor is Meriweather trying to grab one over his shoulder routine either). Frig, if Madison catches that one from Brady we're talking about how badly he choked.

So he's had a crappy career. A collection of eleven human beings is also not quite a deck of cards - there IS a possibility of change, so the metaphor is a stretch. It might be less likely, but the odds are hardly zero. What's wrong with a wait and see approach other than "The general media wants to annoint the kid and FO is probably a touch overzealous in advising against this - we must choose a side now"?

90
by Kurt (not verified) :: Wed, 03/05/2008 - 4:59pm

Any attempt to predict game outcomes with objective statistical analysis is completely and utterly futile.

I'd say there's quite a bit of truth to this, at least until someone can demonstrate that said objective analysis predicts games better than Vegas, Peter King or Bill Simmons' wife do.

“I guess my point is, if your stats grade out the eventual Super Bowl Champs as the worst team in the division, maybe it’s the stats that are flawed rather than the actual games on the field.

Or maybe they genuinely played better in the postseason than the regular season?”

Now that’s just crazy talk. I mean, sure they won playoff games against three teams that they had gone 0-4 against in the regular season. Sure, they outscored those teams by an average of 3 points in the postseason, while being outscored by an average of over 11 points by those teams in the regular season. But they didn’t really play better in the playoffs. The scoreboard just better reflected their dominance when the postseason arrived. Yeah, totally.

Of course the Giants played better in the postseason, which illustrates (1) that "objective statistical analysis" has limited utility in predicting games, and (2) the Giants were probably a pretty good team to begin with, faulty analysis notwithstanding.

#62, 64: All I heard after the game was Samuel’s “dropped” interception. I didn’t remember a drop. Looking at the replay, the ball glances off the fingers of his right handa at 90mph. That’s not a drop.

Thank you. For the last month I thought I was going crazy - I saw a guy jump, reach up, and have the ball glance off the top of his fingertips, and the rest of the world treats it like Jackie Smith part II.

If Betts can score from the one on either of his two chances, the ‘Skins go ahead with about 30 seconds left.

They're going for two, then? That would have been pretty surprising.

91
by Dean (not verified) :: Wed, 03/05/2008 - 5:06pm

RE#77

Tapeh is just a guy. He started and was an acceptable blocker in Philly, but wasn't given the ball much and wasn't a threat to catch out of the backfield. Now, how much of that was his ability and how much was game plan? I can't say. He was drafted for a bigger role then he ever took on.

At the end of the day, how often does Philly let a young player walk?

He's not a bad player or anything - he's just a guy. I don't think there are enough metrics to quantify a replacement level fullback, but that sounds like all he's shown so far in his career.

92
by Footballoutsiderjoe (not verified) :: Wed, 03/05/2008 - 5:31pm

Eli stinks. Worse quarterback ever. 30th best dvoa. Doesn't deserve to play. Makes mistakes. He didn't win the super bowl and isn't as clutch as GOD Brady.

93
by Al (not verified) :: Wed, 03/05/2008 - 5:36pm

Samuels drop was really more an acting job on his part. He was to close to the sideline and in poor position to make the pick. He was just pretending like he had a chance to end the game right there.

A more interesting "drop" would have been Gay's, much earlier in the game. It was competely catchable but Gay missed it because his arm was maimed during an early tackle on Jacobs. Eli got away with that one. It was an example of the Giant's physical play taking a toll on a less physical team.

94
by RoyFlip (not verified) :: Wed, 03/05/2008 - 6:17pm

#92 Jeez, and I thought I watched the game closely... You're right, it doesn't look like he would've come down with it inbounds. Another one was the pass to Burress late that he jumped up to get a finger on and then Rodney Harrision carries on like he had just dropped a pick 6.
All I am saying is that everyone acts as if Eli got away with murder. You don't get 170 yards in the 4th quarter against a very good NE defense (who knows you are passing) by just throwing it up there and trusting your luck.
We could go on with the what-ifs forever I suppose. Is it September yet?

95
by Alex (not verified) :: Wed, 03/05/2008 - 6:19pm

Re 73:
Fair enough. I will say this, the Giants did outplay the Eagles in both games last year, both in terms of the final score and by DVOA (IIRC). But it's the rest of the regular season where, according to DVOA at least, the Eagles closed the gap in performance and outplayed the Giants (slightly).

Also, I do think that the Giants thoroughly and completely earned their Super Bowl win, in every sense of the term. They beat three legit title contenders in a row, two on the road, and held the highest scoring offense in NFL history to 14 points. They had a great postseason. And their DVOA in the postseason was also very good. They are certainly deserving champions.

Still, people saying that DVOA is somehow flawed because it says they only performed at an average level during the regular season are getting a little irrational. I mean, DVOA is based off of the NFL play-by-play, and nowhere in that play-by-play is there a section that says:

"A. Samuel intercepted E. Manning for 12 yards. (But don't worry, that won't happen in the Super Bowl, because Eli is going to suddenly stop throwing killer picks and have a hot streak at just the right time, and the Giants defense is going to hold all four of their playoff opponents to 20 or fewer points, despite allowing the last three of those teams to score 30+ points in each of the 4 regular season games against them.)"

It just doesn't. DVOA is saying that the Giants played like an average team in the regular season because they really did play like an average team in the regular season, and then played like an elite team in the postseason. Just like last year, when the Colts defense really did suck in the regular season, and then played better in the playoffs (and despite all the insults and unflattering comparisons made about him, Jason David does have a Super Bowl ring. Take that, Hole in Zone.).

96
by RoyFlip (not verified) :: Wed, 03/05/2008 - 6:19pm

Sorry, meant to reference #93...although #92 seems to capture it succinctly enough.

97
by Gozer Gozarian (not verified) :: Wed, 03/05/2008 - 6:20pm

Regardless, Who has even said that Eli has emerged as an elite quarterback anyway? It's just the usual "turned the corner" talk from what I've been reading. I don't think it's a ridiculous assumption to believe that he has turned the corner, either, even though Eli will probably have another 4 int/40 incompletion game next season and people will start screaming again. Like someone said above, however, I can handle the 4 int games and bad stats because Eli doesn't operate in the "game manager" style and actually tries to make plays.

98
by Drew (not verified) :: Wed, 03/05/2008 - 8:34pm

This thread was mind-numbing. We get it. Some FO writers are Pats fans, and they are often harsh on their arch-rival Giants. Get over it.

Then in the midst of the hundred-odd "don't ever say anything bad about Eli" posts (as if Giants fans had never bad-mouthed the guy), there was the gem of post #34 with the bible quote of how kids taunted Elsha and called him "baldhead." In response, two bears came out of the woods and mauled 42 of them.

At first, I thought it was appropriate. But then, I realized that Eli isn't bald. (I don't think they sell topuees with the Manning crop-cut). And the Bears were incapable of mauling any children last year, much less 42 of them.

99
by dmb (not verified) :: Wed, 03/05/2008 - 8:45pm

87: I'm not sure that "late game non-heroics" constitutes "a flaw in his game." Does that mean any quarterback who plays for a team that loses a game suffers from a "late game non-heroics" problem? I agree that the picks @DAL and @TB were bad plays, but I guess I don't understand how two plays can be labeled a systemic flaw. (He threw 0 INTs in the Eagles game you referred to.) I feel like I refuted your only example of "mismanagement" pretty effectively...and let's not forget that the 'Skins won two OT games with Campbell at the helm. If he's going to get blame for every close loss, then he should get the credit for close wins, too -- even though in both types of cases, he had a limited effect on the games' outcomes. I guess it could simply be that you think QB play has a greater effect on a team's play than I do (it would explain why you've expressed a high opinion of Eli, while my position was summed up pretty well by David Lewin when he basically said "you can win with him, but he won't win them himself."). Or maybe we're just both homers? : )

90: "They’re going for two, then? That would have been pretty surprising." You're right, that was a dumb mistake on my part. But the point was that Campbell brought them to the Giants' 1, with less than a minute left; and the 'Skins were able to use all their downs before the clock expired (including two failed runs). I don't think you can pin the failure of that drive on Campbell.

100
by Pat (not verified) :: Wed, 03/05/2008 - 9:23pm

78: Great analogy, but if you’re looking at the deck and someone draws 10, 11, 12, diamonds in a row

The 2007 season wasn't about drawing 10, 11, or 12 diamonds in a row. It was about drawing four.

Football is all small sample size. The worse team frequently wins, which is why you actually watch the game.

If the best team in football plays the worst team in football, how often does the worse team win? Probably about 10% of the time or so. If the best team in the playoffs plays the worst team in the playoffs, how often does the worst team win? Probably 30-40% of the time.

And #2 was prophetic. So far, this thread's a complete embarassment to logical thinking.

101
by Matt (not verified) :: Wed, 03/05/2008 - 10:41pm

How frequently? I thought what happened had less than a 1% chance here, and now it's normal all of a sudden? Who crowned Eli an elite quarterback? Who even had time during the Favre lovefest? All I want to prove is that he's not Rex Grossman.

It's my understanding the complaints come from:

1. "Full Kotite."

2. Aaron switching his stance - rather inexplicably - from "The Giants scare me more than the Steelers based on the 4-man rushes that have knocked Brady around" to "this will be a blowout" after the Giants played BETTER and rang up a metric ton of hurries in Week 17 - which normally, by this site's logic, will become sacks next time if the pass rush does that again. The assumption became that Eli would normalize, as if he was everything. How positively ironic.

3. The entire quick reads comment after the Packers game was dedicated to reminding us what happened to Tony Eason. Jake Delhomme put together a heck of a Super Bowl, after all, but immediately in the headline was "don't get your hopes up."

4. No matter how good the methods, prognosticators are expected to eat some crow when wrong (unfair as that may be given the difficulty of predicting so many games). The response has been more about defending the system, which has been attacked many times before - just not from one team's fans like this. Someone use the default FO complaint form already!

5. FOX' unfortunate headline about the fumble luck. We'll be here forever debating who got luckier (the Giants blew a handoff and got a ridiculous catch - does that even out? Does one dropped pick equal a Steve Smith Special? Aaaaggh!) for no good reason and with no conclusion.

102
by John Doe (not verified) :: Wed, 03/05/2008 - 10:55pm

Because I don't like Kiwi at DT and the four aces package has 4 pass rushers on the line. So on obvious passing downs it makes sense to move Tuck to DT so that we can utilize Kiwi and slide a DB into Kiwis spot at LB. Tuck just isn't big enough to be an every down DT, and he's far too talented to be a backup/situational guy. I agree that Kiwi is a better DE than LB, but I don't think he's a better DE than Tuck, at least at this point. One of the things that makes Strahan so great is the fact that he can play the run so well. Tuck plays the run at DE a lot better than Kiwi, if either of them is going to be a situational guy on the line I think it should be Kiwi.

The problem is they're both very good players and this is the free agency era. We won't be able to afford to keep 3 top shelf DEs, if Kiwi moves back to DE we will probably lose Tuck.

Regardless, Tuck has earned the DE spot with his play at DE this year. I don't see the Giants insulting Tuck by benching him in favor of moving Kiwi back to DE. At the very least he'd have to beat out Tuck in camp, and I don't think he's a better DE than Tuck at this point. I think we get a lot of value out of using Kiwi as a hybrid LB/DE and Tuck as a hybrid DE/DT.

103
by Chris (not verified) :: Wed, 03/05/2008 - 11:01pm

99. Did you even watch the games ? In the third game of the year Campbell sure as heck was mismanaging the game at the end ! They could have had plenty of time and they were rushing at the end because of poor management. Just like the 2 picks at the end of the Tampa game, the interceptions at the end of the Dallas game, the mismanagement in the Philly game.

Jason Campbell won 2 games in overtime? You mean beating the 1-15 dolphins @ home in week 1 in overtime?

Campbell DID have more of a limited impact on his games because he was a game manager. He didn't try and win games like Romo/Brady/Manning and succeed, and he didn't try and win games like Grossman and fail.

Nobody did annoint Eli a top 5 quarterback. I was probably the most pro Eli guy out there and I never made that comment. But when people bashed the guy and MADE FUN OF HIM then it makes it look even more irrationale.

Eli did have a terrible day in Minnesota, but then again Peyton had a terrible day in SD, and Romo had a terrible day in Buffalo. That doesn't mean they need to be bashed repeatedly for it. They "win" games.

104
by JasonK (not verified) :: Wed, 03/05/2008 - 11:05pm

#72:

I disagree. The problem about moving Kiwanuka back to DE is that he is not a particularly good run defender when the OT is right on him. (Particularly not with the weight he lost to move to LB.) If you put him in Strahan's spot over the right offensive tackle (usually the offensive strong-side), opponents will run right at him with success. If he's a DE, he's a weakside speed-rushing DE. But the Giants already have Umenyiora filling that role in the base defense.

If Stray retires, the starters will be:
RDE: Tuck
DT: Robbins
DT: Cofield (Although Alford could win this in camp)
LDE: Umenyiora

On passing downs, Tuck moves over to DT, Kiwanuka steps up into the RDE spot, and one of the DTs (Cofield, most likely) goes to the bench. (The "Three Aces and a Jack" formation) The Giants were doing exactly this in the preseason last year before Strahan decided to show up.

105
by Chris (not verified) :: Wed, 03/05/2008 - 11:11pm

102. The thing is that I think Tuck is probably the 2nd best pass rushing DT in the league. If you have that you use it.

Kiwi's length/height and quick first step make him more ideal to be a pass rushing DE. With Tuck at DT and Kiwi at RDE you get the best of both worlds.

With Tuck at DE, and Kiwi at Sam you are losing value. I do think Kiwi is a better end than Tuck but it shouldn't even matter. You lose value by playing Tuck at end because he isn't the 2nd best pass rusher at end and you lose the value of him at DT, and you lose value on Kiwi out of position.

Justin Tuck got signed to the long term extension in the playoffs. The Giants locked him off even before the SB when his stock went up.

You are right, Strahan is good vs the run and pass, which makes him real good. At this point Osi is probably a better pass rusher than Stra but I saw Flozel Adams blow him up on some run plays vs dallas last year. Osi isn't bad against the run, but a 350 pound Flozel is a very tough matchup for him.

I think the Giants will continue to rotate guys around like they did Kiwi's rookie year. Osi and Strah are not only the starting DE's but they are team leaders too. Kiwi and Tuck are still younger players.

People were shocked when the Giants drafted Kiwi out of BC, but Accoursi valued pass rush and he spoke out about " pass rush in the 4th Q".

Value is attached at the margin and if the other team is running the 2 min drill at the end of games when everybody is tired, you will lose close games ( Jason Campbell excluded).

Remember the SB didn't end with the TD to plax. The Giants defense had to take the field and rookie Jay Alford who was rotating in lit up the golden boy for a back breaking sack.

106
by Chris (not verified) :: Wed, 03/05/2008 - 11:25pm

104 - I see your point but
Osi is 6'3 261.
Kiwi is 6'5 265

Kiwi is also younger and can EASILY add weight to that taller frame and the Giants NFL strength program. Going to 275 seems very reasonable, and 280-285 seems within range as well if he really applied himself.

Now you don't want him to bulk up too much because then he could potentially lose speed, but 10 pounds added to a 265 pound frame really isn't much when he can add 5 to his lower body and 5 to his upper body.

Also consider Strahan
6'5 255 which is lighter than both of them and he is actually the best of the 3 vs the run.

LDE - Osi = Pro bowl caliber pass rusher
DT - Tuck = Pro bowl caliber pass rusher
DRE - Stra/Kiwi = PB caliber pass rush

If you go with ...
Osi
Tuck then you would only have 2 PB caliber, with Kiwi as maybe an above average SAM

If you go with 3 aces you have 3 studs instead of 2 on the line... and if Strahan comes back then Donovan better take out another insurance policy.

107
by Maryland Giant (not verified) :: Wed, 03/05/2008 - 11:36pm

Yeah, it's pretty clear that these FO guys are now among the most angst-ridden, victimized, self-loathers on the planet.

To think little old Eli turned this empire into rubble all by his l'il ol' "aw shucks" self makes me shake my head in awe.

LOL...man, these guys struggling to retro-fit the whole deal makes it even sweeter.

108
by Pat (not verified) :: Wed, 03/05/2008 - 11:53pm

I thought what happened had less than a 1% chance here

What?!

No team has a 99% chance of beating another team. Hell, no team has a 90% chance of beating another team.

70%, at best, is more like it. You can pretty much show that from the distribution of wins in the NFL.

109
by Rick (not verified) :: Wed, 03/05/2008 - 11:57pm

Giants fans pissing and moaning about Eli:
1. I am an Eagles fan who lives in NYC. Never have I seen a bigger turnaround in people's opinions about a player than Giants fans and Eli over the period of the playoffs. Like it or not, your love affair with him is relatively recent.
2. He's still not THAT good. Aside from a good 2 minute drill to win the Super Bowl, he still was very mediocre throughout the game. And despite what you all think, it's the defense that won all those games. The defense allowed Eli to "win" by preventing other teams from scoring too much. Eli is incapable of coming back from large deficits, but he didn't have to.
3. Even if he improves, the defense will have a hard time maintaining the level of play the reached during the playoffs. Particularly if Strahan retires. This makes it a zero sum game.

The Super Bowl was a magnificent achievement by this Giants team, I'm not writing any of this to "take away" their accomplishment. Nobody can, or would...it's in the books, and it's still one of the greatest upsets of all time. But remember - it was an UPSET - and EVERYONE called it an upset. Even the Giants fans who believed it could happen. As such, in recognizing the nature of the upset, there is a tacit recognition of the level of play the team put forth throughout the season - patently mediocre. In fact, barely playoff worthy. With one loss - to the Eagles, for example - the Giants could easily have been out of the playoffs. And they did almost lose the second game....and nobody would call the Eagles 2007 campaign remarkable.

So piss and moan all you want about how your team is being "dissed", but sit back and remember all the sturm und drang that you put yourselves through during the season as the team underperformed expectations. Then remember that even as the team was winning in the playoffs, each of the last 3 games came down to the wire. I watched all of them with my Giants friends, and not one was convinced their team would wind up winning as the clock ticked down.

The nice thing about winning is, after the fact, you get to heap praise on your team regardless of the facts, and often in spite of them.

110
by Rick (not verified) :: Thu, 03/06/2008 - 12:14am

101:
I think you've got it all wrong. Nobody, but nobody, was CONVINCED the Giants would win. Even people who said they could realized it was a major upset in the making if it did occur.

As a result, when working from statistics, you have to make your commentary and assumptions utilizing those statistics. But all statisticians recognize the reason they play the game is to generate an outcome - the numbers can't win a game, the players have to. That said, the numbers can tell you ALOT about what will happen most of the time. I've found that by using FO's stats, my betting choices are vastly improved, as are my fantasy teams. So, obviously, stats work. But there are times when they don't. You can't retrofit numbers every time they are wrong, that's poor use of statistics, and I don't think Aaron has said he's doing that. What he has done is gone back to seek out the small instances of information that may have been misread or misinterpreted. From what I've seen, there wasn't much. The Super Bowl was still a huge upset.

I had friends who wanted to bet the Giants and take the points. My response was if the game was THAT close, the Giants would win and it was better to bet the Giants outright and take the odds. Many of my friends understood the logic and made significant sums that way. I came to that conclusion because of FO, not in spite of. Did I bet on the Giants? No, I avoided all their games. While I thought they were mediocre, I also recognized that the most dangerous team is the one that doesn't know how bad it is. I avoid those teams when it comes to betting.

Eli DID normalize in the Super Bowl. His numbers were better than regular season, but not nearly as good as the previous 3 games. It was the defense that deserved the MVP award, the front 4 specifically. Eli got it because of....well, I'm not sure. Probably because he's Peyton's brother and the NFL thought that was kinda cool. Eli didn't break free in that sack, Seubert didn't give up, and his second effort allowed Eli to get out. And that catch by Tyree was a great catch - not a great throw by any stretch.

FO has not dissed the Giants by any stretch. The writers have simply written the facts as they are. But the one fact that simply doesn't jibe with all of this is the fact that the Giants won the Super Bowl. It's known as an outlier - a fact that happened in spite of what the statistics suggested should have happened.

And what's great about that is that isn't it better to have a win like that? I mean, think about the New England Patriots - while it would've been alot of crowing about how great 19-0 is and how the Pats are amazing, isn't it better to be able to shut those fans up?

But in shutting them up, it's wrong to deny the entire season that preceded that moment had any meaning.

111
by Tom D (not verified) :: Thu, 03/06/2008 - 12:34am

Re 10:

"Eli is incapable of coming back from large deficits"

I'm surprised an Eagles fan would say such a thing.

112
by Kurt (not verified) :: Thu, 03/06/2008 - 12:52am

The 2007 season wasn’t about drawing 10, 11, or 12 diamonds in a row. It was about drawing four.

Four? Really? I don't recall that this discussion began in January. My recollection is that they were "drawing diamonds" for quite some time before the playoffs.

Anyway, the deck of cards analogy works great if you accept DVOA rankings as the Holy Tablets sent down from Mt. Sinai, and think that figuring out the "best" and "worst" teams is as simple as rifling through a deck of cards. It also helps if you hate considering new information once your conclusion has been formed. Me, I might take another look at that deck of cards eventually.

I thought what happened had less than a 1% chance here

What?!

No team has a 99% chance of beating another team. Hell, no team has a 90% chance of beating another team.

70%, at best, is more like it. You can pretty much show that from the distribution of wins in the NFL.

I'm pretty sure he's talking about the whole ball of wax - beating Tampa AND Dallas AND Green Bay AND New England.

113
by Pat (not verified) :: Thu, 03/06/2008 - 1:12am

Four? Really? I don’t recall that this discussion began in January. My recollection is that they were “drawing diamonds” for quite some time before the playoffs.

If they would've lost in the first round versus Tampa, I doubt many Giants fans would be claiming that DVOA's view of them was that wrong. It was the four victories in the playoffs which did it.

I’m pretty sure he’s talking about the whole ball of wax - beating Tampa AND Dallas AND Green Bay AND New England.

Uh, okay, if you say so - but that's amazingly stupid, since any team has a small chance of rattling off multiple victories in a row. It's just that there are a lot of teams in the NFL, so when you throw them all together, someone has to win.

By the log5 method and points for/against, which doesn't use DVOA at all ("DVOA handed down from Mt. Sinai" ?! There are plenty of other ways of analyzing football teams' performance - just mildly less predictive - and all of them said basically the same things), the Giants had a 0.7% of making that run.

But 0.7% is about 1-in-140. We're at Super Bowl 42, for crying out loud. Rare events are going to happen. If I threw together a toy Monte Carlo just using pure flat random number distributions, I'd get a team with a 0.7% chance of winning the Super Bowl in the first 42 chances about 33% of the time.

I really don't understand people railing against this. You have to understand: it's a game. If it were perfectly predictive, no one would watch it. That means that bizarre things are going to happen. The Bills are going to beat the Patriots 31-0 in Week 1, and lose to the Patriots 31-0 in Week 17. It happens.

At the same time, however, it's utterly ridiculous to claim that you can't extract a team's true strength better than wins/losses. If you couldn't, no one would watch the game either - you'd have the National Coin Flip League.

Football's a mix of pure luck and pure skill, and it seems very hard for fans to accept that. When their teams win, it must be skill. When their teams lose, it must be luck.

114
by Matt (not verified) :: Thu, 03/06/2008 - 1:26am

107. I'm referring to the "Great Playoff Runs" article, which suggested that if you played the playoffs a thousand times, you'd get seven Giants championships - and that was done without factoring in homefield advantage.

109. I'm not personally attacking DVOA, the writers, or whatever. I rather trust the site - however, it's been noted that certain matchup factors also count. Peyton versus a blitzing 3-4 is usually worth an adjustment in favor of the latter - and as the Colts proved earlier in the season, a 4-3 that can get pressure while dropping a lot of guys into coverage are worth an adjustment away from Brady. This is why, as Aaron noted that a fired-up Giants team in Week 17 were the best chance remaining of ending the Patriots' run - more so than the Steelers.

Week 17 followed a completely flipped script, though. While the pass rush got in Brady's face again and again and knocked him around, the Patriots somehow put 38 on the board and the Giants stayed in it because of Manning, who was having a bad...season, career, whatever, to that point. Elisha's playoff run then became The Story, and the Outsiders rightly predicted he probably couldn't keep it up (although you can argue he did keep it up - but really, it appears he fell somewhere about halfway between his playoff level and the comical "unstoppability" we've all grown to mock). The line mismatch, however, was still in the Giants' favor, and so one could assume that they, like San Diego/Indianapolis/Philadelphia would be able to keep the scoring out of hand if the hurries turned into sacks - certainly a bit different from "biggest mismatch ever" (which from a purely statistical perspective, it of course was.)

Then the Giants won in the face of a somewhat normalized passing game, showing to some extent the fallacy that Eli was everything...and the response has been, from some of the dumber corners of the punditsphere, that he was. Glorious irony.

My post, however, was not about why anyone should have picked the Giants. It's an attempt at an explanation for where all of the attacks are coming from. I agree with FO's reasoning for the most part - I just think that when the mainstream becomes overzealous with a given storyline, the analysis end of the site tends to go just a touch too far the other way sometimes, which started to become a bit grating as the postseason wore on. Add in the same crowd who make the Manning/Brady Debate Threads so much...er, fun, and you had to know this was coming.

I don't know what the dorky little watch salesman's baseline is going to be for the future. I think it's crazy to anoint him a star just yet - but in football five games is a lot more than it is in some other sports, and I think it's enough to suggest that maybe he's not horrible.

115
by johnt (not verified) :: Thu, 03/06/2008 - 1:33am

100: "Football is all small sample size. The worse team frequently wins, which is why you actually watch the game."

If you believe this, then what is the point of this website? The real offense to logical thinking is the idea that somehow FO's statistics can fail to be predictive and yet still of importance. If all the fuss and furor is about whether the Patriots have a 57% or 62% chance of winning the next game, who cares? Are we supposed to be fighting over the bragging privileges of the best DVOA?

Statistics has the concept of the R squared - how much of the variance in outcomes is explained by the model. As time goes on I'm becoming disappointed in the ability of the DVOA model in that respect. The way the FO staff dismisses any criticism of the idea that DVOA approach may not actually be meaningful is expected given how much they have invested in it. It gets to the point where when DVOA and reality diverge, they seem to blame reality rather than acknowledging its increasingly obvious shortcomings - namely that despite all the fancy numbers and tables, it's not all that much better than a skilled observer's subjective impression of what teams are good.

116
by Matt (not verified) :: Thu, 03/06/2008 - 1:33am

Er, "keep the scoring from getting out of hand." Sorry.

117
by Pat (not verified) :: Thu, 03/06/2008 - 3:33am

If you believe this, then what is the point of this website?

Er? Just because a lesser team frequently wins doesn't defeat the purpose of evaluating the true strength of teams. It means it's a game.

This is true for every sport. Baseball, basketball, chess, for crying out loud. The sigmoidal game output function for chess goes to its asymptotes rather slowly for exactly that same reason.

In fact, if I gave you a list of every game where the lesser team beat the better team, I'd probably give you a list of the best football games of all time. That's why we watch.

The real offense to logical thinking is the idea that somehow FO’s statistics can fail to be predictive and yet still of importance.

When did "perfectly predictive" and "predictive" become synonymous? They are predictive. Look at the math, or do it yourself. It doesn't take a lot of effort to realize it.

In any case, predictivity is not the end-all and be-all of stats unless you're obsessed with gambling. The point is to figure out why teams win - to understand what the game is telling you about each team than "on this day on this field this team happened to score more points than the other one."

DVOA itself is a balance between a predictive stat and a descriptive stat. That's flat out stated in the FO FAQ. It helps tell you why a team won, and whether why they won is predictive of future wins.

If you're really curious, go and look up the mathematics of maximum-likelihood rankings of sports teams, or better yet, find a description of Laplace's method for ranking teams based only on binary information. All of them rely on the fact that outcomes are not perfect, but that the outcomes are consistent - that is, the most likely outcome happens most often.

118
by QB (not verified) :: Thu, 03/06/2008 - 6:29am

There's a reason the Giants were underdogs to the Eagles both home and away. Enjoy the variance but don't get too big-headed.

119
by Rick (not verified) :: Thu, 03/06/2008 - 12:01pm

re 111:
Interesting that, rather than address the issue I mentioned, which was stated correctly and backed up with facts, you chose to alter the discussion to who I root for, assuming that's a basis of bias.

My wife is a Giants fan. Living in the NY area, it's not like I have alot of choices for where I can go to get coverage that's unbiased. As a result, I've developed a rather unbiased view of Giants football. I support my wife when the teams aren't playing in games that impact each other, so I was rooting for the Giants to beat the Patriots (I really didn't want a team to go 19-0, either).
I've also gained alot insight into the psyche of the fanbase and the team because it is the primary coverage I get, so I've had to work with what I have.

I wouldn't compare McNabb with Eli. McNabb's statistics are vastly superior in every way but one - and that one is on Eli's finger. However, if you asked anyone over the first 4 years of their career which had better performances, nobody would choose Eli. Except a few Giants fans who bet with their heart.

I admit McNabb has had a few off years primarily due to injury. Assuming he is healthy this year, his last 3 games showed he is very capable and can return to form. In addition, the Washington game showed he still has the ability to lead a comeback from late game deficits that are significant.

So, as far as the team I root for, as opposed to the team I read most about, I am relatively realistic. I want my team to win, but I will base my judgements on reality. Reality, this year, was that the Eagles were only slightly "worse" than the Giants during the regular season. And that "worse" portion showed up in a 2 game differential that could just as easily been 0 games, but for a blown call or a missed FG. These teams, in the regular season, were very similar in performance - inconsistent.

The difference between them is what allowed the Giants to make it into the playoffs and allowed them to make an outstanding run. The Eagles had no such chance, and it's their own fault. They had several opportunities (against Green Bay, the Seahawks, New England, Chicago and the Giants) that they let slip away. Interestingly, McNabb was not the reason they blew any of those games. I'm not sure you could say the same about Eli in every one of the Giants' losses.

Similarly, Eli did not WIN any of the playoff games, except perhaps Tampa Bay. His defense allowed him to perform better than normal, and that was sufficient to win. In fact, Eli's "better than normal" wasn't all that much better. The Giants averaged just under 24 points per game during the regular season. They averaged just over 21 in the post season. So the real thing you can say about Eli is that he didn't LOSE games, but his defense did a tremendous job in helping to win them. During the regular season, the Giants' Defense gave up about 21 points per game. In the post season, they gave up just over 16. That's a 5 point improvement versus the offense's (Eli's?) 2 point DECREASE.

Anyone who thinks Eli was the star of the post season is simply ignoring the facts. He did a good job, pure and simple. He has possibly turned the corner in terms of LOSING games. He has not turned the corner in terms of WINNING them.

120
by Kurt (not verified) :: Thu, 03/06/2008 - 12:14pm

Four? Really? I don’t recall that this discussion began in January. My recollection is that they were “drawing diamonds” for quite some time before the playoffs.

If they would’ve lost in the first round versus Tampa, I doubt many Giants fans would be claiming that DVOA’s view of them was that wrong. It was the four victories in the playoffs which did it.

Giants fans were claiming DVOA was wrong in October. And yes, I'm sure a loss to Tampa, as improbable as that seems, would have quieted many of them down, a least those for whom game results matter. And DVOA fans would happily have pointed to the game as validation.

By the log5 method and points for/against, which doesn’t use DVOA at all (”DVOA handed down from Mt. Sinai” ?! There are plenty of other ways of analyzing football teams’ performance - just mildly less predictive - and all of them said basically the same things), the Giants had a 0.7% of making that run.

But 0.7% is about 1-in-140. We’re at Super Bowl 42, for crying out loud. Rare events are going to happen. If I threw together a toy Monte Carlo just using pure flat random number distributions, I’d get a team with a 0.7% chance of winning the Super Bowl in the first 42 chances about 33% of the time.

But my point is that nobody can *know* what percentage a team has of winning a specific game. Team "ability" (for lack of a better word) is much too fluid to pin down like that.

Plenty of people thought the week 17 game was an indicator that the Giants had "put it together" or whatever and that the weeks 1 and 2 dbacles weren't really true indicators of the team's ability by the end of the season. I wasn't one of them; I didn't think they had more than a puncher's chance in their last three games, but I was wrong. I find it's liberating to be able to admit when I'm wrong.

Also, nobody has ever suggested that the Giants, over the course of the season, were a better team than Dallas, Green Bay or New England. My personal opinion is that the Giants were closer to those teams than most people thought, which made the upsets more likely than others thought they'd be.

I really don’t understand people railing against this. You have to understand: it’s a game. If it were perfectly predictive, no one would watch it. That means that bizarre things are going to happen. The Bills are going to beat the Patriots 31-0 in Week 1, and lose to the Patriots 31-0 in Week 17. It happens.

At the same time, however, it’s utterly ridiculous to claim that you can’t extract a team’s true strength better than wins/losses. If you couldn’t, no one would watch the game either - you’d have the National Coin Flip League.

Funny - for some people, it already is the National Coin Flip League. The Giants didn't actually accomplish anything, they just happened to roll boxcars a few times in a row. And what is "true strenght", anyway? For me, the question of whether the Giants were the "worst" team in the division is simple. At the things that DVOA measures, they were the worst. At winning games, they weren't.

Football’s a mix of pure luck and pure skill, and it seems very hard for fans to accept that. When their teams win, it must be skill. When their teams lose, it must be luck.

Yes, I've noticed that too. When the Giants lose to the Vikings or Redskins turn out the lights, the party's over. When they win, well, statistical analysis says it had a 0.7% chance of happening, so that must have been it. There were people claiming the Giants were lucky to win the Winston Justice game! On the other hand, I've never seen a Giant fan here dismiss any of their losses as "bad luck" (admittedly it would be hard to do, considering the losses). Hopefully everyone else can catch up next year.

I promise you this, from the bottom of my heart - if the Eagles sweep the Giants next year, end up with a better record and win the Super Bowl, you won't hear a peep out of me.

121
by goathead (not verified) :: Thu, 03/06/2008 - 12:48pm

Kurt -

Well put. I'm biased as a Giants fan, but pointed out numerous times why I felt that the Giants DVOA wasn't an accurate indicator of what the team was capable of. And I did this before the SB.

Now, in analyzing the SB, to say the giants simply rolled the dice and pulled off an improbably roll is well beneath anyone who is paying attention to the stats, the players, and the game. The interesting thing about the SB isn't just that the Giants won, its that they frankly did it by going on the field and outplaying a team that was expected to destroy them.

And yes, the FO staff in writing the 'worst SB team ever', 'biggest mismatch ever', etc was essentially going out of its way to antagonize Giants fans. And when they created a new stat after the SB to show how the pats had actually outplayed the Giants, despite what everyone saw on the field, the FO staff lost a ton of credibility.

In past years I've always felt that they did a good job of being objective regarding the pats, bet the thrill of the quest for perfection clearly got the better of them.

122
by BDC (not verified) :: Thu, 03/06/2008 - 1:20pm

117:

"They are predictive. Look at the math, or do it yourself. It doesn’t take a lot of effort to realize it."

They aren't predictive though, that is the thing. I have done the math, and they get about 68% of the games right, which is pretty much exactly what everyone scores. Matching expected results is not proof of predictive powers.

123
by Gozer Gozarian (not verified) :: Thu, 03/06/2008 - 1:50pm

The Giants that won in the play-offs and the SB weren't even the real giants. The whole team was returning video tapes that month.

124
by nat (not verified) :: Thu, 03/06/2008 - 2:49pm

122: About DVOA's 68% rate in picking winners, and how that is similar to what everyone scores.

I'll trust you on that number for now.

But consider this: The oddsmakers and their gambling clients can consider every nuance, every element of moxie, distraction by supermodels, that look-in-the-eye, QB footwork mechanics, injuries, what have you. And yet, using only the play-by-play and statistics, DVOA is just as good at predicting winners.

In my book, 68% at picking the winner PLUS objective analysis is more interesting than 68% plus ad hoc analysis. For all I know, the bookies are using something like DVOA behind the scenes. But since I don't get to see their analysis, I prefer DVOA.

125
by Tom D (not verified) :: Thu, 03/06/2008 - 2:50pm

Re 119:

I wasn't trying to compare Eli and McNabb (sorry for the confusion). Just the fact that Eli has lit up the Eagles after being down big, I think twice in the past two years.

BTW, I'm a Bears fan and have no rooting interest either way, and I enjoy a good Eli joke as much as the next guy.

126
by Joe (not verified) :: Thu, 03/06/2008 - 2:51pm

re: 119

I'm at a loss to understand the reasoning that the defense was solely responsible for the Giants playoff run. The Giants defense definitely played well, but so did all the defenses the Giants faced. Three of those defenses were in the top 10 in DVOA, and even Green Bay was in the top half of the league. The playoff defenses on average were much better than the ones the Giants took on during the regular season.

Moreover, Eli played against and outscored Jeff Garcia, Tony Romo, Brett Favre, and Tom Brady. By your logic these guys are game managers that LOST the games for their teams, since their defenses (much like the Giants) also allowed them to win. So if Eli is outplaying Pro Bowlers and future first ballot HOFers and that isn't all that much "better than normal", I guess you must have a pretty high opinion of him.

127
by John (not verified) :: Thu, 03/06/2008 - 3:12pm

119, I think you missed the point? Not McNabb v. Eli. He was talking specifically about the Stink and the Linc.....remember that one? Actually, Eli has brought his team back. The Bears game comes to mind last year. Granted he played horribly in that game up until the 4q but he scored, I believe 14 in the 4th qtr to win.

128
by Travis (not verified) :: Thu, 03/06/2008 - 3:44pm

No team has a 99% chance of beating another team. Hell, no team has a 90% chance of beating another team.

70%, at best, is more like it. You can pretty much show that from the distribution of wins in the NFL.

The percentage has to be higher, if point spreads are any indication. Since 1994, teams favored by 7 or more in regular season games have won about 78% of those games. (In 2007, it was even higher, 79 of 92, or 85.9%).

Teams favored by 10+ since 1994 have won 83% of those games (in 2007, 40 of 44); teams favored by 14+ 93% (in 2007, 13 of 13).

If the best team in the playoffs plays the worst team in the playoffs, how often does the worst team win? Probably 30-40% of the time.

Playoff data is similar to regular season data; 7+ point favorites have won 74% of their games since 1994.

129
by Gozer Gozarian (not verified) :: Thu, 03/06/2008 - 3:45pm

I wonder how far the revisionism will go with Eli's playoff performances. The Giants scored less points because their running game declined and they faced stronger defenses. I'll admit that he was average until the two minute warning of the SB, but if you look at the numbers for the other games, they are about as good as you can play (remember that in the game against GB, there were a ton of drops). How the hell was he supposed to play better exactly, audible to more pass plays to produce better raw stats?

I can see it two months from now: "BS! Eli did NOTHING in the play-offs!! They only won because Michael Strahan scored all of the touchdowns and he's RETIRED!" 6-10 in 2008!11!

130
by Alex (not verified) :: Thu, 03/06/2008 - 5:29pm

They aren’t predictive though, that is the thing. I have done the math, and they get about 68% of the games right

If they get significantly more than 50% of the games right, they are predictive. They don't have to be more predictive than Vegas in order to be predictive.

Remember, the predictive ability of DVOA is only meant as a way of confirming that DVOA is valuing good teams over bad ones, so that the stats can then be split into Offensive, Defensive, and ST DVOA, and give information about what specific things caused the team to win. Making things even more predictive isn't really necessary or fruitful, because DVOA isn't really meant to be used to predict game winners. As long as DVOA isn't rewarding things that make a team less likely to win, and is valuing things that make a team more likely to win, it's doing its job.

131
by Alex (not verified) :: Thu, 03/06/2008 - 6:02pm

78: Great analogy, but if you’re looking at the deck and someone draws 10, 11, 12, diamonds in a row

I don't see where you're getting 10 diamonds in a row from. The Giants didn't exceed their estimated wins by 10 wins. They exceeded it by 2.1 wins in the regular season. And then they won 4 games in the playoffs as DVOA "underdogs". So, they won roughly 5 or 6 more games than expected over the course of 20 games. Yeah, that's unusual, but it's hardly a cosmic coincidence.

"'Four? Really? I don’t recall that this discussion began in January. My recollection is that they were “drawing diamonds” for quite some time before the playoffs.'

If they would’ve lost in the first round versus Tampa, I doubt many Giants fans would be claiming that DVOA’s view of them was that wrong. It was the four victories in the playoffs which did it."

Giants fans were claiming DVOA was wrong in October.

Yeah, but that might just be because they are Giants fans, and don't like it when DVOA says their team isn't as good as its record (which is, honestly, perfectly normal behavior for an NFL fan). Lots of teams have fans who dispute their team's DVOA when it doesn't agree perfectly with their W-L record, and most of the time, DVOA turns out to be the better predictor of future results than "fan-adjusted DVOA" or current W-L record.

Not that DVOA never misses something that fans pick up on, but just saying that DVOA is wrong doesn't make that the case. And just because the team goes on to win the Super Bowl, doesn't mean the fans were right all along that DVOA was underestimating their team's chances (after all, if a bunch of teams have fans claiming DVOA was underestimating them, then the fact that one of them goes on to exceed their FO-predicted end of season result isn't exactly an indictment of DVOA).

132
by MJK (not verified) :: Thu, 03/06/2008 - 7:49pm

One point on Eli, and I've made it before. It's amazing how one small chance in football can affect perception.

One play before the "Great Escape", Eli hurled an ugly, Grossman-like heave-ho well past his reciever. As soon as he released it, the two Giants fans I was watching the game with audibly groaned, even before the camera angle cut to the reciever, and as soon as they say Samuel get his hands on it, they yelled in consternation. Then Samuel dropped it.

Now I'm not arguing that it was a "sure thing" INT--it would have been a pretty spectacular catch if Samuel had made it. And I'm not arguing that Eli is a bad QB on the basis of one throw. Every QB occasionally has a bad throw, and that one was not particularly a gift, and was thrown under relatively desparate circumstances and hence more understandable than most.

But even so, imagine if Samuel had timed his jump slightly better and held on, and the Pats had won. It's certainly within the realm of reasonable possibility that this could have happened. Then the first 58 minutes, in which Eli had played an average or even below average game, would suddenly loom larger, and would have been capped off by "Eli chucking a game-ending interception". The Giants defense would have played the game of their life and held the greatest offense in recent NFL history to just 14 points, and yet their offense, led by Eli Manning, would have been unable to capitalize on that against a good but admittedly not great New England defense. The New York media would be all over him for "choking at the end".

In short, people's perception of Eli would be completely different, because of a small chance outcome that he had no control over once he let that ball go.

133
by Kurt (not verified) :: Thu, 03/06/2008 - 8:45pm

Yeah, but that might just be because they are Giants fans, and don’t like it when DVOA says their team isn’t as good as its record (which is, honestly, perfectly normal behavior for an NFL fan). Lots of teams have fans who dispute their team’s DVOA when it doesn’t agree perfectly with their W-L record, and most of the time, DVOA turns out to be the better predictor of future results than “fan-adjusted DVOA” or current W-L record.

Not that DVOA never misses something that fans pick up on, but just saying that DVOA is wrong doesn’t make that the case. And just because the team goes on to win the Super Bowl, doesn’t mean the fans were right all along that DVOA was underestimating their team’s chances (after all, if a bunch of teams have fans claiming DVOA was underestimating them, then the fact that one of them goes on to exceed their FO-predicted end of season result isn’t exactly an indictment of DVOA).

Sure, no argument. I'm not saying they were automatically right, simply disputing the idea that they drew four "diamonds". The idea that the Giants were getting lucky had cropped up well before the playoffs.

Lots of teams have fans who dispute their team’s DVOA when it doesn’t agree perfectly with their W-L record, and most of the time, DVOA turns out to be the better predictor of future results than “fan-adjusted DVOA” or current W-L record.

I've seen this asserted many times, but never demonstrated.

Not that DVOA never misses something that fans pick up on, but just saying that DVOA is wrong doesn’t make that the case. And just because the team goes on to win the Super Bowl, doesn’t mean the fans were right all along that DVOA was underestimating their team’s chances (after all, if a bunch of teams have fans claiming DVOA was underestimating them, then the fact that one of them goes on to exceed their FO-predicted end of season result isn’t exactly an indictment of DVOA).

There weren't a bunch of teams like that this year though, were there? There was one. And I'm not trying to "indict" DVOA - I'm not one of those people who think the Giants proved DVOA to be worthless. But if it's not falsifiable - if every result which contradicts it is simply written off as luck, while every matching result (Tampa, Green Bay) is regarded as a success - then DVOA isn't obvjective analysis, it's an article of faith.

134
by Chris (not verified) :: Thu, 03/06/2008 - 9:19pm

It isn't just the dvoa prediction, they were dogging the Giants before the season even started.

DVOA hasn't thrown out "predictions". The only predictions I saw were from Barnwell and you could have easily flipped a coin and beat him.

Of course 4 D-lineman were more valued than Eli in the super bowl, but last time I checked the MVP award was for 1 player. No one player was more important than Eli putting up a cool 14 points in the final 2 drives of the game.

135
by Pat (not verified) :: Thu, 03/06/2008 - 10:00pm

Sure, no argument. I’m not saying they were automatically right, simply disputing the idea that they drew four “diamonds”. The idea that the Giants were getting lucky had cropped up well before the playoffs.

What are you talking about? In October, DVOA had the Giants basically as a top-10 team. They didn't start dipping to the middle or lower-half of the rankings until November/December, and no one could really dispute that - not after the Vikings just destroyed them.

In articles near the end of the season, several Giants fans were saying that they didn't expect the Giants to do anything in the playoffs.

I’ve seen this asserted many times, but never demonstrated.

It's in the FO FAQ. It's easy enough to do - you look and see which correlates better with end-of-season winning percentage: estimated wins percentage at midseason or actual winning percentage at midseason? Estimated wins does.

Practically, it happens all the time. The Vikings in 2006, the Falcons in 2005.

136
by Pat (not verified) :: Thu, 03/06/2008 - 10:09pm

The percentage has to be higher, if point spreads are any indication.

You're probably right, but point spreads are self-selective: 14+ favored teams don't exist early in the season, and late in the season point spreads exist for other reasons (i.e. a team packing it in).

70% is probably too low. It's probably more like 80-85%.

Since 1994, teams favored by 7 or more in regular season games have won about 78% of those games. (In 2007, it was even higher, 79 of 92, or 85.9%).

79 of 92 is not 85.9%. It is, at best, 85.9+/-1%. It's probably more along the lines of 85.9%+/-10%, since counting statistics follow a Poisson distribution. Not including the error just seriously dilutes the point.

Playoff data is similar to regular season data; 7+ point favorites have won 74% of their games since 1994.

That, however, is biased, and probably low sample-size. How many 7+ point favorites have existed in the playoffs? How many years go by without one 7+ point favorite at all?

137
by Pat (not verified) :: Thu, 03/06/2008 - 10:27pm

They aren’t predictive though, that is the thing. I have done the math, and they get about 68% of the games right, which is pretty much exactly what everyone scores. Matching expected results is not proof of predictive powers.

I don't have the energy to explain this fully. Let me just say two things. First, no one will ever get better than about 70% in the NFL. The game itself doesn't predict itself better than that, and win distributions and beatloop frequency also tell you your best prediction percentage is around 70%.

Second, win prediction is not the end-all, be-all of predictivity. Given the fact that a simple maximum-likelihood ranking technique usually gets you to 65% or so, binary win prediction is actually pretty useless. You can do it with virtually no stats at all, so who cares?

What does matter more is how good you are at predicting the actual likelihood of the results (i.e. how many games did you think you're going to get right?), how fast within the year you reach maximum predictivity, and how much additional information is actually available (that is, how much can you break down the data without the fidelity being lost)?

Finally, your final statement makes, literally, no sense. Anything over 50% would be predictive. Being near the maximum predictivity allowable is certainly sufficient.

If you say "well, sure, it's ~68% predictive, but why should we believe it when we see things we don't agree with, and then we end up being right?", the question is "how often are you right, and how often is DVOA wrong?"

In the specific case here, how many times has a team had a 1-in-140 chance of winning the Super Bowl at the start of the playoffs? In all likelihood, probably about 40-50 of them. Why is it surprising that a 1-in-140 occurrance happens one time in 40-50 trials?

138
by Travis (not verified) :: Thu, 03/06/2008 - 11:29pm

That, however, is biased, and probably low sample-size. How many 7+ point favorites have existed in the playoffs?

Since 1994, 77 (roughly 5 per year, mostly in the divisional round), with 57 winning.

How many years go by without one 7+ point favorite at all?

None, at least since 1983 (the extent of my data set). There have been no fewer than two such teams every year.

14+ favored teams don’t exist early in the season.

True for the first two weeks only. In weeks 1 and 2, there have been only 3 such games since 1994 (Cowboys-Oilers Week 2 in 1994, Packers-Bears Week 1 in 1997, Titans-Browns Week 2 in 1999). However, the percentage of 14+ favorites remains around 3% for the next 15 weeks, as be seen from the following poorly-formatted table:

Weeks...........Games.........14+.......% of 14+
Weeks 1-2.......3.......429.......0.7%
Weeks 3-7.......33......948.......3.5%
Weeks 8-12......28......1012......2.8%
Weeks 9-17......35......1075......3.3%

139
by Travis (not verified) :: Thu, 03/06/2008 - 11:32pm

The "Games" and "14+" columns should be reversed above.

140
by Joe (not verified) :: Thu, 03/06/2008 - 11:38pm

In the specific case here, how many times has a team had a 1-in-140 chance of winning the Super Bowl at the start of the playoffs? In all likelihood, probably about 40-50 of them. Why is it surprising that a 1-in-140 occurrance happens one time in 40-50 trials?
The problem is that the Giants chances weren't that low. The week before the playoffs the Giants lost by three points to the supposed best team in the NFL. They very possibly might have won the game if their #1 corner didn't tear his abdominal muscles during a play that resulted in long TD pass to a suprisingly wide open Randy Moss.

Moreover the Giants receiving corps was crippled by injury for most of the season. Burress played all but one game on a single functioning ankle, Toomer did not get back to full speed until well into November due to his ACL tear last season, and Steve Smith missed about 3/4 of the season with various injuries. With Jeremy Shockey as their healthiest receiver it's no suprise that the Giants passing game suffered during the regular season. It's equally unsurprising that the passing game fared much better when Eli finally got some healthy (and talented) targets to throw to.

Plus, the Giants outplayed Tampa, outplayed Green Bay, and outplayed New England. In each one of those games, the Giants were simply the better squad. If their chances were only 1 in 140 what were those teams'?

141
by Dales (not verified) :: Fri, 03/07/2008 - 12:05am

Whoa, I just looked. I really like the Giants' schedule next year.

I don't think the Skins will be all that. I don't think the Eagles are going to be much better than they were this year. I think we are better than all four of the NFC West teams. I think we can at least go 500 vs the AFC teams we face. The Vikes spanked us and could be a problem, but the Panthers don't scare me.

I think we just may avoid the Super Bowl curse.

142
by Dales (not verified) :: Fri, 03/07/2008 - 12:28am

"Giants fans were claiming DVOA was wrong in October."

As one of those who had been making this argument, I want to make sure that, at least as far as I am concerned, it isn't taken out of context, while I stand by it completely.

I think DVOA was measuring things perfectly. But it was missing something it is not designed to handle. That while the Giants were playing mediocre football, and at times bad football, it was all because of their incredible propensity for shooting themselves in the foot. They were bad on third downs, not all the time, but consistently at the worst times. In past years, we were highly penalized; this year, it was in spurts but usually at critical times. We turned the ball over a LOT. We often played undisciplined. We often played dumb.

All of those things combined to make us look rather bad. But we were not bad. Even our performance overall wasn't bad. It was mediocre. But it was mediocre because despite all of the self-imposed handicaps, we were capable of very high levels of play.

This is what I was saying in October. Teams didn't usually beat the Giants. The Giants usually beat the Giants.

Somehow, in the playoffs, we didn't beat ourselves quite as much. We caught some breaks. We had our drops, but other teams did. We fumbled away the ball when we could have put games away, but other teams let opportunities slip away, too. We didn't play perfect, but neither did our opponents. And we didn't shoot ourselves in the foot as often as we had previously. And suddenly, we were a lot better than the Bucs, and just as good as the Pack, the Boys, and the Pats.

FO may have been saying all along that there was no way to predict this was going to happen. I kind of agree. But I also say that they should have seen it coming more than they did. This has been a talented football team for years now. They are good at most positions. Not great (except at DLine) but good-- and that includes the erratic and inaccurate Eli, who does enough other good things that move him into the good category despite his stats. But even if you don't agree at that, you have a team that has good receivers, good running backs, a good offensive line, a great defensive line, a good albeit thin linebacking corps, and a secondary that (with help from the DLine) was showing signs of being good as well.

A pretty good football team, after all.

143
by Dales (not verified) :: Fri, 03/07/2008 - 12:39am

"Lots of teams have fans who dispute their team’s DVOA when it doesn’t agree perfectly with their W-L record, and most of the time, DVOA turns out to be the better predictor of future results than “fan-adjusted DVOA” or current W-L record."

True. And I am not sure that DVOA should be changed.

But I have, at least, suggested something to look at (if it hasn't been already) statistically to see if maybe it can be an indicator. I think what I was seeing might be something that could be measured.

Drops. Penalties that wipe out first down plays on third down. Plays that give the opponent first down on third down. Unforced fumbles. Delay of game penalties. Too many men on the field penalties. Big plays given up when there are too few men on the field (this we probably can't get from the play-by-plays). But all of these *count* against a team, and they should because teams that do them tend to keep doing them.

But *if* a team could stop doing them, what level would they be at? This would be a measure of how dangerous a team is. I could easily imagine where a team is near the top of DVOA ratings but does none of these things often, but is playing a team that does these things often, and if the other team doesn't do them that game, it is an even matchup.

Sort of like the Super Bowl.

It wouldn't be a DVOA improvement. It might be a separate metric-- DVOA assuming none of these things happened, potential DVOA if you will. But it would measure the very thing I was suggesting before the Giants started their run.

144
by Dales (not verified) :: Fri, 03/07/2008 - 12:45am

"79 of 92 is not 85.9%. It is, at best, 85.9+/-1%. It’s probably more along the lines of 85.9%+/-10%, since counting statistics follow a Poisson distribution."

Pat, you are letting your statistical knowledge get the better of you.

There is no error in counting in this case (unless the person writing the post typed wrong or was making stats up). The spreads were what they were, the outcomes were what they were, and the 'winning' percentage can be calculated directly without a confidence interval or a margin of error.

145
by Dales (not verified) :: Fri, 03/07/2008 - 1:39am

Continuing on the last post-- the universe of games that fit the criteria was exactly the same as the sample size, ergo no statistical error to calculate or measure.

There is no unnamed universe of games fitting that criteria that we are just counting a portion of.

In the future, there may be more games fitting the criteria, but that isn't what was claimed; no one said that because results were 85.9% in the past, that will remain the case going forward. The way teams play can change, the way spreads are made can change, who knows.

146
by nat (not verified) :: Fri, 03/07/2008 - 1:43am

No ranking system will match the actual results all the time in the NFL. It is impossible. Here's why:

Look at the divisional play in the 2007 regular season. There were 16 pairs of teams that split their games. That means every ranking system in the world would get 16 of those games right and 16 games wrong. There was also the NFC North, with its crazy rock/paper/scissors results. Every ranking system would get at least 2 of those games right and at least 2 games wrong. (because CHI swept GB who swept MIN who swept CHI) That's 36 results that can't be improved upon.

That leaves 60 divisional games that can be usefully compared to your rankings. If you get all 60 games correct, you score 81% accurate. That's the best you can do.

DVOA rankings matched the results of 54 of those games. That's 90% of the games that matter, and an overall accuracy of 75%.

There were just 6 upsets that a different ranking system could have predicted: SF over ARI (twice), DET over CHI (twice), and NYG over PHI (twice).

Personally, I chalk these up to "any given Sunday". The Giants and the Eagles are pretty close in DVOA, so only the terminally insecure will obsess about their rankings or consider the games major upsets. The other two pairings are just "four of those things that happen".

If anyone has a single, simple theory about how DVOA could be changed to successfully predict all three pairs of upsets, I'd love to hear it. Otherwise, accept that DVOA is just about as good as it gets - in fact as good as it CAN get - in an objective ranking system.

147
by nat (not verified) :: Fri, 03/07/2008 - 10:22am

Here's a follow-up on theorized "better" rankings than DVOA.

If you try to tweak DVOA's rankings by swapping the Giants and the Eagles, you don't improve the rankings' accuracy (in terms of matching wins) at all.

For example: The Eagles beat the Vikings who wumped the Giants. If you rank the Giants above the Eagles, the ranking system gets those two games wrong.

The naked reality is that in the NFL at least 20% of the time and probably much more than that, the weaker team beats the stronger team. This can happen because a team genuinely changes in strength, it plays "above itself" for a few games, has good matchups, plays in conditions that favor its strengths, gets favorable officiating, is healthier, is really quite close in strength, or just plain gets lucky.

DVOA is pretty good at jugding teams' strengths over the season - almost as good as it could possibly be. Live with it.

148
by Kurt (not verified) :: Fri, 03/07/2008 - 11:54am

What are you talking about? In October, DVOA had the Giants basically as a top-10 team. They didn’t start dipping to the middle or lower-half of the rankings until November/December, and no one could really dispute that - not after the Vikings just destroyed them.

Actually the Giants never ranked higher than 12th in October, and there was an extensive discussion of them at the back end of the week 6 thread.

In articles near the end of the season, several Giants fans were saying that they didn’t expect the Giants to do anything in the playoffs.

And I'm sure I was one of them. I certainly didn't expect much, at least beyond the Tampa game. Any Giants fan who had predicted a SB win in December would have been (correctly) dismissed as a homer or a loon.

Nobody is saying that DVOA or any other system should have predicted the Giants to win the Super Bowl. Any time a 10-6 team beats 2 13-3 teams and a 16-0 team like that, it's a huge surprise. But DVOA looked into the numbers and saw a team that was *worse* than a "true" 10-6 team, and gave the Giants less of a chnace to win each playoff game than most observers.

This isn't to say that DVOA stinks, or that it should be "tweaked". The lesson here is that any statistical analysis will necessarily be limited. There were thousands of factors which contributed to the playoff run (some luck-based, some not) which couldn't possibly be taken into account by any statistical analysis. We can make our best guesses, but nobody can possibly know that, say, the Giants had a 1 in 140 chance to win the Super Bowl.

149
by Kurt (not verified) :: Fri, 03/07/2008 - 12:04pm

sorry - I wrote the second paragraph in my last post; it shouldn't have been italicized.

This isn’t to say that DVOA stinks, or that it should be “tweaked”.

Now, having said this, I will make what I hope is a constructive criticism. The game which really dropped the Giants in the DVOA rankings wasn't the Viking game, it was the following week, when they beat the Bears despite losing the turnover battle 4 to 0. DVOA *hates* turnovers, as it should, but I still wonder, as I did after the Tampa game, how useful they are as a predictive tool. Two reasons for this - it's an easier problem to fix than just about any other, and since it happens less than 5% of the time to even the worst teams, it's easier to get through a whole game without a turnover just by random chance, than it would be for example for a team with a poor OL and a slow runner to gain 150 rushing yards by chancing into a bunch of 5 yard runs.

I still can't think of a better explanation for the Tampa game. Tampa in the regular season had a much better turnover differential than the Giants but a worse record, which indicates that the Giants did a lot of other things better than Tampa. So when the Giants won the turnover battle 2-0 (I don't count the last hail mary), Tampa basically had no chance to win the game.

Really, Gerry comes out looking better than anyone - he was the one who thought the Giants had a lot of latent ability, and could be dangerous if they would stop shooting themselves in the foot. When they did, they were more than dangerous.

150
by MdM (not verified) :: Fri, 03/07/2008 - 12:34pm

Is anyone still reading? I've read about 114 posts... too much. I'm aghast at the Giants fans coming out of the woodwork to rail at the FO crew like comment-writers on the Fox website. Stop writing these irrational comments from your cell phone, Giants fan!

Seriously, your team just won a "Superbowl." You should be "happy", not filled with righteous indigation. Maybe someone will make an inspirational movie about Eli(sha)'s special month. Enjoy it--stop filling bandwidth with hate!

151
by Kurt (not verified) :: Fri, 03/07/2008 - 1:09pm

And post 149 goes double for the Viking game. Did the Vikings destroy them by marching up and down the field every time they had the ball, or by preventing the Giants from moving the ball an inch? No - they destroyed the Giants by returning three interceptions for touchdowns, and a fourth inside the 10. Was that predictive? The Giants didn't have an interecption returned for a touchdown in the eight subsequent games they played, so I would have to say no.

152
by nat (not verified) :: Fri, 03/07/2008 - 1:28pm

148 kurt:
I don't think you can blame the Giants' lackluster DVOA on four turnovers in one bad game. The Giants were the third most consistent team in the league.

DVOA had them pegged correctly for the regular season. They beat everyone that the regular season DVOA says they were better than, and lost to everyone they were ranked below, except the Eagles.

Who knows why that happened? Chance? Maybe they match-up well against the Eagles? Maybe they had their few "up days" against them? Maybe the Eagles had "down days"? But in a sixteen game season, you'd expect at least two upsets anyway. So they went the Giants way. Not a big deal.

The NFC playoff games don't require too much explanation. The Giants beat teams that were very inconsistent over the season, ranking near the top in variance. Did any of them have an "up day" in the playoffs? Nope.

153
by RoyFlip (not verified) :: Fri, 03/07/2008 - 1:31pm

#150 No hate here, maybe indignation; but I wouldn't categorize it as righteous, more mild indignation. I think it speaks very well of this thread that there is virtually no namecalling or any "we" beat your sorry butt kind of talk. I have abandoned other threads because of the puerile and hostile remarks. People here at least know the difference between "your" and "you're" And one can always assume that a misspelling(?) is most likely a typo. I sympathize with your having to listen to Giants fans babble on, but it only lasts a year. I am a Giants fan in Baltimore and I still have to listen to Ravens fans tell me how "we" kicked your butt....7 years ago....I think we are still talking about it because we really still don't believe it. Every time I watch the replay I expect Burress to trip in the end zone and have the pass intercepted.

154
by mrh (not verified) :: Fri, 03/07/2008 - 1:49pm

Re Eli and dropped passes (#43, #49, #59 et al). All numbers taken from Stats, Inc and not tainted by FO bias. I looked at all QBs with over 100 attempts (I chose this cutoff to include both Campbell and Todd Collins but coincidentally it is the cuttoff for the main QB DRAP/DVOA table at FO) - that was 51 players in 2007.

Eli had the most drops: 40. Brees had 37, Peyton had 35, and Brady had 31. Yet those those 3 qbs were 3 of the top 4 in DPAR and 13th, 5th, and 1st in DVOA (Eli was 30th and 35th). It doesn't seem likely that 3 to 9 more drops can be the explanation for such a gap in DPAR/DVOA.

Eli had the 3rd highest pct of drops/att, 7.6%. This was comparable to Alex Smith 8.0%; Rex Grossman 7.8%; John Beck 7.5% and David Carr 7.4%. Are drop rates (not totals) a reflection of qb skill as well as receiver ability? Probably not. They have a low correlation with poor throw rates (0.12 among my sample). Poor throw rate would be one -not the only- indicator of poor QBing. Pat or some other statisitcian can advise on the validity of correlating those two rates.

Eli was 4th (tied with Hasselbeck) in number of passes (80) assessed as Incomplete due to "Poor Throws". Anderson led the league with 93, then Palmer (85), Kitna (82), Romo (76) and Brady (73). That's not bad company, but whta about the rate of bad throws (poor throw/total att)?

Eli had the 18th highest (tied with Roethlisberger at 15.1%) rate of poor throws of my 51 QBs. Quinn Gray was #1 with 22.9%, then Alex Smith at 22.3% The names between Eli/Ben and Quinn were mostly poor last year with one exception: Moore, Croyle, Anderson, Lemon, Redman, Clemens, Dilfer, JMcCown, Bulger, Frerotte, Culpepper, Beck, Testaverde, Harrington, and Edwards. Ben's DPAR/DVOA ranks were 9th/6th, Anderson's 11th/15th so poor throws alone doesn't mean you will finish in the 30s in those stats.

Eli's cmp rate of 56.1% ties him for 40th with Edwards and Beck among these 51 QBs. If we throw out all the drops assessed to all the qbs, Eli's adjusted cmp rate (=cmp/(att-drops)) is 60.7%. This raises his ranking to 35th. If we assume that all the drops were completions ((cmp+drops/att)), Eli's 63.7% is 31st. So if we throw out the dropped passes for all these QBs, Eli's cmp% ranks are almost EXACTLY THE SAME AS HIS DPAR/DVOA RANKS.

Eli's "problem" with regular season stats, completion rate or FO stats, no matter how you cut it, was NOT due to dropped passes. BTW, I've argued elsewhere on this site that low cmp rate is not the only indicator of qb ability, so I'm not saying that is the only factor to consider in discussing Eli's ability.

155
by Kevin from Philly (not verified) :: Fri, 03/07/2008 - 5:24pm

Wow, this may be the first discussion of the Eagles offseason that doesn't talk about them trying to get Larry Fitzgerald from the Cards. My worst fear is that they'll package Lito and the 19th pick to move up and draft another stiff Offensive Lineman.

156
by Tom D (not verified) :: Fri, 03/07/2008 - 7:46pm

Re 140:

You're right DVOA doesn't have a injury adjustment, that's probably it's biggest flaw for predictive value.

157
by Jon (not verified) :: Sat, 03/08/2008 - 5:37am

150, I am happy. I'm thrilled. Then again, I always thought Eli and the team were a lot better than how they were playing in the regular season, and that Week 17 onwards was more representative of the team.

What I don't understand is how we can still possibly have this Eli debate now. Phil Simms had a mediocre career, but that perfect SB game memorialized him forever. Give up. It's over. Regardless of his future performance, the Eli Manning debate is definitively settled.

158
by MJB (not verified) :: Sat, 03/08/2008 - 10:53am

Sorry to interrupt the Irrational Eli Manning debate. But, in other NFC East news, the Eagles cut Takeo Spikes the other day. So does this move mean that Eagles are going to move Stewart Bradley into the starting rotation? And if so are they going to insert him in at on the weak side? Or do you all think that they will move Omar Gaither over from the middle to the weak side and put Bradley in the middle?

159
by LI Matt (not verified) :: Sat, 03/08/2008 - 8:42pm

What I don’t understand is how we can still possibly have this Eli debate now.

For me, the most fascinating thing about Eli's career is how emotionally invested people are in his failure. It is REALLY REALLY IMPORTANT to them that they can "prove" San Diego got the better of the deal, that Rivers is a better QB, etc.

Even now, after Eli has done the greatest thing a quarterback can do -- a last-minute drive for the winning score in the Super Bowl -- FO insists on telling us "he's not really that good."

As someone who was rooting for the Giants before Eli was born, I don't particularly care whether he has "good numbers" or "bad numbers". I just want my team to win. The Giants haven't had a losing record since his rookie year, and have made the playoffs in three straight years for the first time since 1986, so he must be doing something right.

160
by Matt (not verified) :: Sun, 03/09/2008 - 12:32am

Trent Dilfer, Rex Grossman, etcetera.

161
by Dean (not verified) :: Mon, 03/10/2008 - 9:23am

RE#158 - I'm sure this will get swallowed up by all the angry Giants fans - and really, can we PLEASE have an Irrational Eli Manning thread? - but the Iggles have said that Gaither will move out to WILL (good - he's better there) and Bradley will play MIKE (we'll see).

Suddenly, we're awfully thin at linebacker. AGAIN.

162
by Pat (not verified) :: Mon, 03/10/2008 - 5:24pm

But DVOA looked into the numbers and saw a team that was *worse* than a “true” 10-6 team, and gave the Giants less of a chnace to win each playoff game than most observers.

What would 'most observers' say that the Giants chance was of winning the Super Bowl? Certainly less than 1 in 12, since that's even among all playoff contenders. What about 1 in 50?

That's still optimistic. We can look at Vegas, for instance, and combine the odds there. That's a good estimate of what the public thought the Giants chances were. For the first 3 games, the aggregate Vegas odds were 39-to-1 against. The Vegas odds for the Super Bowl were 4.5-to-1 against.

That's an aggregate likelihood of 1 in 175.5. Nearly 1 in 200! And these are Vegas odds, which are basically representative of the general public. (Keep in mind this is not the same as the Giants chances of winning the Super Bowl at the outset of the playoffs, since the opponents weren't known).

Now, to remind you what Vince Verhei said the Giants chance was (and this wasn't DVOA, this was log5 and Pythagorean wins: if you use estimated wins, which uses DVOA, rather than Pythagorean wins, the results will change, but not that much) 1 in 143.

Yes. That's right. Vince actually said that the Giants had a better chance than the public gave them. If you use estimated wins rather than Pythagorean wins, the results would probably drop down a bit. Not as much as you'd think, because three of the Giants opponents - TB, DAL, GB - all had lower estimated winning percentages than Pythagorean, just like the Giants. But we're not talking about an order of magnitude change here. It wouldn't drop down by a lot.

So please, keep it in perspective. "Most observers" didn't give the Giants a chance in hell of winning the Super Bowl. DVOA basically gave exactly the same thing.

163
by Pat (not verified) :: Mon, 03/10/2008 - 7:19pm

Pat, you are letting your statistical knowledge get the better of you.

No, I'm really, really not.

and the ‘winning’ percentage can be calculated directly without a confidence interval or a margin of error.

No, it really, really can't. That sample size is chosen randomly: the question was asked this year, not ten years ago, and his sample set goes back to 1983, which is a random value. The number of random events that occur in that sample size clearly fluctuates. So the question of "is 79 of 92 statistically different from 75%?" and the answer is "... moderately, but not a lot."

ergo no statistical error to calculate or measure.

The statistical error is in the number of games counted, which is restricted by the observation point. It still has error if you want to compare two percentages to see if they're different.

If you ask the same question in a year, or last year, you'll get a different answer, even though you're presuming that the object you're measuring is intrinsic. That's pretty much the definition of statistical error.

164
by Alex (not verified) :: Tue, 03/11/2008 - 1:53am

There is no error in counting in this case (unless the person writing the post typed wrong or was making stats up). The spreads were what they were, the outcomes were what they were, and the ‘winning’ percentage can be calculated directly without a confidence interval or a margin of error.

Yes, and when you flip a coin 15 times and it comes up heads 9 times, do you suddenly think that it's not a fair coin, or that the chance of getting heads was really 60%? Of course not, the chance of getting heads was 50%, there just wasn't a large enough sample size to tell. When you're dealing with a small sample size, there's going to be some error compared to the actual probability. A small sample of games (which 92 is) isn't enough to show whether the teams would average an 85.9% winning percentage over the long run, or whether they were just lucky, and would actually win 75% of the games over a larger sample. Unless you think that it's impossible for a team with a 75% chance of winning each game to win 79 games in a 92 game stretch.

165
by Matt (not verified) :: Tue, 03/11/2008 - 8:30am

Everyone seems to think that the giants won on a miracle throw by ELI.Nobody seems to acknowledge the fact, that Tyree is a horrible receiver and he made the catch look harder than it was. He jumped maybe a foot off the ground, so it wasnt overthrown, he was standing in place so it wasnt like he had to run it down and catch it. He had both hands on it to began with, but somehow both hands lost the ball and the helmet came to play. It was only a miracle b/c Tyree caught it. They guy flat out drops more passes then he cathces.Any other receiver would have got the ball with both hands.

166
by ToastPatterson (not verified) :: Tue, 03/11/2008 - 1:53pm

I've read most but not all of this thread and I just had a couple of thoughts regarding all of the FO defenders who say that the Giants fans are teh stupid for ever challenging the supremacy of DVOA.

1. The Giants didn't win SB XLII because of luck. The Giants flat outplayed the Pats. Particularly the Giants defense front vs. the Pats O-line. They were getting pressure on Brady all game. Justin Tuck beat pro bowl guard Logan Mankins like a rented mule.
There may have been some luck, calls and non-calls that went both ways but LUCK did not decide that game.

2. Eli did not play well in "just the last 2 minutes" as many here have claimed. He played very well for the entire 4th quarter. His stats for the 4th quarter were 9-14, 152 Yds, 2 TD, 140.5 Passer Rating. He led his team on two 4th quarter go-ahead TD drives. The only other QB to do that in the Super Bowl is named Montana. Give the man his due.

3. The Giants did not "pull 4 diamonds," again implying random luck. They played better than their opponents in four consecutive games.

I don't know why it's so hard for FO devotees to accept that the Giants simply played better at the end of the season than they did earlier. This is not an indictment of DVOA. I think DVOA is a very good tool at showing how teams measure up based on past performance. But as everyone knows past performance is not indicative of future success or lack therof. I think the Giants just show DVOA's limitations as a predictive tool. The Giants stayed relatively healthy and played better over their last five games than they had previously. They were also better than their opponents over their last 4 games. Is that so hard for people to admit?

167
by Pat (not verified) :: Tue, 03/11/2008 - 2:53pm

They were also better than their opponents over their last 4 games. Is that so hard for people to admit?

I think the Giants actually did play better than their opponents over the last 4 games, but you don't have to play better than an opponent to actually win. Worse teams win games all the time.

That's part of the skill of being a coach, actually. If you manage the clock really well, you can beat an opponent even though they outplayed you.

168
by Pat (not verified) :: Tue, 03/11/2008 - 2:59pm

I should also note: the "four diamonds" bit that you're ragging on, you're completely misunderstanding.

When I said the Giants pulled "four diamonds" out of the hat, what I meant was that they pulled four excellently-played games out. I don't mean they got lucky. I mean they put together a great four-game stretch.

Average teams can put together four-game stretches of great play. That doesn't make them a great team, because those four-game stretches will be sandwiched between three-game losing streaks or something similar.

Any team, on any week, can put out a great game. What makes a team great is how often they can do that. During the regular season, the Giants couldn't do it too often. They still did it occasionally, like the last game versus the Patriots, or the first game against the Eagles.

During the playoffs, they played great for four weeks. That's not lucky, but it wasn't likely, which is the entire point.

169
by ToastPatterson (not verified) :: Tue, 03/11/2008 - 5:08pm

re # 168 "That’s not lucky, but it wasn’t likely, which is the entire point."

Fair enough on pulling diamonds being intended to describe something that's unlikely as opposed to lucky.

Still seems like an inapt analogy since pulling cards from a deck depends entirely on luck unless you're Penn Jillette or Josef Kavalier. Football actually requires some skill.

Maybe a better analogy would be winning the Triple Crown. Possible; very unlikely; and very cool when it happens.

170
by Pat (not verified) :: Tue, 03/11/2008 - 6:03pm

Still seems like an inapt analogy since pulling cards from a deck depends entirely on luck

Nono - pulling cards from a deck is entirely random. That's the point.

And from a fan's point of view, it is entirely random when an average team plays like a great team. It happens. Infrequently, but it happens - and "when" it happens isn't something you can really predict.

Fans like to think you can predict it, of course: pointing to the Eagles, Ravens, and Giants stepping up against the Patriots and saying "see, teams step up against the Patriots." Or with the Giants, saying "the team knew they could hang with the Patriots, and so they played like it afterwards." But the truth is that teams stepped up against crappy teams, too. The Eagles utterly destroyed the Lions, for instance, and there were about a hundred examples last year, too. And the Eagles hung with the Patriots too, and then promptly self-destructed against the Seahawks, completely nuking their season.

I don't even think from the player's point of view it's anything but random. Sure, they give lots of "man, we really felt the momentum" stuff, but that's self-selective. Go back and interview the Eagles from last year when they lost to the Saints. They were riding on a ton of momentum... and still lost.

It's just a question of everyone starting to play their best at the same time. None of them can control when the others play their best, and so again, it's just random when it happens to all come together.

171
by Kurt (not verified) :: Tue, 03/11/2008 - 7:39pm

Pat, the problem with the deck of cards analogy is that we don't know what's in the deck. Some people think they do, but those people are wrong. The deck is constantly changing, and the best evidence of its contents is what we see one week at a time, unlike a real deck of cards.

*When* an average team plays a great game it might be random but great games themselves are not random - the higher the quality of the team, the more likely a great game is. When a team plays five great games in a row it should alter the overall assessment of the team, even if it doesn't make them a great team. Note that a deck of cards is exactly the opposite - when you draw 4 diamonds in a row it makes the next card *less* likely, not more likely, to be a diamond.

172
by Pat (not verified) :: Tue, 03/11/2008 - 11:33pm

Pat, the problem with the deck of cards analogy is that we don’t know what’s in the deck.

By the end of the season, we know at least 16 of the cards in the deck. The chance that those 16 are close to the mean and the next 4 are not is much greater than the chance that the next 4 are close to the mean and the first 16 were not.

That assumes a little about the distribution of performances, but it doesn't assume a lot. The only way to violate that would be if teams weren't really different from each other. This can *also* be inferred from win distributions and maximum likelihood analyses.

It also assumes that the mean is constant (teams don't change), which is obviously wrong. But if that's the zeroth-order assumption, then the first-order assumption is something like weighted DVOA, which, doesn't really help in this case - the Giants weren't exactly "getting better" until the last week of the season (at best).

And the faster and faster you let the mean change, the more you move towards the assumption of "the strength of a team is unknowable from week to week," which is also clearly not true given the fact that we can predict most teams pretty well.

173
by Pat (not verified) :: Tue, 03/11/2008 - 11:34pm

when you draw 4 diamonds in a row it makes the next card *less* likely, not more likely, to be a diamond.

Random draws of cards assume replacement. Otherwise it's not random, since it's not a memoryless process.

174
by BDC (not verified) :: Wed, 03/12/2008 - 2:34am

124:

"In my book, 68% at picking the winner PLUS objective analysis is more interesting than 68% plus ad hoc analysis."

Yes, but at the end of the day, they all get the same results. Remember, DVOA doesn't claim to be AS GOOD as everything else, it claims to be better. Simple analysis shows this to be untrue.

130:
"If they get significantly more than 50% of the games right, they are predictive. They don’t have to be more predictive than Vegas in order to be predictive."

No, but they do have to be more predictive then Vegas to be useful. Anyone can can get more then 50% of the games right. That is the whole reason point spreads exist; because everyone knows picking the overall winner of a game is easy. Incidentally, if I tell you I have the ability to predict what number a die WON'T land on, and it turns out I am right oh say, about 5 out of 6 times, am I predictive? Am I useful?

137:
"Second, win prediction is not the end-all, be-all of predictivity. Given the fact that a simple maximum-likelihood ranking technique usually gets you to 65% or so, binary win prediction is actually pretty useless. You can do it with virtually no stats at all, so who cares?"

Well that is kind of the point isn't it? If the stats here don't tell us anything more predictive then anything else out there, they aren't predictive, now are they?

"Finally, your final statement makes, literally, no sense. Anything over 50% would be predictive. Being near the maximum predictivity allowable is certainly sufficient."

No, it isn't. If you provide results that are exactly as expected from any other method, you are no more predictive then that method. To reuse my previous example, I predict that the following die roll will not be a 6. I predict that on every die roll. I will of course be right 5 out of 6 times. Am I giving you any useful information? I mean, I am right far more often then I am wrong. Does that mean anything?

"If you say “well, sure, it’s ~68% predictive, but why should we believe it when we see things we don’t agree with, and then we end up being right?”, the question is “how often are you right, and how often is DVOA wrong?”"

Well that is the whole point, isn't it? DVOA is right exactly as often as just about every other system, statistical or not, out there. So what then makes DVOA more predictive then those other systems out there that seem to produce the exact same results?

175
by Kurt (not verified) :: Wed, 03/12/2008 - 11:59am

By the end of the season, we know at least 16 of the cards in the deck. The chance that those 16 are close to the mean and the next 4 are not is much greater than the chance that the next 4 are close to the mean and the first 16 were not.

I'm not arguing that. What I am arguing against is the idea that the first 16 games *is* the mean, and everything that happens after that is an aberration. Again, I'm not saying you're doing that, but some are.

That assumes a little about the distribution of performances, but it doesn’t assume a lot. The only way to violate that would be if teams weren’t really different from each other. This can *also* be inferred from win distributions and maximum likelihood analyses.

It also assumes that the mean is constant (teams don’t change), which is obviously wrong. But if that’s the zeroth-order assumption, then the first-order assumption is something like weighted DVOA, which, doesn’t really help in this case - the Giants weren’t exactly “getting better” until the last week of the season (at best).

They were also getting better over the course of the playoffs. Put it this way - the Giants' last five games (week 17 + playoffs) also happened to be the their five best games. The chance of this happenening by pure chance - that out of 20 games, the last 5 would also be the best five - is by my calculations about 1 in 15,000. So as unsatisfying and unprovable assertions like "they put it together" or "the week 17 game gave them confidence" are, such explanations are more likely than the distribution of performances happening by pure chance.

And the faster and faster you let the mean change, the more you move towards the assumption of “the strength of a team is unknowable from week to week,” which is also clearly not true given the fact that we can predict most teams pretty well.

I would say that the *precise* strength of a team is not knowable, but you can establish a general range.

Random draws of cards assume replacement. Otherwise it’s not random, since it’s not a memoryless process.

Under that scenario drawing a "good" card would have no effect on the next card. This is also not true in football - a great performance is evidence of team quality, and affects, at least a bit, the chances of the next performance being good.

176
by TracingError (not verified) :: Wed, 03/12/2008 - 1:55pm

I can't even read this whole thing but someone explain to me how two near interceptions and one Houdini escape to throw the ball up for grabs constitutes a good 2-minute drill. The job got done, but this wasn't exactly Montana against the Bengals.

177
by Gozer Gozarian (not verified) :: Wed, 03/12/2008 - 3:16pm

ahh, I see it's now two near interceptions.

178
by Pat (not verified) :: Wed, 03/12/2008 - 3:19pm

Under that scenario drawing a “good” card would have no effect on the next card. This is also not true in football - a great performance is evidence of team quality, and affects, at least a bit, the chances of the next performance being good.

Uh, no, it doesn't. A great performance increases the best estimate for the true strength of the team, but it has no effect on the actual true strength of the team (which is presumed fixed) and therefore doesn't change the next likely outcome.

If you've got an unknown quantity of diamonds in a deck, consistently drawing diamonds increases the best estimate for diamond fraction in the deck, but it has no effect on whether or not the next card is a diamond other than its effect on the estimate.

Like I said before, a lot of that hinges on the assumption that the true strength of a team is fixed, sure - but the extreme opposite assumption is that the true strength of a team is completely mutable, which would mean that football isn't predictive at all, which we know is false.

The next-to-most extreme assumption is that the true strength of a team is best estimated by the previous week's performance, and that alone - and that also isn't indicated in the data. What is indicated in the data is that the best estimate of a team's strength is the average over the past 10 or so weeks (or so) - and that doesn't help the Giants.

(Note that, again, as a disclaimer, none of this is meant to impugn the Giants. All I'm saying is that their performance wasn't predictable. Fans should rally around that, in my opinion, but for some reason people treat it as an insult.)

179
by nat (not verified) :: Wed, 03/12/2008 - 3:40pm

The Giants success in the playoffs was not predictable from their overall regular season success (DVOA, net points, even win/loss) nor by trends in their regular season play (DVOA, net points, win/loss...). Why should that bother anyone? It's both obvious and not insulting.

Their success is due to either their becoming a better team, or due to the "any-given-Sunday" effect - some combination of the Giants playing above themselves,their opponents playing below themselves, and the usual random stuff like fumble recoveries, timing of plays, etc.

Accept it. The Giants were not very good (or bad) in the regular season. They were certainly more successful and probably an improved team in the playoffs.

Good for them. They played their best football in the playoffs. Who would have guessed it? Among those guessers, who would have had compelling statistics to convince anyone else it was going to happen?

Don't you love it when your team gives you a pleasant surprise?

180
by Alex (not verified) :: Wed, 03/12/2008 - 9:21pm

“In my book, 68% at picking the winner PLUS objective analysis is more interesting than 68% plus ad hoc analysis.”

Yes, but at the end of the day, they all get the same results. Remember, DVOA doesn’t claim to be AS GOOD as everything else, it claims to be better. Simple analysis shows this to be untrue.

DVOA doesn't claim to be better at picking winners. DVOA claims to be better at explaining why teams win by breaking down the contributions of the different parts of the team.

“Second, win prediction is not the end-all, be-all of predictivity. Given the fact that a simple maximum-likelihood ranking technique usually gets you to 65% or so, binary win prediction is actually pretty useless. You can do it with virtually no stats at all, so who cares?”

Well that is kind of the point isn’t it? If the stats here don’t tell us anything more predictive then anything else out there, they aren’t predictive, now are they?

First of all, you seem to be using the word "predictive" in an awfully strange way. Predictive doesn't mean "the most predictive thing out there". Adjectives of the form X-ive don't generally mean "the most X-ive thing in existence". They just mean X-ive. Predictive just means that when DVOA says a team is better than its opponent, that team wins more than 50% of the time (i.e. it really is better). That's all. It's not a competition.

And keep in mind, DVOA comes straight from the play-by-play. It doesn't know what the Vegas line is, so replicating the predictive success of the Vegas line would be evidence that DVOA is at least as predictive as Vegas.

As for what the point of it all is, it depends on what you're trying to do. If you're trying to predict winners better than Vegas does, then yeah, FO's stats aren't going to help you a whole lot.

But that's not what FO is trying to do. FO's goal isn't to use their analysis of statistics to pick a higher percentage of winners. FO's goal is to use their analysis of statistics get a better idea of the true strength of NFL teams, and better understand the contributions that individual players/units give to teams. Not "use statistical analysis to get a better idea of the true strength of NFL teams, and then use that to pick a higher percentage of winners than Vegas". FO wants to know more about why teams win, not for the sake of improved betting strategies, but just because they're curious about it as football fans.

FO wants to know whether Joe RB really helped his team win with his 330 carry, 1000 yard season by getting crucial gains that sustained drives in tough situations, or if he was just compiling a bunch of yards because he had so many carries. Vegas can't tell you that. Point spreads don't give you any information about which players contributed the most to the team's victory, or which units were most detrimental to the team's efforts. And that's what the point of DVOA is. Vegas might be (slightly) better at picking winners than DVOA, but DVOA is way better at telling you what each player/unit contributed to their team's wins.

For instance, if the Vikings were about to play the Bengals, and Vegas had the Vikings as favorites, you'd know that the Vikings were probably better than the Bengals. Great! But what if you wanted to know which team had a better QB? Suddenly, Vegas isn't very useful. I mean, Tavaris Jackson probably won't ever be even close to as good as Carson Palmer. So, to figure out who's the better QB, you might use their statistics. But which statistics do you use? How do you know that they're the right ones?

Well, you might develop a model to approximate how many yards in a certain situation turn into how many points, on average, and how those points typically turn into wins. Then, you could analyze every play of the NFL season, and you'd have your numbers! But wait, what if some part of your model were wrong? What if you'd originally thought that getting 5 yards on 2nd-and-12 against an average defense was a good play, and rewarded QBs for doing that? What if you were wrong about all sorts of situations, and were rewarding people for plays that actually hurt their team's chances of winning? How would you know that your model is picking out the right things to reward (after all, we don't want to reward a QB for throwing for a lot of yards if it doesn't make his team more likely to win)?

Here's how: you'd treat each team as if it were just one player, and figure out how your model would rate that player. Then, you'd make sure that, when teams with a higher rating play teams with a lower rating, the higher rated team usually wins. That way, you'd know that whatever it was that got a player a higher rating in your system was increasing his team's chances of winning.

Then, you might keep track of what each team's rating in your system would be if it were treated as one player, and notice that it's pretty steady, year-to-year, compared with most other measures of team strength, like W-L record. You might think of it as a better approximation of a team's true strength. And of course, you'd regularly tweak your model by checking whether certain changes to the model would make teams with higher ratings win more often. Eventually, your model might even get as many games right as Vegas does.

But here's the catch: even if you knew the exact true strength of every NFL team at the beginning of the year, you still wouldn't be able to predict games all that much better than Vegas.

After all, who's better, the Giants or the Eagles? Giants? Ok, so since Dallas beat the Giants twice, we have to put them higher than the Giants, otherwise we'd be wrong in two games already. But if we do that, we'd be wrong when the Eagles beat the Cowboys. And if we don't put the Vikings over the Giants, we get their game wrong. But if we do put the Vikings over the Giants, we'd be wrong when the Eagles beat the Vikings. So no matter which team has the higher true strength, you're going to be wrong on at least two games right there. And then, what about the Redskins, are they better than the Giants? If they are, then we'd get the first game between the two wrong. But if they aren't, we'd get the second game wrong. Etc.

So, even if you could go into the future and find out all the game outcomes before the year started, and then use that to rank the teams from best to worst, you'd still have a list that would be wrong on a bunch of games. Add in the fact that we don't have all the game outcomes (and play-by-plays) and we're going to get a few games wrong in the beginning of the year because we won't have enough information yet.

But who cares? We're not trying to beat the Vegas line! We're trying to figure out just how bad Tavaris Jackson is.

(We then realize that there are some truths in life that are better left undiscovered. But that's neither here nor there.)

181
by Pat (not verified) :: Wed, 03/12/2008 - 9:39pm

No, but they do have to be more predictive then Vegas to be useful.

Is there a "bolder than bold" HTML option? They don't let me use the blink tag, or any of the size tags, so...

Useful for gambling, maybe, but some of us don't give a *!#( about gambling. Vegas gives aggregate odds. Vegas doesn't tell you why a team is likely to win. It doesn't tell you what to look for in the game to get an idea as to whether or not a team will win, or how it is winning.

Vegas doesn't give you a way of really isolating the offense, defense, and special teams. I don't know of any statistic which even attempts to do that other than DVOA.

Predicting the game outcome is about the least interesting thing in football to me, because I know the final answer is going to be something like "a team like this wins 60% of the time" or some crap like that. Big deal.

182
by BDC (not verified) :: Thu, 03/13/2008 - 5:07am

180:

"DVOA doesn’t claim to be better at picking winners."

Yes, it does. Do you really need me to cite where they have done this?

"First of all, you seem to be using the word “predictive” in an awfully strange way. Predictive doesn’t mean “the most predictive thing out there”."

Yet that is exactly what this site has made claims of. Again, do you want me to cite? I will if it is needed.

"It doesn’t know what the Vegas line is, so replicating the predictive success of the Vegas line would be evidence that DVOA is at least as predictive as Vegas."

Yes, but remember, DVOA claims to be the BEST, not simply just as good as everything else. Again, do I need to cite?

"But here’s the catch: even if you knew the exact true strength of every NFL team at the beginning of the year, you still wouldn’t be able to predict games all that much better than Vegas."

Agreed. Yet that is what DVOA claims to be able to do. One more time, need a cite?

183
by nat (not verified) :: Thu, 03/13/2008 - 6:06am

182: "Do you really need me to cite where they have done this?"

Well, since citing would have taken you about 10 seconds to type in, we all know you don't have the cite in your hand already.

How about reading "Method to Our Madness" which explains what DVOA and DPAR are?

The biggest advantage of DVOA is the ability to break teams and players down to find strengths and weaknesses in a variety of situations.

Vegas odds don't do that.

or talking about Fantasy Football...

DVOA does a better job of distributing credit for scoring points and winning games.

Vegas, odds, and point spread are not even mentioned. Odds don't assign credit to players or units.

...or from the FAQ

There are many websites that provide gambling services. Football Outsiders is not really one of them. We know that many people use our stats to influence their gambling decisions, but our stats and analysis are primarily here to enhance understanding of how to win football games and how to build a Super Bowl champion.

Remember, this site began with the question "Do you need to establish the run to win?" (see What are we talking about here? in "about this site")

So if your whole rant is about how DVOA has failed because it doesn't pick your bets for you, I can only feel sorry for you.

184
by Alex (not verified) :: Thu, 03/13/2008 - 3:48pm

“DVOA doesn’t claim to be better at picking winners.”

Yes, it does. Do you really need me to cite where they have done this?

That'd be interesting, because they'd be contradicting themselves. I can cite where they have specifically stated that DVOA is not better at picking winners. They claim that it is the best at distributing credit to the players/units that most deserve it.

From Methods To Our Madness:

In the aggregate, DVOA may not be quite as accurate as some of the other, similar "power ratings" formulas based on comparing drives rather than individual plays, but, unlike those other ratings, DVOA can be separated not only by player but also by down, or by week, or by distance needed for first down.

185
by Gerry (not verified) :: Thu, 03/13/2008 - 10:49pm

"Of course not, the chance of getting heads was 50%, there just wasn’t a large enough sample size to tell"

Not what I was arguing, nor what you said.

You said this:

“79 of 92 is not 85.9%. It is, at best, 85.9+/-1%. It’s probably more along the lines of 85.9%+/-10%, since counting statistics follow a Poisson distribution.”

Except for the fact that the entire universe was just 92. No more, no less.

If you pull 100 jelly beans out of a jar, and 97 are red, then guess what-- you got 97% red. Not probably more along the lines of 85.9% +/-10% since counting statistics follow a Poisson distribution. You can argue, however, that the likely percentage of red beans in the jar is along the lines of 85.9% +/-10%.

Furhter, if you pull 100 jelly beans from a jar that had 100 jelly beans in it, and you got 92% red, then if you want to argue that the "percentage of red beans in the jar is along the lines of 85.9% +/-10%", then you are being obtuse.

There were that number of games that fit the criteria. The entire universe being measured was 92. The percentage can be calculated precisely, as to what happened. It might be a case where past performance does not indicate what future performance would be (count on it,even) but to argue that it wasn't the exact percentage stated is taking statistical concepts and using them to prove you don't understand them.

186
by Chris (not verified) :: Fri, 03/14/2008 - 8:36am

The entire year Eli was the butt of a lot of jokes here, but in the end is the the DVOA makers that must answer the questions on the defensive.

I was right on Eli, Vick, and T-Jack.

187
by Alex (not verified) :: Fri, 03/14/2008 - 1:26pm

If you pull 100 jelly beans out of a jar, and 97 are red, then guess what– you got 97% red. Not probably more along the lines of 85.9% +/-10% since counting statistics follow a Poisson distribution.

Fair enough, but I think you're taking Pat a little too literally here. He didn't mean the past percentage wasn't exact, he meant the actual likelihood that it was supposed to represent wasn't exact.

188
by Joe (not verified) :: Fri, 03/14/2008 - 4:06pm

So do any of you agree with the notion that Jason Campbell is a significantly better QB than Eli Manning?

189
by nat (not verified) :: Fri, 03/14/2008 - 6:08pm

Back in 146, I explained why there were a minimum for 18 upsets in divisional games - where I defined upset as a lower ranked team beating a higher ranked team. It didn't matter HOW you ranked the teams for the season. As long as you used a one ranking for the whole season, these (almost 20% of games) upsets could not be avoided.

Now I've looked at the whole season, both with and without the playoffs.

For the regular season there were at least 48 upsets out of 256 games. If you include the playoffs, there were at least 52 upsets out of 267 games.

What did DVOA think? DVOA rankings called 59 games upsets in the regular season and 64 games upsets if you include the playoffs.

To put it another way: DVOA thinks 24.0% of all games were upsets. We know that at least 19.5% were upsets, and we know that there could have been more - in the sense that a "correct" ranking of strength doesn't have to be the one that calls the fewest games upsets. Weaker teams do beat stronger teams now and then.

I suspect that adding back into DVOA the actual value for fumble recoveries and fumble and interception returns, and giving special teams credit for opponents' missed fieldgoals and kickoff/punt lengths would give an even better result. DVOA excludes them because they are not predictive, not because they don't cause wins.

Well, that's about it. DVOA gets it pretty good. It's at most 12 games away from "as good as it gets" for the season, and probably closer than that to measuring the "true" strength of the teams.

190
by Pat (not verified) :: Fri, 03/14/2008 - 11:03pm

If you pull 100 jelly beans out of a jar, and 97 are red, then guess what– you got 97% red.

Yes. And the answer to "what percentage of jelly beans in the jar are red?" is 97%. Exactly.

If the jar is filled with jelly beans from a dispenser which dispenses a percentage of red beans and a percentage of non-red beans, however, and you count the number of red beans in the 100 you get from the dispenser and get 97, the dispenser does not put out 97% red beans. It puts out something like 97 +1.5% or so depending on what kind of confidence intervals you want to use because you're too close to a zero count in non-red beans, so the definition of a confidence interval depends on the presumed shape of the distribution.

The original question was not "what percentage of 7+ point favorites in the playoffs have won?" it was "what percentage of 7+ point favorites in the playoffs win."

The first is your example. The second is mine.

191
by Pat (not verified) :: Fri, 03/14/2008 - 11:25pm

#189: Theoretically some of the upsets could be avoided if you included home-field advantage, since that's an additional variable. However (from experience) I know that that actually doesn't work, and if you want to make HFA fixed for a team, it actually makes things worse on average.

In fact, you could also add a second ranking to create possible loop structures (like, say, rank Team A (1,2), team B (1, 1), and team C (2, 1) and say something stupid like "if rank 1 or rank 2 of both teams are the same, the higher rank 1 or lower rank 2 wins, otherwise, the lower rank 1 wins", which would mean B beats A, C beats B, and A beats C.

However, you can also show that the win data for the NFL doesn't support a system like that.

Part of the reason why I'm saying this is that you're implicitly assuming that "if DVOA(A) greater than DVOA(B), DVOA(A) wins," whereas there are actually three (main) rankings for each team, and so you could actually do something much more complicated.

For instance: call the win function W(A,B), and if it's greater than 0, A wins. Instead of "W(A,B) = DVOA(A)-DVOA(B)", you could have some function where W(A,B) is (ODVOA(A)-ODVOA(B)-DDVOA(A)+DDVOA(B))+f(SDVOA(A)-SDVOA(B)), where f is a coefficient that depends on the first sum of terms (offense+defense, basically). If that sum is positive, f is near zero. As that sum goes negative, f grows to 1.

Basically, you'd be saying "if the game is going to be mainly offense, special teams doesn't matter much," and "if the game is going to be mostly defense, special teams matters a lot."

However, in the end, it's not going to net you much, if anything at all. Part of the problem is that coaches can choose the pace of the game, and changing the pace of the game can stress certain components of the team. A many-drive game can put a lot of emphasis on special teams. A short-drive game can put a lot of emphasis on pass defense.

In that case, you'll never be able to make a perfect ranking that correlates with wins because game plans (which are 'random' from a true strength standpoint - that is, they can change week to week freely) can easily change the outcome of the game. And you'll never be able to measure the likelihood of a coach choosing some strategy.

192
by nat (not verified) :: Sat, 03/15/2008 - 12:32am

191:
Excellent comments all.

My point was to see how good a ranking system straight DVOA was by seeing how close to perfect it got. "Perfect" in this case meant "fewest possible disagreements with actual game winners".

The answer is "pretty close".

It's also useful to prove that at least 19% of all regular season games are true upsets. Since the average rank difference in a regular seasn game is about twice that of the average playoff game, I know to expect more playoff upsets - about half way between 19% and the 50% that equal teams would have. That's about 34% upsets or about 4 games each playoff season.

This year was a bit higher with 5 upsets by DVOA rankings. I would be worried (a little) if the rankings had predicted the playoffs with 100% accuracy. It would also be of concern if DVOA had missed 6 or more games. 5 upsets is not too bad, considering.

It's interesting to me that DVOA missed 5 playoff games, while the "best possible" ranking for the entire season missed 4 playoff games (and would call the NYG-TB game correctly). It just goes to show that the Giants' playoff run was a miraculous overachievement, regardless of the causes, and not just a misperception of their regular season strength.

193
by Jerry (not verified) :: Sun, 03/16/2008 - 4:48pm

Regarding the Giants and predictive statistics, I've always thought that the Giants are a very difficult team to get a quantitative handle on.

The reason for that is that they're a team with a high variance. They're a high-ceiling/low-floor team on both sides of the ball, especially offense. While the overall #s might aggregate to an average-ish team, they are capable of playing at a very high level.

Nobody embodies that concept more than Eli Manning. Yeah, his overall numbers pretty much suck. But he's capable of playing really well. Not "game manager" good. Really great elite "carry the team" good. Every year he plays a couple of games where he's out of this world. The other QBs in his statistical strata don't do that.

Defensively, the Giants are in a similar (but less pronounced) boat. The pass rush can be devastating. When the front 4 is getting to the QB, they can win the battle on their own. When they're not, the back 7 is exposed and the defense looks awful.

Put it together and you have a team that can get thrashed by the Vikings or beat any of the top teams.

With the Giants, anything can happen. A team like the Redskins probably looks similar in aggregate, but has a lower variance. The Redskins play a lower risk game and you basically knew what kind of a game you were getting with the Gibbs-Skins -- close to the vest, close game that could go either way at the end. But they weren't in a position to play exceptionally well (and they didn't lay many massive eggs either).

I love stats and I really like this website, but sometimes you have to say that stats are better at analyzing certain groups of events (teams) than others. IMO the Giants fall into that category.

194
by nat (not verified) :: Sun, 03/16/2008 - 10:17pm

193:
Jerry, what's your theory on why the "high variance" Giants had the third lowest DVOA variance in the NFL for the regular season?

Variance: "You keep using that word. I do not think it means what you think it means."

195
by nat (not verified) :: Mon, 03/17/2008 - 1:40pm

More on 192...

For the last five years, we've averaged more than 4.5 upsets per playoff season. That's about 41% upsets, higher than I had suggested in 192.

My earlier estimate was based on playoff teams being equally distributed between the average and the best seasonal DVOAs, with an essentially random paring for each game.

That's not a very good model. The playoffs are designed to bring in the best teams, so they tend to feature most of the top 12 teams with one or two outliers. The playoffs themselves are arranged to pair up roughly equal teams for the conference championships and Superbowl, and exclude the best teams from the first round. So it's perhaps not a surprise that upsets are common.

So here are some rules of thumb for how often to expect upsets:

1) Regular season games with widely separated ranks - less than 20% upsets.
2) Average regular season game - 20% upsets.
3) Regular season games between teams close in DVOA - more than 20% upsets.
4) Playoff games - around 40% upsets.

Despite the Giants' historical playoff run, this year was the median year for playoff upsets in the past five seasons.

As noted earlier, the 20% upset rate in the regular season is not a sign that DVOA is broken. Those are true upsets against the season rankings, caused by variations in teams' play, or by other unpredictable factors.

196
by RoyFlip (not verified) :: Tue, 03/18/2008 - 4:29pm

176, 177 Yes, the Irrational Trash/Defend Eli debate rages on!
As time passes, some people will claim he threw more than 12 "near" interceptions If it wasn't caught by a Giant, it will qualify as a "near" interception. As someone else here put it so well, there seem to be those who are so invested in Eli's failure that they now will denigrate his performance almost to the point of saying the Super Bowl MVP played terribly. As to the throwing it up for grabs claim: Watch the play again- After he slips away from a sack,he runs right and has Vrabel (I think) bearing down on him. He has to put a lot of air underneath it or it gets blocked. He hits Tyree 20+ yards downfield while on the run and being rushed. His receiver is covered. Would you have him ask Vrabel to wait until someone is open? He has to hit Tyree up high or Harrison either levels him or knocks it away. Add to this down, distance, time remaining, Super Bowl, and a very good defense that knows you have to throw. I think he threw for something like 150 yards in the 4th quarter. That is somehow throwing it up for grabs? Also, as great as the Giant defense was, they couldn't stop the Patriots from driving down the field to go ahead, putting Eli in the position of having to score a touchdown to win. Remember that when you say that Eli was the benefactor of the Giant's defense. He certainly was, but it was the other way around at the end.

197
by RoyFlip (not verified) :: Thu, 03/20/2008 - 11:36am

"Eli was the benefactor"
I meant beneficiary. Although its not like anyone else is still reading this. I need to get a life.

198
by PW (not verified) :: Fri, 03/21/2008 - 5:31pm

I read it 197...

Good points

199
by Alex (not verified) :: Tue, 03/25/2008 - 10:38am

I think it bears mentioning that if you add in the postseason games, the Giants' DVOA goes up to about 9%, which would put them in the top 10 in the NFL, and second in the NFC East. So saying they were "the worst regular season team in the NFC East" isn't the same as saying they were the worst team in the NFC East this year. They just didn't have their best games until the playoffs started.

200
by miami (not verified) :: Wed, 04/02/2008 - 3:17pm

'But my point is that nobody can know what percentage a team has of winning a specific game.'

Sure they can. Vegas does it every week. Your comment, like many here, shows a shocking lack of understanding of variance and statistics.

NYG was ~20% to win the SB and they did. The ex-ante estimate was still correct, given the volumes of info and $Billions of dollars wagered on the game.

Mario Mendoza occasionally did hit a double in the gap. You don't turn around and say that's impossible because he's a. 200 hitter.

A 3-13 team wins about 20% of their games, and each time will almost always be an upset. Etc.

201
by Austinfan (not verified) :: Sun, 04/06/2008 - 4:19am

It's amusing to read all the Giant fans trying to rewrite reality. The Giants were a mediocre team that got hot, and got the breaks. The Eagles were a mediocre team (mostly due to McNabb being injured) that didn't get the breaks and never got it together until it was too late.

Go look at the Giants' second half, after going 6-2 against the soft half of their schedule, they went 4-4 the second half, when good teams hit their stride.
In the playoffs, they weren't exactly a dominant team, the Bucs had no business being in the playoffs, they just had a cupcake schedule in that division.

The three real wins, all by 3 points, not exactly a dominant performance.
In other words, what you'd expect when a mediocre team has a hot streak, they win some games they normally would lose, but they don't win them by big margins because they're not innately superior.

Nothing the Giants did negates DVOA except in the minds of Giant fans, toss a coin four times and it will come up heads four times in a row 6% of the time - so if you see four heads in a row it doesn't means the odds changed from 50/50, you're just witnessing an unusual event.

The fun of sports is you get to witness these unusual events, as much as I dislike the Giants (mostly because NY fans are the biggest ass wipes in God's creation, worse than Cowboy front runners), I enjoyed watching the Patriot smugness being deflated. However, I never confuse the long shot event with a trend, going into 2008, the Giants are still a mediocre team. And Manning is a mediocre QB who has pretty much reached his peak, the fact that he failed to "choke" doesn't suddenly make him a top 10 QB.

I think the primary explanation for the Giant success was integrating Spag's defensive schemes with the right personnel, but before teams had enough film to counter those blitzes. Next season, you'll see a lot of adjustments to exploit those LBs when the Giants bring max pressure.

The NFL is a game of move and counter move, you can gain an advantage in the short run, but unless you actually have the talent in hand, that edge will soon evaporate.

202
by Seepferdchen (not verified) :: Wed, 04/16/2008 - 7:42pm

He threw right to Asante Samuel who then dropped it. That is not good work Eli, that is thank F ing God.

I’m a fan of Eli, but he is streaky. When he’s bad, he is very very bad, not like bad meaning good, but like Alex Smith bad.

Seepferdchen

203
by RoyFlip (not verified) :: Tue, 04/22/2008 - 10:43am

"It’s amusing to read all the Giant fans trying to rewrite reality."

It is equally entertaining to have the Eli and Giant Haters act as if they somehow didn't win. They beat the "Greatest-Team-In-NFL-History". Because Samuel has a pass go through his fingers (as opposed to "dropping it") doesn't change the outcome. Eli took them 83 yards in 2 minutes. Margin of victory doesn't count. You win or you lose. I don't see anyone saying that Eli is now great and the Giants will run off 4-5 championships. The Giants won with a touchdown in a 2 minute drill.

BTW, if you want asswipes for fans, try living between Baltimore and DC, now. No Dancing in the Streets, just purple camo pants and hopelessly living in the past.

204
by masini :: Sun, 02/14/2010 - 12:20pm

When I read this article now remember what we went through then. I hope not to have part of it.vanzari auto dezmembrari masini piese auto ieftine

205
by Glandos (not verified) :: Fri, 11/12/2010 - 10:44am

Great website , i love your css job , i`ve sent you a mail !

best rakeback offers