Writers of Pro Football Prospectus 2008

Most Recent FO Features

BradyTom00-10.jpg

» Week 4 DVOA Ratings

Five different teams from last year's DVOA top eight rank in the bottom half of the league through four weeks of 2014. What can we learn from other teams with similar starts in the past?

06 Oct 2005

Washington Bandwagon Headed for a Crash

by Aaron Schatz

After three weeks of the 1999 season, the New England Patriots were 3-0. In two of those wins, Drew Bledsoe led the Patriots down the field on a game-winning drive during the final two minutes. Though each game was decided by just three points or less, the Patriots showed that they knew how to win when the game was on the line, and were a leading Super Bowl contender.

After three weeks of the 2004 season, the Jacksonville Jaguars were also 3-0. In two of those wins, Byron Leftwich led the Jaguars down the field on a game-winning drive during the final two minutes. Though each game was decided by just three points or less, the Jaguars showed that they knew how to win when the game was on the line, and were a leading Super Bowl contender.

What do these two teams have in common? Each team began its season with three straight close wins that could have easily been losses. And each team failed to make the playoffs. The 2004 Jaguars were 6-7 the rest of the way, while the 1999 Patriots were 5-8.

The 1998 Oakland Raiders and 1999 Miami Dolphins also each won three straight close games during the first half of the season, then each went 2-6 over the second half of the season. That Dolphins team actually backed into the playoffs, only to famously get crushed 62-7 by Jacksonville.

These are the cautionary tales for all those who would quickly jump aboard the bandwagon of the 3-0 Washington Redskins. The storyline is tempting: an experienced head coach and a veteran quarterback who last year each looked like the game had passed them by, seeking redemption with clutch victory after clutch victory. The reality is a deeply flawed team that has narrowly beaten three other flawed teams thanks to some good timing and a lot of luck.

According to Football Outsiders' Defense-adjusted Value Over Average ratings (DVOA) -- which break down each play of the season and compare it to the NFL average based on situation and opponent, explained here -- Washington has been the 22nd best team in the NFL so far this season. They rank 16th in both offense and defense. They are 29th in special teams because they only have positive yardage on three of 16 punt returns and have returned only one kickoff past the 30-yard line.

How has Washington won three games with such mediocre performance?

Washington won its first game against offensively-challenged Chicago, which had a rookie quarterback making his first NFL start.

Dallas outplayed Washington for 56 minutes before 39- and 70-yard touchdown passes to Santana Moss. Long-bomb touchdowns of this type, though great plays that deservedly gave Washington a one-point victory, do not usually indicate that a team will win games in the future.

Washington won its third game in overtime when Seattle's Josh Brown missed a field goal that would have won the game at the end of regulation. In fact, though the league's kickers have hit 81 percent of field goals this year, those kicking against Washington have only hit three-of-six. In overtime, after Seattle outplayed them for 60 minutes, the Redskins exposed Seattle's fatal flaw by converting three straight third-and-long situations to set up the winning field goal.

Despite its 3-0 record, Washington has actually seen a decline from its best unit, the defense. Washington is allowing 5.8 net yards per pass, up from 5.4 last year. The run defense is allowing 3.7 yards per carry, up from 3.3 last year. And while Seattle is a strong offense, Chicago and Dallas most assuredly are not. Washington's defensive DVOA ranked third in 2004, but so far is 16th in 2005.

While the defense has only sacked the opposing quarterback four times, Washington's spotty offensive line has allowed ten sacks. And running back Clinton Portis, who saw a historic drop in his yards per carry average last year, hasn't been that much better this year. He's averaging just 4.1 yards per carry and has only one run longer than 13 yards.

There is one place where the uplifting Washington storyline is really true: Mark Brunell really has played better this year. His completion percentage of 57% and average of 7.0 yards per pass are his best numbers since 2002, and he has thrown only two interceptions. (His DVOA was -22.8% in 2004, but so far in 2005, it is 20.5%, currently 11th in the league.) Of course, that completion percentage is actually lower than last year's 62% completion percentage from Patrick Ramsey, the quarterback Brunell replaced, and the high average of yards per pass will drop as the touchdowns that beat Dallas get smaller and smaller in the rearview mirror.

Meanwhile, Washington's schedule is about to get a lot more difficult. The next two weeks, they go on the road to face two powerful AFC West teams, Denver and Kansas City. They still have two games against defending NFC champion Philadelphia and two games against the newly-powerful offense of the New York Giants, plus games against Tampa Bay and San Diego.

The Redskins may soon learn that clutch performance is a harsh mistress. After starting 3-0 and then beating the expansion Browns, the 1999 Patriots lost their next two games by a combined three points. The 2004 Jaguars, after winning their first three games close, later lost at home to Tennessee by three points and to Pittsburgh by just one point.

With three close wins already in the bag, the Redskins have a reasonable shot at the playoffs. But when their luck evens out -- as it inevitably will -- that difficult schedule also makes them a reasonable candidate for a collapse.

This article appeared in Tuesday's edition of the New York Sun.

Addendum: Since 1996, seven teams have strung together three straight regular season wins of three points or less. Four are mentioned above, and the other three were cut from the Sun article for space:

  • The 1998 Arizona Cardinals were 6-7 going into Week 15 and then won their final three games by a combined eight points over teams with a combined record of 14-34. They managed to beat Dallas in the first round of the playoffs, lost to Minnesota in the second round, and the following year regressed to 6-10.
  • The 2002 Minnesota Vikings were 3-10 going into Week 15, but won their final three games by a combined six points. They went 9-7 the following year.
  • The 1999 Tennessee Titans were 3-1 going into Week 5. They won their next three games by three points apiece, and in the last one handed the undefeated Rams their first loss. Steve McNair, who missed Weeks 2-7, returned for that game, and stayed healthy as the Titans went 7-2 in their last nine games to finish 13-3 and eventually make the Super Bowl.

Washington fans who disagree with the premise of this article are hereby invited to make the case that this year's team has more in common with the 1999 Titans than it does with the other six teams mentioned above.

Second Addendum: Another interesting thing about the Washington Redskins is that our counterparts at Baseball Prospectus spent the season making this exact same argument about the Washington Nationals. For a comparison, read this article written by Joe Sheehan in mid-June. At the time, the Nationals were in first place with a 37-27 record despite being outscored by opponents. The Nationals went 44-54 the rest of the way and finished in last place, albeit in a division where no team had a losing record.

Posted by: Aaron Schatz on 06 Oct 2005

195 comments, Last at 16 Oct 2005, 9:53am by Mykhael

Comments

1
by charles (not verified) :: Thu, 10/06/2005 - 2:34pm

Pete Carroll, Jack Del Rio, Joe Gibbs.
Which is these does not belong?

2
by Chris (not verified) :: Thu, 10/06/2005 - 3:35pm

Re:This Article
Reads exactly like a hit piece -I ususally like the articles here, but this one seems fairly shallow, and as poorly supported as a typical hit piece.

Re:The Argument
According to DVOA they are sixteenth on offense and defense, twenty-ninth on special teams.

It's only been three games, and fluky special teams are a fact of life, so that one doesn't really matter in a discussion of how good a team is at the moment. On Offense and Defense they've done averagely according to DVOA, but they've played against extremely good defenses (Chicago), and otherwise very capable offenses (Seahawks) and the defensive adjustment has been minimal (if 40% changes a VOA of -22.1% into -12.0%, wouldn't a full adjustment change it to about 0%, and given the extremely minimal number of games, 0% could mean a great team that has played great teams [which may very well be the case] that are underrated due to playing against the good teams they lost to [or, in other words, are underrated due to your underranking of Washington, due to your choosing to use defense adjusted value over average as a middleweight, and if they teams they played were rated properly, they'd be far better than 0%}.

So let's look closer at the units. A great halfback. A resurgent old QB who was never great, but looks to be decent this year. A good h-back. An improved o-line. A recieving corps capable of breaking open a game. An underrated Dline. Extremely effective linebackers. An even more effective secondary. Looks like they'll still be great at the end of the year from every perspective that considers where the numbers really ome from, and isn't just used to hating the Redskins (believe it or not, I'm not actually a Skins fan, though that seems to be all I get to post about since so many people are plain wrong about them). (Yes, I am accusing you of bias, since that is the only accusation that makes any sense given that you took the time to write this hit without considering everything).

Seattle did not outplay Washington for most of the game, as Washington simply underperformed and still managed to tie. You failed to mention that at the range Seattle was kicking from, a miss is just as likely as a good field goal. Plus, according to your statistics Seattle is a top ten team, despit the fact you spite one of their four opponents tremendously.

The Bears managed to be number eleven, and it was also a fairly logical win against a team you admit is good.

In the Dallas game, they should have gotten those touchdowns earlier, but they didn't, and still came up with a win. In part, Dallas is spited for playing the Redskins. Since you aren't prepared to admit Washinton is good, that pulls down the Dallas rating considerably. (I am aware you use statistics generated by formula for the DVOA, but the way you are doing it is very unfair toward Washington, and any team that played Washington, and you don't seem prepared to accept that).

History does not control the present, especially when you are only choosing two characteristics to choose that history, and the selections so few.

The only decent argument made in the entire article is that Washington has a brutal schedule, but look again, because they have already proven they can deal with it. If Washinton sweeps the Eagles, WSashington will win the Division, and Phili a wildcard. If they win one and lose one against the Eagles, they will get either Division or Wild Card, and Phili the other playoff spot. If they lose both againbst the Eagles, then the Eagles get division, and homefield advantage, while the Redskins get a wild card spot.

3
by Dave (not verified) :: Thu, 10/06/2005 - 3:46pm

If Washinton sweeps the Eagles...

aaaaaaaaaahahahahahahaha ... if washington sweeps phila ... then they'll ... oh, dear. (wiping tears)

4
by Ken (not verified) :: Thu, 10/06/2005 - 3:48pm

I'm interested to know what the average margin of victory is for teams that make the playoffs - perhaps particularly interesting if only victories are considered?

My personal opinion is that to be a great team, the ability to close out tight matches is highly important. But then again, a great team won't have to rely on that ability too often. Certainly, in any game decided by a score or less, the bounce of a ball can determine the outcome of a game. It's one of the reasons football is such an exciting sport.

My guess is, in reference to the second comment, that reasons of space stopped a fuller analysis. Certainly the conclusion seems to be sound; that you can't rely on clutch performance over a long period of time (although I understand that in a game like football, clutch performance is difficult enough to measure anyway due to small sample size.)

I suppose the one corollary is that constantly winning tight games can give a team a confidence that it might not otherwise possess.

5
by James Gibson (not verified) :: Thu, 10/06/2005 - 3:51pm

A hit piece? I don't think so. I've been feeling this way even without DVOA. Teams just don't continue to win narrow games like that. I predict it starts falling apart this weekend at Denver. So far, they have played - Chicago, who I don't think is a good team, Seattle, which can't ever seem to win on the road, and Dallas, who seems ok not great, and that win had a really fluksih nature to it.

6
by TBW (not verified) :: Thu, 10/06/2005 - 3:57pm

Good article. The Redskins certainly seem like a fraud at this point. Personally, I don't think the Skins will ever be better than mediocre until Dan Snyder sells the team or removes himself completely from running it. Is there a bigger boob in the NFL today than Snyder ?

7
by Pat (not verified) :: Thu, 10/06/2005 - 4:02pm

wouldn’t a full adjustment change it to about 0%

Opponent adjustments are small because if they were large right now, almost every team would be average. Especially the teams that have already had byes, as there's less data.

The Dallas game was the telling game for Washington. They were outplayed the entire game save the last five minutes. Against a team with a secondary that actually knows how to cover receivers, they would've lost.

Hint: there's another team in their division with a top rated secondary.

8
by SJM (not verified) :: Thu, 10/06/2005 - 4:06pm

I've known aside from DVOA that the Skins are not as good as their record, but that doesn't mean they won't have a good season. I'm not going to say that the Skins are like the '99 Titans, but the special teams will get better. Moss will return punts, and hopefully Frost will be a better punter than Groom.

It's true, the defense isn't as good, but there is room for improvement (Sean Taylor will only get better if his shoulder injury is not serious, Marcus Washington has not played as well as last year, Arrington is not contributing at all, Rodgers will also improve).

Finally, Portis looks better than last year, thanks to a better deep passing game, a better designed running game and more relief from Betts. He's not at the level he was in Denver, but he could easily break a lot more long runs.

And I would like to say that the Moss for Coles trade certainly looks pretty good right now (even before the Jets lost their QBs).

9
by Drew (not verified) :: Thu, 10/06/2005 - 4:07pm

Pete Carroll, Jack Del Rio, Joe Gibbs.

Jack Del Rio. The other two are championship coaches.

10
by princeton73 (not verified) :: Thu, 10/06/2005 - 4:09pm

Pete Carroll, Jack Del Rio, Joe Gibbs.
Jack Del Rio. The other two are championship coaches.

the other two also don't keep axes in the locker room

11
by Drew (not verified) :: Thu, 10/06/2005 - 4:16pm

the other two also don’t keep axes in the locker room

So the question then becomes, are they winners because they don't keep axes in the locker room, or do they not keep axes in the locker room because they're winners?

12
by Pat (not verified) :: Thu, 10/06/2005 - 4:20pm

SJM:

The main problem is that their weakness on defense currently doesn't stack up very well against their own division - it's just not a good thing to not have a top-ranked pass defense in the NFC East, especially with the Giants entering the "hi, we can shred you to pieces" as well.

Washington's pass defense right now is looking mediocre at best. Both Seattle and Dallas threw the ball very well, and Washington won both of those games on a missed FG (Dallas missed one on their first drive). In both of those games, it helped tremendously that the other team lacked any attempt at a defense.

To be honest, Washington might even win the next 3 games: Denver's pass offense depends on their rush offense, and Washington can shut that down, and Kansas City's pass defense is just as bad as Seattle's.

But the games after that don't look so pleasant: NYG, PHI, TB, OAK, SD. They'll be lucky to go 1-4 in that stretch.

13
by Aaron (not verified) :: Thu, 10/06/2005 - 4:21pm

The funny thing about the accusation that I am somehow biased against Washington is that I'm the one who spent the off-season arguing that Laveranues Coles really wasn't any better than Santana Moss (click link on my name). Also, I've added an interesting second addendum to the article above. Another link of interest.

14
by Rick "32_Footsteps" Healey (not verified) :: Thu, 10/06/2005 - 4:28pm

Trust me, I'm an Eagles fan. If this was a hit piece, I'd have seen many more cheap shots at Washington (like making the references to the Giants and the Eagles into cruel jokes). Aaron might have his faults, but using an article to grind an axe is certainly not one of them.

While I certainly have my own Philly-based bias in regards to the team, I do think Aaron has a great point - if the vast majority of previous examples point to the result he's predicting, and the current indicators predict a slide, then I think the Redskins and their fans have every right to be nervous about upcoming weeks.

15
by zip (not verified) :: Thu, 10/06/2005 - 4:29pm

Re: #2

That's an interesting view you have of things! Let's break it down:

"On Offense and Defense they’ve done averagely according to DVOA, but they’ve played against extremely good defenses (Chicago), and otherwise very capable offenses (Seahawks) "

Giving one example of a good defense they played and one example of a good offense does not make your statement true. The Seahawks and Dallas D is not "extremely good defenses" and I wouldn't describe Dallas and Chicago as "very capable offenses."

" and isn’t just used to hating the Redskins "

"Yes, I am accusing you of bias, since that is the only accusation that makes any sense"

"Since you aren’t prepared to admit Washinton is good,"

"I am aware you use statistics generated by formula for the DVOA, but the way you are doing it is very unfair toward Washington,"

I never get sick of people on this site accusing people who analyze DVOA of bias.

"So let’s look closer at the units. A great halfback. A resurgent old QB who was never great, but looks to be decent this year. A good h-back. An improved o-line. A recieving corps capable of breaking open a game. An underrated Dline. Extremely effective linebackers. An even more effective secondary.

Using a phrase to describe each unit is not "looking closer," it's called "glossing over." And you mentioned how awesome Portis is twice, even though he wasn't awesome last year and hasn't been this year.

"An underrated Dline"... with 4 sacks in 3 games.

"Seattle did not outplay Washington for most of the game, as Washington simply underperformed and still managed to tie. "

Um, how is one team "underperforming" different than being "outplayed." That's like saying "you didn't beat me, i lost."

"If Washinton sweeps the Eagles, WSashington will win the Division, and Phili a wildcard. If they win one and lose one against the Eagles, they will get either Division or Wild Card, and Phili the other playoff spot. If they lose both againbst the Eagles, then the Eagles get division, and homefield advantage, while the Redskins get a wild card spot. "

Washington still has 11 other games to play not against the Eagles. But obviously they will crush the Giants, and Tampa Bay/Carolina/Atlanta will also not compete for the wild card. Obviously.

16
by mawbrew (not verified) :: Thu, 10/06/2005 - 4:38pm

Washington may make the playoffs. A 3-0 start is a big help in getting there. But the commentary in post 2 reminds me so much of some preseason claims I heard:

1. Pack fans claiming Jim Bates was THE answer for their defense.

2. Viking fans claiming Culpepper wasn't going to miss Randy Moss.

3. Jets fans claiming there was no reason to assume Pennington would miss time due to injury this year.

Except, of course, that post#2 wasn't from a Washington fan.

17
by Leeroy (not verified) :: Thu, 10/06/2005 - 4:44pm

Washington wasn't outplayed for 60 minutes vs Seattle. They dominated the 1st half, Seattle had the ball for less than 10 minutes. You win what you win, but after 3 games, I'd say sample size is a major issue.

18
by princeton73 (not verified) :: Thu, 10/06/2005 - 4:45pm

my reaction to the article was "is this trip really necessary?"

I mean, come on, is there REALLY a Redskins bandwagon? Does ANYONE really think they're an elite team?

Maybe there is; I haven't sensed the groundswell

19
by Pat (not verified) :: Thu, 10/06/2005 - 4:47pm

Keep in mind that when you talk about "winning those tight games" - in the Seattle and Dallas case, they won both of those games because both the Dallas kicker and the Seattle kicker missed makeable field goals. You will lose those games if you play them over and over. A missed (makeable) FG is just good luck.

The Chicago game I could see being consistently repeatable. They only won by 2 points, but the Bears offense was only in the Washington red zone twice, and once was due to a turnover, and once was due to good starting field position due to a penalty.

But the Dallas and Seattle games were won on sheer luck. This doesn't mean that Washington is a bad team. It just means they played three games, won 1, and got lucky on two others. In a stretch of 3 games, that'll work. In a stretch of 13 more? Not so easy.

20
by mawbrew (not verified) :: Thu, 10/06/2005 - 4:55pm

Re: 19

Heck, even the Bears had the ball (first and ten) in (admittedly long) FG range midway through the fourth quarter. A FG then puts the Bears ahead in that game. Of course, the Bears committed three consecutive false starts (that has to at least tie a record, doesn't it?) in route to a 4th and 38 punt.

21
by Leeroy (not verified) :: Thu, 10/06/2005 - 4:56pm

You know the Redskins also missed a FG in the game. And yes kickers have missed three FGs against the skins, but the misses have been from 47, 47 and 41.
Its not like they are missing chip shots.

22
by DK (not verified) :: Thu, 10/06/2005 - 4:57pm

Yes, they are playing over their heads a bit, but this is not a terrible team.

It's easy to look at the three wins and say that they were all luck, but that's not really true.

They completely outplayed the Bears with twice as many total yards, but they had an ill-timed turnover and stalled in the red zone, which made the score closer than it should've been.

They crammed all of their offense into 2 minutes against the Cowboys, but I don't see how that's any worse than scoring those touchdowns 30 minutes apart. In the end, the Redskins had nearly identical game stats against a decent Cowboys team on the road.

It's popular to say that the Skins won the Seahawks game on a missed field goal, but that field goal wouldn't have been so decisive if the Skins hadn't missed a 39 yarder or the Seahawks hadn't made a 53 yarder earlier in the game. Plus, the Sewahawks final field goal attempt came on an interception off of a fluky tipped pass. It's just as easy to say that it's a fluke that the Seahawks even had a chance to win that game.

This team does a lot of the little things well - they rank near the top of the league in time of possession and third/fourth down conversions and near the bottom in penalties.

They probably only have average talent, but they are well-coached, and they're definitely a legit playoff contender with those three wins in the bank in a weak NFC.

23
by zip (not verified) :: Thu, 10/06/2005 - 5:00pm

#16 said what I was trying to say in a much more concise fashion.

24
by mawbrew (not verified) :: Thu, 10/06/2005 - 5:02pm

Re: 22

Is there such a thing as a well-timed turnover (at least when you're the team giving it up)?

25
by MdM (not verified) :: Thu, 10/06/2005 - 5:08pm

Chris, man you are in denial, my friend! Sweep the Eagles?

It seems to me that teams that win by a large margin are better and thus more likely to go to the playoffs than teams whose wins are subject and perhaps owing to random bounces of the ball!

Isn't Carolina an example of this? Except their luck ran an entire season--followed by a season and a half of mediocrity.

We will have to see if Atlanta regresses to the mean.

There may be an argument that some teams are simply designed to become embroiled in, and win, close games. How did the Patriots manage to win 3 Superbowls by 3 points each?

Somehow, I don't think Washington will reach that standard of excellence, and it seems ridiculous to me to call it a "hit piece". Aaron's article seemed sensible to me, and I would guess anyone who's not a Redskins homer.

26
by DavidH (not verified) :: Thu, 10/06/2005 - 5:09pm

In fact, though the league’s kickers have hit 81 percent of field goals this year, those kicking against Washington have only hit three-of-six.

This would be more convincing if you took distance into account and said something like "using average kicking over the past few seasons as a baseline, you would expect Washington's opponents to be X/6 from the distances they've kicked from. Instead they're 3 of 6."

As #21 says, the FG's were not short.

27
by DavidH (not verified) :: Thu, 10/06/2005 - 5:27pm

You know the Redskins also missed a FG in the game.

That's another point against the Skins. The point is that when a team misses a FG, it is their fault, and can be used to try and predict their future kicking performance. When a team's opponent misses a FG, it is usually the opponent's fault, not due to good "FG defense" or something. So you shouldn't expect the opponents to keep missing.

Oh, and sorry about forgetting to close my tag in the last post.

28
by james (not verified) :: Thu, 10/06/2005 - 5:32pm

1. A total sample size of 3 past teams and 3 games? Can a conclusion truly be drawn from this. Any high school teacher would give this report an F just for that.

2. Just as with Michael Vick the Washington D is misunderstood by DVOA. Their success is not predicated on causing turnovers. I suspect that is the problem with the data.

3. Clinton Portis is measured as 10th in success rate which IMO is the only stat of rb measurement that is reliable as far as performance. The other stats are hit or miss with rbs.

4. An offense hitting 50% of 3rd downs is ranked middle of the road. Again DVOA misunderstands the MO of this offense. Move the chains keep the ball. They are making 50% bc they are always in 3rd and short not because they are hitting a ridiculous amount of third and longs. This is exhibited by Brunell's DVOA somehow. I will never understand how a guy throws two perfect lasers to a receiver perfectly in stride and it gets called flukey. Perfect hindsight analysis that dallas should have been in prevent whereas if the skins had methodically moved the ball downfield against a prevent then there would be those who say "Why were they in prevent?".

5. Penalizing WAshington for kicking mishaps when John Hall one of the best legs in the league has been sidelened for over two games is just ridiculous. You figure they get better in that department the moment he steps back on the field. They have missed as many fg's as their opposition.

In conclusion:
There is more wrong with the stats and the method of analysis three games into the season then there is wrong with Washington.

They give up less than 40 points in 3 games(against two teams averaging more than 20 no less) and somehow are a middle of the pack defense. My ass!

29
by Pat (not verified) :: Thu, 10/06/2005 - 5:33pm

#26:

Between 40 and 49 yards, NFL kickers average 70% success so far this season.

The made field goals were 33 yards, 41 yards, and 53 yards.

The missed field goals were 41 yards, 47 yards, and 47 yards.

If you want to make it more dramatic, you could say "NFL kickers average 70% success so far this season between 40 and 49 yards, yet against the Redskins, they're just 1 for 4."

30
by Pat (not verified) :: Thu, 10/06/2005 - 5:37pm

1. A total sample size of 3 past teams and 3 games? Can a conclusion truly be drawn from this. Any high school teacher would give this report an F just for that.

Followed by...

They give up less than 40 points in 3 games(against two teams averaging more than 20 no less) and somehow are a middle of the pack defense. My ass!

I give you an F!

Hint - if you criticize someone because of small sample size, you can't turn around and use that same sample-size limited data to support your own statement.

31
by marc (not verified) :: Thu, 10/06/2005 - 5:52pm

The 1999 Green Bay Packers started 3-1, with three close comeback wins solely because Brett Favre is a badass.
The 1999 Green Bay Packers ended 8-8 solely because Ray Rhodes is a dumbass.
There is a very good reason that the playoffs begin after week 17, and not after week 4.

32
by Pat (not verified) :: Thu, 10/06/2005 - 6:02pm

They have missed as many fg’s as their opposition.

You're wrong.

Nick Novak has had 1 FG blocked this year, and made the other FG he's tried.

Even in the Seattle game alone, Seattle had two missed field goals to Washington's 1 blocked field goal.

33
by Kibbles (not verified) :: Thu, 10/06/2005 - 6:08pm

Guys, DVOA doesn't "misunderstand" teams. Every team's goal is to alter the point differential in their favor. They do this by scoring points, and by not allowing the opponant to score points. DVOA measures how well they're meeting those two goals.

Yes, Washington's defense isn't predicated on causing turnovers. HOWEVER, causing turnovers both kills the drive (stopping opponant from scoring points) and gives their offense the ball in great field position (scoring points). Therefore, it accomplishes the two main goals of a football team. You can never say "Oh, hey, our defense is better without turnovers. We don't need turnovers, our defense isn't based on turnovers." That's like saying "You're underrating our team. Our offense's goal isn't to score TDs, it's to score FGs, and we've scored the third most FGs in the entire NFL, you and DVOA just don't understand what our team is trying to do here!"

34
by JoeGibbsIsSexy (not verified) :: Thu, 10/06/2005 - 6:10pm

"the accusation that I am somehow biased against Washington"
Aaron, you're worse than Pastabelly with your blatant hatred of the Skins!
Now all of us diehard Skins fans that ignore facts and evidence (2 winning seasons in the past 12 years should be enough to convert you blind haters!) need to come up with a derisive nickname for you!
How about Aaron Fatz? Because anyone that doesn't worship the Skins is FAT!

35
by wrmjr (not verified) :: Thu, 10/06/2005 - 6:23pm

#18 makes a good point: is there really a bandwagon for the 'Skins? I'm a Redskins fan and I'm not on the bandwagon. At the beginning of the season, I though 6-10 would be a good record for them. I now think 8-8 is possible, which may be enough to get them into the playoffs in the poor NFC. I don't expect them to go very far.

An article on the Cincinnati bandwagon-- now there's one I'd like to see.

36
by Catfish (not verified) :: Thu, 10/06/2005 - 6:25pm

I didn't see the game, but didn't Chicago return an interception for a TD that got wiped out by a penalty? That would have won it right there.

History does not control the present

This is certainly true, nothing can change the fact that the Redskins are 3-0. What history can do, however, is help gives us a glance at what the future will hold. Right now, history says that the Skins are not likely to continue this success.

I'm a believer that a team's record in blowouts is a much better indicator of team quality than their record in close games. It seems much less likely that a mediocre team will win three straight in blowouts than a mediocre team winning three straight in close games.

37
by Matthew Furtek (not verified) :: Thu, 10/06/2005 - 6:28pm

I hate that football teams get credit for close losses but discounted for close wins. You know at the end of the year the Bears, Cowboys and Seahawks will all be bemoaning their losses to the Redskins and people will say "They are an up and coming team because they played so many close games."

Aaron,
Good article. I'd agree that the sample size is a little small, maybe you shouldn't break DVOA out until week 5 and 6. I'm content to be as happy Redskins fan for now, but like everyone is waiting for the Chiefs and Broncos games to happen before saying anything.

38
by B (not verified) :: Thu, 10/06/2005 - 6:32pm

"but didn’t Chicago return an interception for a TD that got wiped out by a penalty?"
Yea, but without the penalty, Washington probabally (You never know with Ramsey at QB) wouldn't have thrown the interception. It wasn't one of those plays where the penalty ended up not affecting the outcome.

39
by Pat (not verified) :: Thu, 10/06/2005 - 6:34pm

I didn’t see the game, but didn’t Chicago return an interception for a TD that got wiped out by a penalty? That would have won it right there.

Wasn't returned for a touchdown. Was returned for 55 yards. But the penalty was defensive pass interference, so it likely could've been the cause of the interception.

Like I said, I'd be willing to believe that the Chicago game is predictive, because the Redskins offense was far more productive than the Chicago offense that game.

But missed field goals are "points that our defense gave up, but their special teams gave away."

40
by zip (not verified) :: Thu, 10/06/2005 - 6:37pm

An offense hitting 50% of 3rd downs is ranked middle of the road. Again DVOA misunderstands the MO of this offense.

Converting 3rd downs is important for success, however, not settling for FGs (Chicago game) or not getting all your points from two long passes (Dallas game) are also important for long-term success. No one game stat is proof of the effectiveness of an entire unit.

Note: DVOA is not a "game stat"

41
by SJM (not verified) :: Thu, 10/06/2005 - 6:39pm

The criticisms of stat analysis in here are silly, because people are making absurd accusations like "the rating system misunderstands the team." If the Skins aren't producing turnovers, that's bad no matter how you look at it.

However. There is one problem with stat analyis early in the season, and I'm not talking about small sample size. The methods used here do not take into consideration the possibility that teams can get better or worse over the course of the season. This is why WEIGHTED DVOA exists. If the Skins are an average team right now with a really good record, they could stay average, they could get better, or they could get worse. But their good record gives them a leg up on the competition while each team tries to improve and jockey for the playoffs.

Let me name you two people who don't think the Skins are as good as their record: Joe Gibbs and Gregg Williams. They are working their asses off to make sure the team doesn't need lucky breaks to win an doesn't have to win by slim margins.

Let's not go overboard on DVOA. Each team right now has the potential to get much better or much worse. The Skins just happen to have an early advantage.

42
by EorrFU Skins Fan (not verified) :: Thu, 10/06/2005 - 6:40pm

As a fan of he Skins I have wached every snap this season. I agree with this asessment of the team. Certainly they have some holes in the defence this year. Their Secondary is a shadow of the Smoot and Bailey days. Yet, Shawn Springs, Carlos Radgers, and the cariest SOB in the league Sean Taylor are all going to contribute. Ultimitely the Offansive line is a sieve except the gaurds. The backs are good, the recievers are better, the LB's are good (if not as good as hoped) and the D-line is under-reated (they don't get penetration but are good run stoppers)
I think they will be lucky o pull out 2 wins in the next 6 weeks. Denver is a better team but a bad matchup for the skins. They can't exploit the weaknesses of the Skin's that many other teams can. Yet, the Denver D will be a hard nut to crack for that skins offense. If portis can run up the middle the skins can win. But that is a big if. The real key may be how they can screw up Plummer's timing with their blitzing. The snake does not inspire confidence against any exoctic scheme.

If they stay lucky they have a chance against the Broncs and the Bucs. Based solely on the matchups not on being better teams. Otherwise I think the 49's are a win and the Chiefs, Giants, and especially the Eagles are all teams that can exploit the RedsKins faults. If they end better than 8-8 the season must be seen as a astounding succcess.

43
by james (not verified) :: Thu, 10/06/2005 - 6:40pm

Oh god Pat,

Its showing divergance. Which data is right? I suspect that teams that give up 15 points a game for a season and 3 games don't suddenly start giving up 20 like its their job. I suspect that DVOA will catch up with the success the Skins are having. That was my point.

Good day sir

44
by putnamp (not verified) :: Thu, 10/06/2005 - 6:42pm

One place sample size is a major factor is the fact that the Redskins went 13-18 against the Seahawks on third down, and something like a ridiculous 8-9 on 3rd-and-long. That's going to lead to a bloated 3rd down conversion rate, I would bet.

Take it from a Seahawks fan, guys, going 3-0 to start the season doesn't mean a damn thing. We've done it several times in the last 4 or 5 years, and it has netted us nothing.

45
by B (not verified) :: Thu, 10/06/2005 - 6:46pm

Re #43: And I suspect teams that score 15 points a game don't suddenly start scoring 20.

45
by Jerry P. (not verified) :: Thu, 10/06/2005 - 6:46pm

"A total sample size of 3 past teams and 3 games? Can a conclusion truly be drawn from this. Any high school teacher would give this report an F just for that."

Aaron noted four other examples at the end of the article that did not appear in the New York Sun version of the article. Maybe you missed it or it was added later by Aaron, but there is definitely more than three examples. He even challenges Redskins supporters to describe how they are like the one successful example, Tennessee. No one has yet.

47
by james (not verified) :: Thu, 10/06/2005 - 6:47pm

kibbles no. 33,

You missed my point. Just as DVOA has a hard time with Vick there has to be some other units/players who aren't correctly measured. I believe that Washington is one of them. Don't know how to make it better. Turnovers were a guess.

48
by Jerry P. (not verified) :: Thu, 10/06/2005 - 6:49pm

Sorry, Aaron noted three other examples in the addendum. There were actually four in the article. Still, seven total.

49
by Trogdor (not verified) :: Thu, 10/06/2005 - 6:53pm

How does DVOA have a hard time with Vick? It has him pegged as an incredible runner and below-average passer. Is there any reason to think differently?

50
by james (not verified) :: Thu, 10/06/2005 - 6:55pm

pat,

I agree noone said that. Not part of the argument. Noone is saying the skins O is going to suddenly be great.

If you go back and read I state that Skins D isn't going to start giving up 20 like its their job.

You have nothing to say for or against that statement. Your comment shouldn't start with re. You didnt reply to anything.

51
by james (not verified) :: Thu, 10/06/2005 - 6:58pm

re:chicago game

Quick to mention the intercepiton that got called back yet never mention that a fumble caused by a clothesline to the qb's head in the red zone was held up. Those were potential points the skins were robbed of.

re: Dallas game
noone is mentioning that Dallas got a long bomb on a trick play as if that is conventional or would work 100% of the time. The deficity should have been 6 not 13 according to your the logic that you have to consider what would happen if the fgs were made.

52
by Kibbles (not verified) :: Thu, 10/06/2005 - 7:06pm

Re #47: You think Washington is a great defense, DVOA disagrees, so suddenly Washington is one of those teams that DVOA can't accurately measure?

First off, why exactly can't DVOA accurately measure a player or a unit again? Even Michael Vick. I think DVOA does a great job of measuring Vick. It analyzes how good his passing is. It analyzes how good his running is. It can analyze how good the entire team is with him, and how good it is without him. I fail to see how DVOA fails to capture Michael Vick.

Second, why exactly does DVOA underrate Washington? I mean, you disagree with DVOA, so you think DVOA has to be wrong in this instance. Give me support. Give me a reason why you think DVOA is wrong. If you're going to accuse DVOA of being wrong, give a reason why you think that, rather than just saying "I think this, DVOA thinks that, so DVOA obviously is missing something".

I think we can all agree that homers of a particular team are going to tend to think that our team is better than it really is when it's playing well. I know I'm guilty of it, too. In fact, I'm pretty sure I pioneered the whole "DVOA underrated our defense because there's too much emphasis on turnovers" arguement about Denver. I argued that a stop on 3rd down was as good as a turnover, and Denver was the best in the league at getting stops on 3rd down. I argued that the offense was designed to sustain long drives, and the lack of turnovers wasn't as big of a problem as DVOA made it out to be. And then I actually ran some numbers and looked at Denver's scoring offense and defense when they got turnovers compared to their scoring offense and defense on drives where they didn't, and I came to the conclusion that turnovers really are as important as DVOA makes them sound.

53
by James Gibson (not verified) :: Thu, 10/06/2005 - 7:07pm

Why shouldn't teams both get credit for close losses and have their wins devalued for close wins? History in all sports suggests that teams that have W/L records significantly different from their Pythagorean W/L records will spring the next season toward their pythagorean projections. I beleive this trends true during a season as well. The web archive seems to be down right now or else I'd show you one of the most ridiculous cases of this I've ever season - the Colts in the early '90s, which I wrote about on my old football analysis web page.

54
by John Gach (not verified) :: Thu, 10/06/2005 - 7:07pm

Anent Washington's mediocrity -- which in this case seems the exactly correct term -- see Sagarin's ratings at http://www.kiva.net/~jsagarin/sports/nflsend.htm. Sagarin spots the same thing that FO does. Washington technically ranks 12th in his current rankings, but that masks a large disparity between the Elo Chess system, which counts only winning & losing, and Pure Points, which counts winning margin and has much greater predictive power. The Redskins are 7th in Elo and 16th -- == precisely mediocre -- in Pure Points. Doesn't make them a poor team like hapless Houston & SF. Does suggest, though, that they're not an elite team, for just the reasons that Mr. Schatz wrote about.

Interestingly the Eagles, Patriots, Broncos, and Bengals all have Elo & Pure Points rankings that differ little. Indianapolis, on the other hand, has a spread like Washington's, which I strongly suspect means the opposite of what it means for Washington, viz., that Indianapolis's offense is likely to play the rest of the season more as it did against Tennessee than as it did against Cleveland and that its winning margin will likely increase (and thus the gap between Elo & Pure Points will diminish).

Sagarin's rankings look about right to me, while FO's seem a bit wacky at the top: Giants #2, Indy #6, & Miami #7. Obviously this will change as more games are played and the quality teams play well. But why is the bottom of the FO DVOA rankings about right, save just possibly for the Ravens (if their offensive line can remember that they once knew how to block)?

Speaking of the mysteriously awful Ravens (though FO did predict a decline in both offense & defense, if I recall correctly): close examination of the current DVOA stats identifies the team's two main problems -- and neither is the quarterback.

1) The Oline has been the worst in the league in adjusted line yards, averaging a horrible 2.77 yards per play. Interestingly, the Ravens's Oline is #4 in power plays, which suggests that so far it really has been good for about 2 1/2 yards a play, whether the team needs 10 or 2(though one shouldn't overlook its #26 ranking in getting stuffed and #20 in sacks allowed).

2) Jamal Lewis has been seriously awful with a DVOA of -49.3%, the lowest rated back in the rankings. Chester Taylor's -12.2% DVOA does suggest, however, that much of the Ravens's ineptitude running stems from the poor play of the offensive line. This week's game against the Lions will likely foretell whether the Ravens have any chance at the playoffs. If they can't run successfully against Detroit (as they couldn't against the Jets, although they certainly kept trying), then they can't run, which means they can't pass, which means the defense will continue to stay on the field and have to play exhausted in the second half of games, which means they will lose a lot of games, especially once they get past the current pattycake part of their schedule (Titans, Jets, Lions, Browns, Bears).

55
by Theo (not verified) :: Thu, 10/06/2005 - 7:11pm

According to history, the Washington Redskins are going 5-0 and then 5-5.

56
by charles (not verified) :: Thu, 10/06/2005 - 7:13pm

Nobody on FO has a bias against the redskins. If you click on my name Aaron has a some info on the 2001 redskins who became dvoa darlings in the second half of the season but missed the playoffs because of some bad breaks. The 2005 redskins are a bad dvoa that has caught good breaks all year. What i need is the redskins to have a good dvoa and catch some lucky breaks. That might mean a deep run in the playoffs. But i'm happy they beat dallas, i really want them to play jason campbell now so they can go to the next level. BTW, the giants being no. 2 in dvoa is a joke. And the redskins will beat the eagles once this year. more on that later.

57
by Richie (not verified) :: Thu, 10/06/2005 - 7:15pm

That Dolphins team actually backed into the playoffs, only to famously get crushed 62-7 by Jacksonville.

Not before winning at Seattle, for their first playoff win on the road since 1972.

58
by ABW (not verified) :: Thu, 10/06/2005 - 7:17pm

What is up with this "DVOA doesn't know how to rate Vick theory"? It's not that DVOA doesn't know how to rate him, it's that he's NOT A VERY GOOD PASSER and DVOA shows that(as does every other statistic). He is a damn good runner and that shows up in his very good rushing DPAR and DVOA.

DVOA rates Vick right where it should. Atlanta wins because of it's running game and D. And before you even say it, I'm sure teams end up weakening their run D to account for Vick's running ability sometimes too - and, well gee golly gosh, makes perfect sense since he's clearly the best running QB in the league by DVOA and every other analysis. I'd probably game-plan for the best running QB in the league too if I was a coach.

No statistical analysis, DVOA, yards, TDs, passer rating, whatever, will ever rate Vick as a good passer, because he's NOT. This does not mean that the statistics are broken - it means Vick does not throw a football particularly well.

59
by Catfish (not verified) :: Thu, 10/06/2005 - 7:50pm

Quick to mention the intercepiton that got called back yet never mention that a fumble caused by a clothesline to the qb’s head in the red zone was held up. Those were potential points the skins were robbed of.

My apologies, I didn't realize that. Like I said, I didn't see the game. Also, the Chicago Tribune made it sound like the penalty didn't have anything to do with the interception, hence the benching of Mike Green.

60
by Ryan (not verified) :: Thu, 10/06/2005 - 8:00pm

RE: #46
First of all I am a biased Redskins fan lets just get that out of the way up front.

But I do believe that a "case" could be made like you asked to compare this Redskins team to one of the successfull teams that Aaron later added to his list. Of those teams you could argue (albeit thinly) that the resurgence of Mark Brunell could be similar to the return of Steve McNair (no I am not arguing that Brunell = McNair in terms of comparing QB's). One thing noticed almost universally by those who have watched the Redskins and Brunell specifically all note that he looks like a completely different QB this year than he did last year. Although he won't admit that the injuries were bothering him the performance on the field sure seems to suggest that there was something seriously wrong with him last year.

Now I don't expect the Redskins offense to suddenly turn into an Indy/StL/NYG offense, but even "mediocre" play from their offense would be a huge improvement from the "horrid" rating they got last year.

Do I think the Skins are legitimate superbowl contenders, no but I do believe that like many other posters here have stated, having a 3-0 record after 3 weeks improves your chances of going to playoffs quite a bit more than having an 0-3 record.

61
by james (not verified) :: Thu, 10/06/2005 - 8:16pm

Lets all agree to disagree about the skins when looking DVOA and look at a different set of data that correlates just as well to winning and losing. This data set is the Drive Stats found on football outsiders.

When you look at this data you get a much clearer picture of why the Skins are winning games with a much simpler means of measurement when compared to DVOA.

In my opinion the most important measurements of team performance are yds/dr, dsr, and pts/dr in reverse order of importance.

When we analyze the skins this way we see a much different picture.

1. Defensively they are ranked in the top 10 in every category. 9-yds/dr 6-pts/dr 6-dsr. Bolstering this early season data is the correlating data of last years defense which had ranks of 3,1, and 2 respectively. Each measurement is within 5-6 ranks of last year which MAY be assumed to make this years numbers stand up a little more.

2. Offensively the team is showing an ability to move the ball. They are 5th-yds/dr, but 21 pts/dr, and 14 in dsr. As compared to last year when they were 29,31,32 respectively. These numbers lead me to believe that the offense is on the rise as they have easily moved the ball all 3 games(over 300 yds each game). Also I considered, that the Ramsey's play for game 1 is included in the data to skew it a little more to the down side.

Every team with a top 10 scoring defense (measured in pt/dr) last year had a .500 or better record with the exception of the only two that did not have a top half(15 and above) scoring offense(TB and WAS).

There was less of a correlation between top 10 scoring offense and winning percentage as Minnesot,KC, GB, and Carolina finished .500 or worse.

From last year's data, the biggest difference in terms of rank between yds/dr and pts/dr offensively was 11.

Now consider that redskins have the 5th rated offense in yds/dr which with even with the extreme derivative data point of 11 ranks away from their current yds/dr rank of 5 would place their scoring offense in the top half(barely). If you used the median and mean, and mode difference 3,2.96,3 then you are much more likely to see a top 10 scoring offense from the skins which puts them in elite company according to last years stats.(only New England, Philadelphia, and Pittsburgh achieved that distinction last year)

Also consider that their offense stands a good chance of staying at that rank after having faced two opponents who rank in the top half of defense in yds/pt(13 and 4th ranked teams). The average defensive rank of opp face in yds/dr is 13.3. The average rank of their remaining opponents is 18. If I take the seemingly logical step of assuming that against lesser competition they will perform better than 5 seems to be the least of what they will end up measuring this season. I will continue to use 5 however.

IMO what this points to is a great improvement in the overall quality of the team.

The skins with a top 10 scoring defense measured in pts/dr, and a top half scoring offense measured in pts/dr should finish no worse than 9-7, if last year's data was significant.

If you believe the Skins can finish around the same area in yds/dr(avg rank of remaining schedule is 18) then its no strecth that they at least finish in the top half and if they behave according to most teams then they will be in the top 10 in pts/dr.

Without counting the Chicago game(valid excercise since Brunell did not start or practice to start) the redskins have had 19 drives and 34 points which is yds/pt of 1.78 which would place them 13th in the league as of this week.

This analysis points more towards a playoff run for the skins then just looking at DVOA.

Thanks for reading.

62
by JonL (not verified) :: Thu, 10/06/2005 - 8:17pm

This discussion is all well and good, but...is anyone outside of Washington jumping on the team bandwagon?

63
by Pat (not verified) :: Thu, 10/06/2005 - 8:19pm

James (#43):

Um, was this comment supposed to be coherent? Because it... wasn't.

Was it regarding the "you get an F" comment? That was because you criticized Aaron for making a conclusion based on 3 games. You then went and made a conclusion (Redskins have a good defense) based on 3 games. If you're going to be of the opinion that 3 games is too small a sample, fine, but you need to be consistent with that.

Extending the data to last year seems a little questionable, given the amount of turnover on the team.

Especially because the Redskins didn't stop the Seahawks from scoring more than 20 points. The Seahawks did that all on their own. If your defense allows another team into field goal range, and they don't convert, that's luck, not quality defense.

64
by Jerry P. (not verified) :: Thu, 10/06/2005 - 8:26pm

"But I do believe that a “case� could be made like you asked to compare this Redskins team to one of the successfull teams that Aaron later added to his list. Of those teams you could argue (albeit thinly) that the resurgence of Mark Brunell could be similar to the return of Steve McNair."

To be clear, it was Aaron that invited people to compare the Redskins to the '99 Titans. Not a successful team he listed, the only successful* team of the seven listed.

Your entire comparison seems to be "McNair coming back from injury in the 1999 season is similar to Brunell coming back from injuries last year". You also admit that it's a very thin case. You make no other comparisons. But technically, you have offered the only comparison so it is the best.

If that's the best that's been offered up, doesn't that go to Aaron's assertion that the Redskins will most likely end the season how most of those teams did and not be dominant like the Titans were?

*Successful means top tier, like some believe the Redskins are instead of just good enough to get a Wild Card spot.

65
by Sid (not verified) :: Thu, 10/06/2005 - 8:35pm

Washington is a fraud. They'll probably still finish 7-9.

66
by james (not verified) :: Thu, 10/06/2005 - 9:14pm

Pat,
Seattle only scored 17. What are you talking about.

Read number 61 to see where I am coming from.

I meant no disrespect to Aaron by making the F comment. It was meant to point out that 3 teams and 3 games are not enough to make a valid conclusion. It was not a knock on Aaron he is obviously the man.

67
by Richie (not verified) :: Thu, 10/06/2005 - 9:45pm

Seattle only scored 17. What are you talking about.

His point is that Seattle would have scored 20 had they made their field goal. The fact that they missed the field goal had more to do with Seattle's failure than with Washington's success.

68
by Charles St. B (not verified) :: Thu, 10/06/2005 - 10:23pm

Your skins analysis is simply flawed. There is simply no way you can assume this team is going to crash and burn through the rest of this season. Sure all of the skin’s games have been close this year: which in fact is a continuation of last year. Last year 11 of 16 games were won or lost by 1 touchdown or less and 13 of 16 games were won or lost by 10 or less points. And while Washington has needed a Quarterback for a long time, nothing is wrong with our defense. In fact, forget DVOA, the skins defense has been one of the best, in term of points and yards allowed since last year. Last year even the best teams Washington played had trouble running and scoring on the skins. You talked about being lucky, well that what it is about sometime. It seems luck was against Washington last year, the way we were edge out in almost every game. If it was not for many turnovers and penalties we could easily have won 11 games. The way the refs were calling against Washington last year, you would have thought they were out to get the skin (and they were called at the worst time).

And you are definitely wrong about the Seattle game. You admit that Seattle is offensively very good. However they certainly did not outplay Washington in that game. The skins completely contain them for three quarter. Seattle was not able to run the ball effectively until the forth quarter. All afternoon the skin had no problem moving the ball from one end of the field to the other; our third down efficiency was great; I think we had the ball for over 21 minute in the first half. We were running and throwing well against Seattle, Washington’s problem was turnover and not being able to score in the red zone, otherwise we would have put them away early.

Last year I would have said there is no way the skins can come back on teams like this, However, maybe the they are just doing what really good team do, find a way to win !!!

69
by NF (not verified) :: Thu, 10/06/2005 - 10:32pm

#28:

"I will never understand how a guy throws two perfect lasers to a receiver perfectly in stride and it gets called flukey. Perfect hindsight analysis that dallas should have been in prevent whereas if the skins had methodically moved the ball downfield against a prevent then there would be those who say “Why were they in prevent?"."

About the first "perfect laser," it's possible that it was a bad pass that a reciever without the top speed and turning ability of Santana Moss would not have caught. The pass was thrown as Moss was going straight near the right sideline, but went towards the middle of the endzone, so to catch it Moss had to slant towards the middle of the endzone at top speed with two or three defenders behind him. This may have been the intended play, but I'm not sure that Moss didn't just make a heads-up play by having the awareness to change his route to the location the ball did go to. On the second pass he did catch the ball in stride running down the right sideline.

Also, TMQ had an item about how, even disregarding what play was called, there was a man-coverage breakdown on both touchdown passes, with the cornerback who was supposed to cover Moss letting him get past then trying and failing to catch-up with him.

70
by thad (not verified) :: Thu, 10/06/2005 - 11:15pm

re 28
Are you or real? You really do not understand why those two passes were called flukey? Did you see Brunell play last year? He sucked. He especially sucked on deep throws. And finally, mercifully, for those of us that live in DC, he got benched. Now those two throws were very good. But i never thought Brunell could make them.

71
by Goldbach (not verified) :: Thu, 10/06/2005 - 11:16pm

After reading this discussion, I'm convinced that Aaron is biased. He's biased against mediocre teams.

72
by Fnor (not verified) :: Thu, 10/06/2005 - 11:45pm

As my wife would say when playing Ogre Battle... "Let's look at the map!" (drive chart).

Normally I would agree with the time of possession argument. However, it fails to address that the majority of that big time (14 minutes of it) was spent on drives that produced 0 points. It was the first quarter, so they weren't exactly running the clock down. So what exactly were they doing, oh-so-slowly marching the ball against a below-average defense?

Let's have a look at how many 3rd downs the redskins accumulated and how many of them were 3rd and long.
3rd and short: (1-5) 5
3rd and long: (6-10) 9

Okay, so they didn't do so great on downs 1 and 2. In those first two time-consuming drives, WAS converted on 4 3rd downs and had a fifth handed to them by DPI. So, we have 177 yards of total offense, 5 3rd and longs, and 7 points and a complete meltdown right at the red zone. How is that finding a way to win? It just looks like dead clock and unimpressive yardage against a bad defense, coupled with a whole lot of lucky breaks to me. If this were a Manning or a Brady, I might just buy it, but not from Bulger. In the seattle game, at least, they were relying to the nth degree on Bulger Favresquely getting them out of holes. He's not going to be able to come close to keeping that up.

And you can't say "ignore DVOA" because, even with a small sample size, stats are much better at this sort of thing than us making assertions about teams is. The fact of the matter is that when you have a 3rd and 9, a 3rd and 12 and two 3rd and 13s against a bad defense, you're in a heap of trouble against good or even average ones, because the magic first down fairy isn't going to keep showing up.

73
by David (not verified) :: Thu, 10/06/2005 - 11:48pm

Fnor - You mean Brunell, not Bulger, right? Otherwise I'm very confused.

And...wife...playing...Ogre Battle? Your words are strange and disturbing to me.

74
by Fnor (not verified) :: Thu, 10/06/2005 - 11:53pm

Yeah, Brunnel. I'm really tired....

Ogre Battle is a tactical RPG series. You get your little army and beat the crap out of other little armies. She's a tactical genius, so she loves that sort of stuff. Carries over into football and chess. Anyway, in the beginning of each fight in the Nintendo 64 version the advisor would meet you in the briefing room and it would zoom in on the map, even though he never said anything useful, so "looking at the map" to us is looking at a less-than-completely-useful representation of the data and trying to come up with some meaning from it.

Guess I should've explained the injoke. Then again, the explanation was almost as long as the post....

75
by BillinNYC (not verified) :: Thu, 10/06/2005 - 11:57pm

I can't believe no one has gone after MY favorite line, "In the Dallas game, they should have gotten those touchdowns earlier, but they didn’t, and still came up with a win."

So we're good, but we couldn't get TDs that we should have. Hmmmm.

And I couldn't care less about WAS.

76
by Josh, FL (not verified) :: Fri, 10/07/2005 - 12:03am

RE: 2

I disagree with a few things. First of all, yes Portis is a great back, but he doesn't fit the system at all. He's not the kinda guy you should give the ball to 30 times a game. He'll wear down bigtime. He's more of a cutback runner, which is why he was so successful in Denver. If Dan Snyder knew anything about football, he would have foreseen this and wouldn't have shelled out huge bucks for Portis.

Mark Brunell once was one of the better quarterback in the leaugue, but after having taken so many hits throught out his career, I just don't think he's got much left in the tank. Yes he has played fairly decently as of late, but at the same time, he hasn't played a great defense yet. The Bears have a good defense, but theire lack of offense reall hurt them, and Roy Williams allows a coupld of huge plays every game so he was due to have some late in the game, and I don't believe they have a great defense.

What does a great H-back have to do with winning 3 games? Not much.

Other than Chris Samuel and Jon Jansen, who are both pretty good, their line is ok.

Santana Moss is about the only WR capable of breaking open the game, and once defenses start double covering him every play, the barely existent offense will no longer be in existence.

I think on defense it's more what Gregg Williams has done than what players they have. And since they don't have a pass rush except for when they blitz, than later on in the season after offensive lines have jelled together a little more and will be more capable of picking up the blitz effectively, than weaknesses in the secondary will start to show up.

I'm kinda thinkin that teh Redskins' 3-0 run is a fluke and they'll be lucky to get past .500, and I also hate to say it, but I think Joe Gibbs shoulda stuck with NASCAR. He's a great guy and all, but today's game is just too different from what it was when he coached.

77
by BillinNYC (not verified) :: Fri, 10/07/2005 - 12:14am

Oh, and I love the DVOA misunderstands x, y or z. Its math. And not even all that complicated math. As it says, its Value over Average.

(as copied from Stats Explained)
How has this person/team done versus the average of all others in similar circumstances based on down, distance, time, etc. Rushing plays are compared to other rushing plays, passing plays to other passing plays, tight ends get compared to tight ends and wideouts to wideouts. If Player A gains three yards under a set of circumstances where the average NFL running back gains only one yard, it can be argued that Player A has a certain amount of value above others at his position. Likewise, if Player B gains three yards on a play where, under similar circumstances, an average NFL back would be expected to gain four yards, it can be argued that Player B has negative value relative to others at his position. If you divide a player's total success value by the average success values of all players in the each of the situations he faced, you get VOA, or Value Over Average.

I recall way back in High School, I got a D on an algebra test. I should have just argued that the teacher just misunderstood my answers!

78
by Supreme Wu (not verified) :: Fri, 10/07/2005 - 12:24am

"In overtime, after Seattle outplayed them for 60 minutes"

The writer clearly did not watch that game. Not even the insanely hardcore Seattle fans think they outplayed anyone that first half.

79
by james (not verified) :: Fri, 10/07/2005 - 12:51am

nf,

The qb will throw the ball like that on purpose because it puts the db at a disadvantage. His hips are turned and the reciever has every chance to advance. It was a great throw by Brunell. If Brady throws that there would be 5 articles about how great a throw it was.

Is Brunell Brady? Hell no! But that was the best pass I've seen all year.

For everone else,
Read 61, its long, but illustrated my points much better than my earlier "non level headed" argument about how dvoa "misunderstands". That is not what I meant to be taken out of my argument. What I meant was for everyone to not look at DVOA as the lord savior and last word. There are other ways to measure good/bad teams such as Drive Stats which show a different picture as I tried to illustrate.

80
by David (not verified) :: Fri, 10/07/2005 - 12:55am

Fnor -

I do know the Ogre Battle series, although it's mainly through Tactics Ogre, which I don't like much. It's just that you don't run into guys whose wives play strategy/RPGs much, especially when you're talking football.

81
by Fnor (not verified) :: Fri, 10/07/2005 - 1:00am

Your argument, however, is placed squarely on the shoulders of yards per drive. Sure, that's great, but that can be inflated by things that are not actually in line with the team's performance. See my argument about third downs. If they can keep converting 3rd and 10+, sure, they're going to be on the rise. But they won't. No one can keep that up.

82
by james (not verified) :: Fri, 10/07/2005 - 1:09am

My argument relies mostly on points per drive using yards per drives as another measurement which usually has a certain correlation.

They have been playing above average teams in the yards/dr department, and will now be playing mostly below average teams.

Why would they perform worse against below average teams? Because the great passes that Brunell is throwing are suddenly going to stop.

Brunell was a good qb for years in the league. He has one bad year, probably because of injury, and is being judged on 8 games and not over 100.

Yes he is older but his arm is not. He is throwing great passes. I hope you are watching the games and commenting and not just looking at the stats and commenting.

The skins offense leaves alot to be desired, but it is hardly been underperforming. They are doing exactly what Gibbs asks. Take care of the football, get in third and short.

83
by Sergio (not verified) :: Fri, 10/07/2005 - 1:14am

Aaron,

You really like it when you mention Miami's humilliating defeat against the Jags, huh???

;)

84
by Fnor (not verified) :: Fri, 10/07/2005 - 1:48am

But they're not getting 3rd and short. In fact, they're getting third and short exactly as often as they're getting 3rd and 10+. The drives are staying alive because out of those 19 3rd and 10+s, they've converted on an astounding 12. That's 52.5%. Interestingly enough, that's the exact same percentage they have on 3rd and short. On short that mostly makes sense. But think about it: through three games, this year's redskins have a higher 3rd down efficiency on 3rd and 10+ than last year's most efficient team had with all 3rd downs, 3% higher than Minnesota's total. The league average was 37%, about 15% lower than the skins' mark this year.

And yes, I've watched two of the three games. I've been thoroughly unimpressed.

85
by james (not verified) :: Fri, 10/07/2005 - 2:02am

Not saying you should be impressed. The skins are a 8-8 shell that does the little things right to be better.
That is my point. They are winning by just being themselves. Hopefully they get better but they are doing things the right way now.

Instead of saying Joe Gibbs is calling good plays( evidenced by wide open recievers every third down) or Brunell is getting the job done by scrambling twice in crunch time for long third down conversions, everyone says its luck. Why are those runs so easy? Because they've been converting the whole game. Sorry to inform you but they are getting the job done and you guys are just hating.

"They can't keep completing 52% on third down noone is that good"

Of the three teams that aaron reference how many were 52% on third down and had as many yards as Washington. I'm not sure what the answer is and it is an honest question.

They earned every single victory though if replayed them might be 0-3, 1-2, or 2-1. The point is they put themselves in the situation to be able to win because of their defense which has allowed 3 touchdowns all year.

They took their last two wins, with the offense(which supposedly sucked so bad, moving the ball easily when it counted). Drove on Seattle in OT, Seattle never touched the ball. But, yeah their offense sucks, sure.

Two long bombs on Dallas in the 4th quarter. Yeah that got lucky against that all pro safety two times in a row. That Moss-Brunell combo is no good.

You want to talk about an undefeated team that is fraudelent. You need to look in Florida.

86
by DavidH (not verified) :: Fri, 10/07/2005 - 2:14am

From last year’s data, the biggest difference in terms of rank between yds/dr and pts/dr offensively was 11.

Now consider that redskins have the 5th rated offense in yds/dr which with even with the extreme derivative data point of 11 ranks away from their current yds/dr rank of 5 would place their scoring offense in the top half(barely).

Umm... if you are trying to bring yds/dr and pts/dr together, you should move them towards eachother, not keep one fixed and move the other one. Why would you think that their ranking in yds/dr is more accurate than their ranking in pts/dr? 5th and 21st becomes 8th and 18th that way. (Or 12th and 14th if you want to use the mean/median numbers you quoted.)

Also, I like that you
A) said pts/dr is the most important number
B) told us the Skins were ranked 21st
C) claimed they have a good offense

87
by james (not verified) :: Fri, 10/07/2005 - 2:26am

They've had 300 plus yards every game. That is good.

Why should the yards stop coming. That is what they are doing well and against better than average competition.

You could look at it the stats whichever way you choose. If you think that yardage will come down even though they will be facing lesser competition than they ahve already faced then you are welcomed to do so.

I choose to think that they will begin to score more points to match their yardage effort.

We will see who was right at the end of the season.

Everything is for arguments sake and is nothing personal.

I picked the Skins as a playoff team at the beginning of the season and they are doing almost exactly what I thought they would. Win with defense,not make mistakes on offense.

You can disagree and write a convincing argument. You do that well.

But, when it comes down to it, the proof will be in the pudding.

I guess we shall both see.

88
by james (not verified) :: Fri, 10/07/2005 - 2:49am

What is funny is that this site was created to show an alternate side to conventional wisdom. It was created to show what conventional wisdom (avg. yds by running back, qb rating, etc) could not.

Now those tools have become conventional wisdom themselves with too many people "afraid" to look beyond those stats for the next level of thinking.

Just like the "old timers" who are pay too much attention to qb rating maybe, just maybe some on this site are too quick to use the new conventional wisdom to explain what they haven't seen before as an anomoly.

My point is unless you are willing to think differently then you ever have before you should not be on this site.
What sense does it make to condemn conventional wisdom by utilizing DVOA and then let DVOA become conventional wisdom itself.

Don't tell me that the skins are lucky to convert on third and long 52% of the time. Give me a play by play description of why that play shouldn't have worked. Did a defender miss an assigment. Did a receiver make a once in a lifetime catch. Were 10 tackles missed on one guy for a lucky play. Did a cornerback fall down. Did a lineman miss a sure sack. Until you can do that you are just reverting back to conventional wisdom. Call me stupid but this site is for those who don't believe in conventional wisdom.

What you all aren't realizing is that you are watching the man who once reinvented offense(Gibbie) MAYBE doing it again. Do you think he is somehow not creative enough to do so? We saw another guy come back and do the same. Dick Vermeil. But then again conventional wisdom once said that guy could never turn around a sorry Rams franchise that was going nowhere.

See where I'm going.

89
by A A (not verified) :: Fri, 10/07/2005 - 3:12am

A few quick notes:

1. To say that Seattle outplayed Washington for 60 minutes is ridiculous. Anybody who watched that game knows the Redskins were absolutely dominant for the first 3 quarters but unfortunately let Seattle back in the game with two late, long touchdown drives.

2. The Redskins have not only won 3 consecutive games by 3 points or less, they've also done it WHILE LOSING THE TURNOVER BATTLE IN ALL THREE GAMES. They are -4 on turnovers this season and that is a primary reason why all their games have appeared to be so close. In actuality, the Bears game was nowhere near as the final score indicated (look at the total yards) and the Seattle and Dallas games were also won in spite of losing the turnover battle. Meanwhile, teams like the Bengals and especially the Giants are achieving great succcess and notoriety because their opponents have been sloppy and turning the ball over like crazy and they've been able to capitalize on that and run up the score.

No Redskin fan thinks this is a Superbowl contender or even necessarily a certain playoff team. All that we care about is that we're 3-0 and our team is finally showing signs of heading out of the wilderness thanks to some solid coaching.

90
by Lafcadio (not verified) :: Fri, 10/07/2005 - 6:56am

I think Washington will beat the Eagles at Lincoln, because it will be the last game of the schedule and Reid will ask McNabb, TO, Westbrook, Dawkins... to stay at home in order to be ready for the play-offs. The team which is going to have a wild-card is the one which will win the Giants-Redskins game of week 16. (why is there nobody to ask if Eli is a fluke ? He had good games against the very impressive Cards, Saints and Rams, all at home. And the chargers aren't well known for their scary secondary.)

Writing this I'm checking the schedule, and what I see is scaring ! The Giants will make the play-offs and Eli go to the Pro-Bowl. The only defenses he will face are Philadelphia and Washington. Indeed, only is own division could stop him (I don't know what to think about Dallas' D).

Scarying... Giants 10-6

91
by charles (not verified) :: Fri, 10/07/2005 - 9:09am

The redskins won each of those three games because the other coach got conservative in crunch time while the redskins got aggressive.
After Orton threw that pick around the redzone it was nothing but running plays for chicago.
After Bledsoe threw the flea flicker it was nothing but checkdowns and jones up the middle for dallas. Does anybody realize that dallas got the ball back two times after washington took the lead including once at midfield but couldn't get a field goal. Why? because they are playing and coaching scared this year.
Seattle could have beaten washington when herndon picked off brunell. They had just under a minute and were near the 30. Hasselback and Jackson were having a good fourth quarter but Holmgren got scared. As soon as the redskins got the ball in overtime they went back to pushing the ball down the field with moss and won the game.

Looking at the redskins schedule the rest of the year i'm looking at aggressive coaches and conservative coaches.
Denver-conservative even if they get an early lead on the redskins they won't let plummer air it out to try to lengthen the margin so the game will come down to the fourth quarter again where the redskins will pull it out.
kc- aggressive redskins will probably lose unless they get a boatload of turnovers or if kc doesn't put surtain on moss.
san fran - conservative. come on now
ny giants - aggressive now but will not be when eli starts throwing picks and barber starts fumbling again. in other words they are aggressive when the play bad defenses but coughlin plays close to the vest when they play good defenses. See redskins -giants games from last year.
Philly-aggressive clearly their style of play makes them the best team in the nfc but almost lost to skins last year by playing too conservative. I think it had to do with pinkston and owens getting the yips in that game from #21.
tb-conservative come on now they beat tb last year.
oakland-conservative after moss catches his one random 75 yard pass play in the first half.
san diego-conservative marty ball, they didn't want to score 41 points on the patriots but tomlinson kept breaking tackles and eventually popped up in the end zone.
st. louis, arizona- both aggressive but nfc west teams aka the division where everybody makes their playoff runs on.
dallas, ny giants, philly again- parcells won't let bledsoe air it out on a windy day in d.c., neither will coughlin, and philly will have clinched home-field by then so they will be resting their starters.

That's it the biggest difference this year is that gibbs is coaching more aggresive this year so they will win games they are not supposed to win.

If anybody disagrees i will have michael westbrook punch you in the face.

FREE TIMMY SMITH.

92
by james (not verified) :: Fri, 10/07/2005 - 10:41am

Sign me up for a free timmy shirt!

93
by EorrFU (not verified) :: Fri, 10/07/2005 - 11:01am

I believe that in football that there is a large luck component. Just as in Baseball scoring is dependent on how well a series of players hit and winning depends on a timely but random cluster of walks and hits.
The drive stats back up what I have observed, that the Redskins have been effective moving the ball but have been unable to 'capitalize' in the red zone. While some of this may be due to a different approach in the redzone, I find it much more likely that some bad luck has kept the team from scoring. Much Like success on 3rd downs can over-rate a team, an unusual series of failed plays in the red zone can create a lack of scoring and DVOA. Mutch like cluthch hitting, perhaps success in the red zone has a higher component of luck in it than we often presume. Thus, the small sample size creates statistics which were traditionally thought to be meaningful but in reality have more to do with how successful plays are bunched together.

94
by Pat (not verified) :: Fri, 10/07/2005 - 11:14am

It was meant to point out that 3 teams and 3 games are not enough to make a valid conclusion.

Sooo... why are you making conclusions, then?

That's my point.

I don't think that 3 games (by week 4) aren't enough to make a valid conclusion. At that point, you've got a decent amount of statistical information (based on play-by-play data) and you're starting to have enough independent data sets that you can actually see something.

Seattle only scored 17. What are you talking about.

Seattle scored 17. Washington allowed 23. Or, if you want to be even more pedantically statistical, Washington allowed 17+0.7*3+0.7*3 = 21.2 points. The fact that Seattle didn't score those 6 points doesn't change the fact that Washington allowed them. Unless the Redskins have a field goal force field, the ability to score those points or not was entirely contained within the Seattle special teams.

As a similar example, consider a team that scores 3 touchdowns on the Redskins, yet misses all 3 extra points. Are you going to say that Washington's defense held them to less than 20? No, they didn't. They held them to 21, and the other team generously handed back 3 points.

This is why points are such a biased metric: because the fact that Seattle scored 17 and not 23 (or at least 20) has absolutely nothing to do with Washington's defense.

Likewise, Washington scored 17 (in regulation), but Seattle allowed 20 (or 19.4 if you want to be pedantic).

95
by Xao (not verified) :: Fri, 10/07/2005 - 11:35am

re: 88
What is funny is that this site was created to show an alternate side to conventional wisdom. It was created to show what conventional wisdom (avg. yds by running back, qb rating, etc) could not.

Now those tools have become conventional wisdom themselves with too many people “afraid� to look beyond those stats for the next level of thinking.

The problem with this claim is that you're not delving to the next level. You're falling back to simpler stats in an effort to gloss over the failings that DVOA has pointed out.

Don’t tell me that the skins are lucky to convert on third and long 52% of the time. Give me a play by play description of why that play shouldn’t have worked. Did a defender miss an assigment. Did a receiver make a once in a lifetime catch. Were 10 tackles missed on one guy for a lucky play. Did a cornerback fall down. Did a lineman miss a sure sack. Until you can do that you are just reverting back to conventional wisdom. Call me stupid but this site is for those who don’t believe in conventional wisdom.

Looking at play by play descriptions to form a cumulative analysis is exactly what DVOA does, yet you're willing to throw out the results without being able to refute them. In any event, it's not necessary to evaluate each play to know that converting better than 50% of your third and long plays is an unsustainable trend. Just take a quick look back over the league's history and see how many teams have been able to do it in, say, the last twenty years.

96
by Justus (not verified) :: Fri, 10/07/2005 - 11:36am

Re james' #88

Extraordinary claims require extraordinary proof. Saying that Gibbs is reinventing the NFL offense and that 50+% conversion rates on 3rd and long are susstainable requires a tremendous level of proof. From *you*. Not from the people who think it is unsustainable. They don't have to break down each of those 3rd and long situations, you do.

Unfortunately for all sides involved the sample size is currently limited, even more limited than football stats usually are, so homerism overwhelms the discussion as people have a tendency to cherry pick statistics that favor their point of view. If you are serious about wanting to somehow revisit the composition of DVOA or suggesting the creation of additional metrics...well you are simply crazy to suggest that on the basis of three games. Again, extraordinary claims require extraordinary proof. The insufficiency of DVOA cannot be demonstrated in three games.

97
by James Gibson (not verified) :: Fri, 10/07/2005 - 11:42am

James (#61) - looking at the drive stats in isolation is a mistake, in my opininon simply because they aren't adjusted for turnovers. And Washington's defense is towards the bottom in both fumbles/drive and INTs/drive. Fumbles tend to be worth about 50 yards and INTs worth about 45 yards.

An example - in 2000, Tennessee had a better defensive straight up drive ranking than Baltimore, but Baltimore's D was better after turnovers were accounted for (non schedule-adjusted). Click my name to see what I compiled for that year. Baltimore's non-adjusted VOA is significantly better than Tennessee's that year too (which I found on this site).

98
by Charles Fischer (not verified) :: Fri, 10/07/2005 - 11:45am

"In overtime, after Seattle outplayed them for 60 minutes, the Redskins exposed Seattle’s fatal flaw by converting three straight third-and-long situations to set up the winning field goal."

Did you actually watch the game?

Seattle didn't "outplay" Washington for 60 minutes before OT. Possibly one of the most ridiculous things I've ever read.

99
by MRH (not verified) :: Fri, 10/07/2005 - 12:05pm

Re sample size discussion. Ithink the point has been made already, but I thought I'd repeat it:

One reason DVOA is a better metric to use than team record or total pts or even drive stats at this point in the season is that DVOA is based on plays while the other the others are based on number of games (3) or drives (32). All these samples are small relative to their total at the end of the season, but DVOA is explicitly meant to overcome the problems of the inherently small sample size of NFL games:

Football statistics can't be analyzed in the same way baseball statistics are. After all, there are only 16 games in a season. Baseball has ten times more, and even the NBA offers five times more. The more games, the more events to analyze, and the more events to analyze, the more statistical significance. That is true, but the trick is to consider each play in an NFL game as a separate event.

-- From FO's explanation of their methods

Additionally, DVOA uses opponent quality to further correct for statistical illusions found in those other methods. The drive analysis above mentions opponent quality but most other comments have ignored or played it down.

100
by Pat (not verified) :: Fri, 10/07/2005 - 12:06pm

Mutch like cluthch hitting, perhaps success in the red zone has a higher component of luck in it than we often presume.

Red zone performance is used in the "estimated wins" calculation.

Estimated wins are predictive. Thus, red zone performance is likely to be predictive - that is, poor red zone performance is likely indicative of a fault in the team, not just luck.

101
by Pat (not verified) :: Fri, 10/07/2005 - 12:26pm

Of the three teams that aaron reference how many were 52% on third down and had as many yards as Washington. I’m not sure what the answer is and it is an honest question.

Any team that is 3-0 playing above its means is going to have a stat that's out of line with reality. As Fnor put it, this is Washington's.

The 50% conversion rate on 3rd and Long isn't predictive. It's not a "strength of Washington's offense." It's luck.

If you wanted to, you could predict what Washington's 3rd and long conversion percentage should be, by using the 3rd and Short conversion efficiency, and using the dropoff for the rest of the league from 3rd and Short to 3rd and Long, and apply that, and you'd see that Washington's overconverting 3rd and Long.

It's like flipping a coin 10 times, and getting 9 heads and a tail. You don't expect on the next 10 flips to get 9 heads again. You expect 5.

102
by james (not verified) :: Fri, 10/07/2005 - 1:17pm

"Looking at play by play descriptions to form a cumulative analysis is exactly what DVOA does, yet you’re willing to throw out the results without being able to refute them. In any event, it’s not necessary to evaluate each play to know that converting better than 50% of your third and long plays is an unsustainable trend. Just take a quick look back over the league’s history and see how many teams have been able to do it in, say, the last twenty years."

No team has ever won throwing it 60% of the time as much as the eagles have yet they have been winning for over 5 years now. Because it had never been done before people said it could not be sustained.

Now, with the redskins converting 52% of third downs despite not playing very well on 1st and 2nd, this cannot be sustained.

That is not necessarily true. Be careful what you ascertain just because its never been done before.

How many 6th round qb's had won 3 superbowl titles before Tom Brady.

How many teams every won a game throwing on their first 25 attempts.

People who think outside of the box get ahead faster than those who only think conventionally.

Maybe just maybe thats what the skins are doing.

103
by B (not verified) :: Fri, 10/07/2005 - 1:23pm

That's a gambler's fallacy if I've ever seen one. No baseball team ever came back from a 3-0 deficit in the playoffs until the Red Sox did it last year. Does this mean that coming back form 3-0 will now be a regular occurrence? Of course not.

104
by Ray (not verified) :: Fri, 10/07/2005 - 1:29pm

C'mon James. Your argument in #102 is a huge stretch and you know it. The Eagles run/pass balance is a strategy choice. Picking a good QB in the 6th round is a personelle choice. Passing 25 times to start a game is another strategy choice. See where the common thread is here? All those things you mentioned are specific choices made by a team.

Converting a 3rd down is not a choice. It's a play in the game that may work and it may not, no matter what the team chooses. If you're putting the Redskins 50%+ 3rd and long conversion rate so far as evidence of a new successful strategy, then I'll say the team is still terrible, because if it was just a choice, I'd choose to convert 100% of those 3rd and long plays instead of 50%.

Historically, there has been a certain percentage chance that a 3rd and long will be converted. Sometimes a team will get lucky and convert more. But at some point odds are that luck will run out, and suddenly things won't go so well anymore.

105
by james (not verified) :: Fri, 10/07/2005 - 1:48pm

Why is everyone calling execution luck?

It's so frustrating.

I am making one point with two "corralaries".
1.There are other ways to analyze a team besides DVOA
1a. even using DVOA 3 games into the season is a little bit of a stretch
1b. using 3 teams that 100% different from the redskins in every aspect of the game except their beginning win streak and point differential is a classic case of apples and oranges. That is unless you can point out other similarities in DVOA, game situations, etc. If this redskins team had started 3-0 in other seasons, winning by less than points each game, and then went on to finish below .500 for the rest of the season then that argument could be valid.

However, since we are using an invalid or statistically insignifant data to measure the redskins, I use other invalid and insignificant data to show a different picture and am automatically attacked for doing so.

Do you see the hypocrisy in listening to one argument that isn't really valid but attacking another that also isn't really valid.

To me there can only be one reason for that and that reason is that your mind is made up before you read what's presented.

An example is how I show a previous year's correlation with a different set of data(Drive Stats) as a point of reference and some readers are bringing up DVOA to disprove that theory.

They have usually matched up by the end of the year, yet it is assumed that Drive Stats have to go down to the level of DVOA and not the other way around. Why? Because DVOA matches your predetermined mindset.

I have been using Drive Stats for two years to predict the success of teams over seasons or just the next game because to me it makes much more sense and a lot of people are going to feel stupid for not at least entertaining the thought that maybe they are saying what DVOA can't right now.

I too considered the skins to be flukey even as a fan, until I started looking at Drive Stats closer. Then I looked at Portis's succes rate. Then I looked at Brunell's DVOA and his DPAR.

What you have is a possibility, though a bit unlikely, that this is the beginning of a better Skins' offense and not the end of a flukey streak. Am I saying they should continue to try to rely on third down? No I'm saying maybe they won't have to as much if they continue what is an improvement in almost every department from last year.

As far as the field goal argument. Noone is mentioning the skins might have to change their offensive philosophy because their pro bowl caliber kicker has been injured for over two games. So, yes the skins have gotten a little lucky(even though Seattle made a 53 yarder last week) in the field goals against department they have been unlucky in the field goals for department. Yes, they haven't attempted many, but that could be do to the fact that they have a rookie kicker that cannot be trusted.

I am aware that I am making alot of asssumptions from 3 games but so did Aaron's argument.

So I think my argument should be given the same amount of respect(not that I deserve the same amount of respect as Aaron). Just from an argument point of view we both use iffy stats and correlations to make a point.

Most are accepting Aaron's because they are in disbelief after assuming another 6-10 season from the skins. You guys are predisposed to agreeing that the Redskins suck no matter what facts are being used to support that feeling.

thanks for reading

106
by Pat (not verified) :: Fri, 10/07/2005 - 1:50pm

James, B in #103 is exactly correct. Go look up gambler's fallacy if you don't understand it. You can't sustain a conversion rate of 3rd and long higher than (or equivalent to) 3rd and short for long.

So either all the rest of the information on the Redskins offense is wrong, or the 3rd down and Long conversion rate is wrong. Occam's Razor chooses the latter.

If you’re putting the Redskins 50%+ 3rd and long conversion rate so far as evidence of a new successful strategy, then I’ll say the team is still terrible,

Let me state this better for you. I'll say the team is still terrible, because they should be running those 50% successful plays for 10 yards on first down. :)

107
by brasilbear (not verified) :: Fri, 10/07/2005 - 1:53pm

Are the Redskins this years 2001 Bears? Its an honest questions, I'm not trying to troll.

108
by james (not verified) :: Fri, 10/07/2005 - 1:58pm

I know exactly what a gambler's fallacy is.

It works both ways.

It basically states.

Just because something was done before doesn't mean it can be done again.

It also states just because something has never before doesn't mean it can't be done.

By bringing up the gambler's fallacy he is contradicting himself.

109
by james (not verified) :: Fri, 10/07/2005 - 2:01pm

no. 104
Exactly as I suspected.

Before the eagles started winning with a 60% run/pass ratio would you have said it could be done?

Before Tom Brady won 3 SBs as a 6th rd pick would you have said it could be done.

All I'm saying is people spend too much time saying what can't be done instead of saying lets see how it can't be done.

Every team is different. DVOA does a great job of comparing how different teams get the job done. I assume alot of the formulas include scoring plays and thats how most teams DVOA matches their scoring against/for rankings. The most dangerous teams are the ones that are hard to understand how they get it done. It's called being unconventional. It harder to solve a problem that's never been seen before.

What DVOA is great for is showing divergance from conventional statistical measurements. Meaning there become two different measures of the same output. Is DVOA always right? Not at all. After all the NFC runner up was the 19th team overall in DVOA.

110
by Pat (not verified) :: Fri, 10/07/2005 - 2:04pm

Do you see the hypocrisy in listening to one argument that isn’t really valid but attacking another that also isn’t really valid.

Why are you giving information that you don't think is valid? If you don't think that 3 games are enough data to base opinions on, wait for two more weeks. I actually think there's a good chance Washington might win the next two weeks (and still be a crappy team) so wait until then.

Otherwise, drop the sample size argument.

Anyway, I don't buy the sample size argument. 3 games is plenty. You've got something like 12 independent data sets to analyze, and something like 60-70 data points for each set. That's plenty.

As far as the field goal argument.

In the two games that Hall has been out, the Redskins have punted from their opponent's 41 (that's a 59 yard field goal), opponent's 42 (that's a 60 yard field goal), and opponent's 42 (see above). Those are the only times where they punted inside their opponent's half in the last two games. All the other drives they either scored touchdowns, kicked a field goal, or turned the ball over.

They have lost 3 points due to Hall being out. One blocked FG. That's it. No other changes so far.

Most are accepting Aaron’s because they are in disbelief after assuming another 6-10 season from the skins.

That's a very insulting accusation to make. Don't presume to know other people's motivations.

111
by Yakuza Rich (not verified) :: Fri, 10/07/2005 - 2:07pm

The one thing I've been on the lookout when I watch Washington is the play of their secondary in coverage. KC Joyner's "Football Scientist" book, praised the '04 Skins defense. However, he did mention that their secondary allowed for wide receivers to get open at an inordinate rate last year. Joyner later surmised that Washington better get a better pass rusher because sooner or later, opposing QB's will start completing those passes to those open receivers.

Here are the numbers of the last 2 opposing QB's against the Skins defense (Bledsoe and Hasselback):

47 for 74 (63.5%)
503 yards
251.5 yards per game
6.7 yards per attempt
2 TD’s
0 INT’s
92.3 QB rating
1 sack

It's obvious you cannot run against the Washington defense. Perhaps you can if you think pass first.

However, it seems that Joyner was right and you can pass on the Skins defense.

It will take a coach that's smart and has some guts to do it, but once they figure out that you need to probably throw the ball about 40-50 times or more against the Skins, then you'll score points against their defense.

112
by Pat (not verified) :: Fri, 10/07/2005 - 2:10pm

Just because something was done before doesn’t mean it can be done again.

No, it doesn't. And by stating that, you clearly don't understand what it means.

The strict gambler's fallacy is the belief that a random process's results are biased by recent results. That is, if you flip a coin 10 times, and get 9 heads, you're "due" for a tails. That's the strict gambler's fallacy.

The more general "gambler's-type fallacy" is the belief that a statistical result which is at odds with other known statistics, but consistent with a fluctuation, is anything other than a fluctuation. The reason it's a fallacy is Occam's Razor. A fluctuation is the explanation that requires the least additional information.

"This table is hot" or "this coin is lucky" are examples of that.

The belief that a 50% conversion rate on 3rd and long is anything other than a fluctuation is a perfect example of that.

113
by Yakuza Rich (not verified) :: Fri, 10/07/2005 - 2:13pm

****In the Dallas game, they should have gotten those touchdowns earlier, but they didn’t,****

Err....how? When?

Washington could barely get it to midfield the entire game. When they did go deep, guys like Anthony Henry had perfect coverage.

114
by Pat (not verified) :: Fri, 10/07/2005 - 2:13pm

It will take a coach that’s smart and has some guts to do it, but once they figure out that you need to probably throw the ball about 40-50 times or more against the Skins, then you’ll score points against their defense.

Too bad for the Redskins they play Philadelphia twice and St. Louis, then.

115
by TMK (not verified) :: Fri, 10/07/2005 - 2:40pm

Man, Washington acolytes are the most virulent I've seen out here -- and I've read every comment Kibbles has ever posted about Denver.

All Aaron did was find teams that have the same statistical stretch that Washington has put together so far -- and pointed out that only one of them (the 1999 Titans) has amounted to much in the latter parts of the season. He even invited Washington fans to make the case that they match up with that team.

In response, he has gotten flamed and insulted by a bunch of posters who think that volume and shrillness make up for research. For everyone who complains about Carl, at least his posts usually involve reference and statistical evidence, not mindless boosterism.

Is Gibbs a HoF coach? Yes. Does that automatically elevate this team to the one out of seven that compare? No, primarily because the level of competition (Orton's first start, a bad Dallas team, and Seattle coming east, where they always flail about) hasn't met that standard yet. In this NFC, it does mean they're a playoff contender; it doesn't mean they're a good team.

So let's cool all the vitriol, OK? There are enough "true believers" in DC already.

116
by james (not verified) :: Fri, 10/07/2005 - 3:01pm

Pat,
Just come out and say that you won't accept anything besides an argument that states that the redskins aren't good.

re: gambler's fallacy
You are comparing something with an exact scientifically proven probability with something that has an inexact "sort of happens most of the time" probability. Hasn't 3rd down conversion rate been the one thing that varies wildly from DVOA. I think Aaron has stated that many times.

A more productive exercise might be to go back and find teams who have converted 50% third downs over a 3 game stretch and how many have been able to keep that pace. Instead it is assumed that it can't be sustained.

re: invalid arguments

I state that if Aaron can present an invalid argument and draw conclusion than surely I should be allowed to as well since his argument is the basis for this discussion.

Your answer is to not have an argument.

That's ridiculous.

Re: third down conversion rate

Is it just me or is too much being made on this subject. The skins aren't scoring any points so in effect third down conversion isnt having any effect on the outcome of games. Their points are coming off of big plays. Pass, interferance, 40yd td pass, 70 yd td pass. They have only one other touchdown. The third down argument does not apply to this team because they aren't even making good use of it. If you want to argue field position you may have a point. But consider that they're defense is already working with the third worst starting field position in the league.

Different Subject
------------------
Alot of the main posters take offense to what I write.I'm guessing its because I only write when I disagree, and when I disagree it usually something that argues against DVOA as a correct measurement of ALL teams. When I agree I dont write anything, what is the point?

For instance, I wrote on one of the prediction threads and got a lot of heat from B for my predictions. Because I stated that I didn't agree with the prediction method. So far my method is just as correct as the site's, if not more.

I realize that everyone at the site works extremely hard and has put in a lot of work towards their method of statistical analysis. I am however allowed to disagree with some of the conclusions made from that data. Will I always use DVOA sometimes when I need an answer that can't be found by conventional stats the way I like? Yes. Do I always look to DVOa for all of the answers? no. Do I catch heat anytime I point this out? yes. I feel like the kids who weren't allowed to dance in FootLoose. I'm very sorry if my 90% approval rating for DVOA is offensive to some on this site.

DVOA has ranked one team 19th that was an NFC runner up and another 20th that came pretty close to winning a superbowl.

While most of the time correct it does not always do a perfect job of "grading" a team. It is doing a poor job of "grading" the redskins, IMO. I use stats that show a different picture. I share my stats, and am pretty much lambasted as an idiot.

Sorry I don't subscribe to DVOA religously as some do. Sometimes you have to look in more than one place to find your answers.

Just so everyone knows the year Carolina's DVOA was terrible, they ranked 7th in yds/dr 6th in pts/dr and 7th in dsr defensively. Defensive their DVOA was -3.6%. Similar to skins this year, yep. There was probably just as much skepticism about the Panthers as there were about the Skins right now. The point is Drive Stats are a little more predictive than DVOA when considering a team that is winning but has low DVOA.

check it out.

thanks for reading.

117
by DavidH (not verified) :: Fri, 10/07/2005 - 3:10pm

I want to point out one way that I suspect yds/dr may be misleading. The limit to the number of yards you can gain on any drive is not fixed. It is dependent on starting field position. So if teams A and B both have great offenses, but team A also has a great defense that gives them good field position, team B will probably have more yds/dr, simply because they have more possible yds/dr.

So I like to look at percent of possible yards gained (%gained):

(yds/dr)
------------------------------------
100 - (avg starting field position)

Why do I bring this up in this discussion? Because, coincidentally enough, Washington has the worst average starting field position in the NFL (their own 21.77) by a large margin. The difference between Washington and the 29th ranked team (Seattle, 26.02) is more than the difference between Seattle and the 14th ranked team (Buffalo, 30.10) As a result, I think their drives/yd ranking overrates their offense.

Here are the %gained rankings (the last column is the team's ranking in yds/dr):

1 PIT 0.56 (1)
2 SD 0.51 (2)
3 CIN 0.51 (10)
4 PHI 0.49 (6)
5 NYG 0.49 (12)
6 SEA 0.48 (3)
7 IND 0.48 (4)
8 NO 0.46 (8)
9 DEN 0.45 (16)
10 KC 0.45 (9)
11 DAL 0.44 (13)
12 WAS 0.44 (5)
13 CLE 0.44 (7)
14 ATL 0.43 (15)
15 STL 0.43 (11)
16 CAR 0.42 (14)
17 TEN 0.40 (22)
18 NE 0.40 (18)
19 GB 0.39 (17)
20 OAK 0.39 (19)
21 TB 0.39 (20)
22 MIA 0.37 (21)
23 ARI 0.36 (23)
24 MIN 0.35 (24)
25 JAC 0.34 (25)
26 DET 0.33 (27)
27 BUF 0.31 (26)
28 NYJ 0.30 (28)
29 CHI 0.30 (30)
30 BAL 0.27 (29)
31 SF 0.27 (32)
32 HOU 0.26 (31)

Washington is 12th here, more in line with their DSR (14th). Notice that they are essentially tied with Cleveland now. If you compare the rest of their numbers, it looks like Washington is Cleveland, but with more turnovers and worse starting field position. (Cle has more pts/dr, and fewer punts, interceptions, and fumbles).

Also, considering field position helps explain Washington's low pts/dr numbers. It's not really that their offense is only 21st best at scoring. It's that their offense has had a more difficult task in front it if it wants to score. I'm not sure exactly how you would adjust for that, but I imagine a good adjustment would also bring the pts/dr up to around the level of DSR.

Hmm, talking about this is making me think DSR may be a pretty important stat...

118
by JasonC23 (not verified) :: Fri, 10/07/2005 - 3:11pm

DVOA is cool until it disagrees with my evaluation of my favorite team; then it sucks.

119
by B (not verified) :: Fri, 10/07/2005 - 3:11pm

Hey, wait a second. I disagreed with your predictions because you kept insisting Green Bay was going to win thier division. I thought the Vikings were going to win, but that's beside the point.

120
by Daniel Warehall (not verified) :: Fri, 10/07/2005 - 3:13pm

Is Washington a fraud? If the statement is “Washington is 3-0 and has a shot at the playoffs,� the answer is no. If the statement is “Washington has been one of the NFL’s best teams so far in 2005,� the answer is yes.

That's from the header at the beginning of the article...

Why can't we all get along?

121
by DavidH (not verified) :: Fri, 10/07/2005 - 3:16pm

I meant to take out that comparison to Cleveland, as it doesn't really add much. oops.

122
by Pat (not verified) :: Fri, 10/07/2005 - 3:20pm

Just come out and say that you won’t accept anything besides an argument that states that the redskins aren’t good.

I again state that this is incredibly insulting. You don't know my motivations. Stop presuming that you do.

The point is Drive Stats are a little more predictive than DVOA when considering a team that is winning but has low DVOA.

Prove it. Show a correlation plot. Showing individual examples is not a proof - it's cherrypicking.

123
by james (not verified) :: Fri, 10/07/2005 - 3:33pm

tmk,
I in fact wrote a very long post with only statistics and previous correlations.

It was still ripped apart for the words used and not the contents. One response only dwelled on what I said about the offense mainly that I said "it was good", when I in fact did not.

Some other responses have dwelled on third down convesions when in fact, with Washington's offense hardly scoring points is not even an explanation of their success.

One might argue that both sides are equally unwilling to buckle.

I at least have been backing up most of my ascertions with stats that I found on the site.

A perfect example of another successful team that has similar stats to the skins are the Carolina Panthers of 2003. I believe they made it all the way to the superbowl. Similar DVOA similar defensive drive stats. That is an illustration of exactly the point I made. When DVOA doesn't explain a team's success defensive drive stats will.

What this all points to is a team that makes you beat them. If you don't beat them they won't beat themselves and they have a little big play ability.

124
by james (not verified) :: Fri, 10/07/2005 - 3:38pm

B,
First I said these stats aren't always the best to predict and then suggested another method. Then I was ripped, alot!. Then I said why don't we predict.

At which point I predicted playoff teams of
Indy,Cinci,NE, Oak, KC

and NFC of
Phi,GB,Atl,sea,stl

We agreed to disagree about that once I was duly informed of GB's line woes. I then selected Chicago,though it somehow never posted.

125
by james (not verified) :: Fri, 10/07/2005 - 3:43pm

Pat,
I wrote a long article with a correlation plot argument to drive stats.

I guess you have chosen not to read it. It is no. 61 if you want to read it.

I have always considered defensive drive stats to be the most predictive when it comes to succes of w/l record.

They work a couple of ways.

You can look at the bottom half underacheivers in terms of yds/dr rank to pts/dr rank.

The teams that rank lower in pts/dr to yds/dr are some very good candidates to improve.

Two good examples are KC in 2003 and SD in 2004.

I like drive stats alot better than DVOA for some things and vice versa for others.

thanks for reading

126
by james (not verified) :: Fri, 10/07/2005 - 3:45pm

add on to no. 124

I left out Bal and Was as two playoff teams I predicted. ooops

127
by James Gibson (not verified) :: Fri, 10/07/2005 - 3:51pm

James (the other one) - I believe that at the end of the season drive stats may be better than VOA for rating teams, but 1) they need to be adjusted for turnovers which yards/drive are not, and 2) they need to be adjusted for strength of schedule such as the way VOA is turned into DVOA, which is the thing that Aaron and company have stated is the big breakrhough akin to park factors in baseball. The other advantage, of course, is that DVOA allows you to rate individuals which drive stats won't. I think after you adjust the drive stats for turnovers, you'll find they are very similar to the VOA calculations. It's figuring out the "D" part that might be tricky.

128
by Pat (not verified) :: Fri, 10/07/2005 - 4:02pm

I wrote a long article with a correlation plot argument to drive stats.

That wasn't a correlation plot. A correlation plot would be if you plotted drive yards vs. wins (corrected for strength of schedule) for the past 5 years.

As far as I can tell, all that was was two statistics (8/10 of the top 10 teams in scoring defense have greater than 0.500 record, 6/10 teams in scoring offense have a greater than 0.500 record) both of which tell you very little about the strength of the statistics themselves, mainly because of the arbitrary cutoff at 0.500.

129
by james (not verified) :: Fri, 10/07/2005 - 4:04pm

James Gibson,
I completely agree with what you say about the advantages of DVOA over Drive Stats.

One thing I have to do with Drive stats is use all 3 yds/dr,pts/dr,and dsr as a rating(some people on this board think I am suggesting using only yds/dr, which is the worst of the three measurements and is far from what I am suggesting). Second thing I have to do is go over to DVOA and see how they match up.

What I usually find is similar ranking. When I don't I usually trust drive stats over DVOA. I like to compare Drive Stats to the video tape and DVOA as the eye witness. Both pretty reliable but only the tape is indisputable.

This all started when I tried to find a formula to winning SuperBowls with DVOA and found that Drive Stats were much better. I dont remember dsr being there before this year so maybe I never paid attention to it, or maybe its new. What I found were defense drive stats were a better predictor of success. I found that you need to be around the top 10 in most of the categories except yds/dr on defense to have a chance at the superbowl.

Not making mistakes on offense,special teams seemed to be the other requirement, but I never came up with anything in that department because defense was so predictive and offense was not very predictive.

130
by James (not verified) :: Fri, 10/07/2005 - 4:05pm

Pat,
Please reread the article. I said yds/dr is the least important of the stats and would only be used to try and predict where the redskins pts/dr might end up.

131
by DavidH (not verified) :: Fri, 10/07/2005 - 4:07pm

The point is Drive Stats are a little more predictive than DVOA when considering a team that is winning but has low DVOA.

I think this is close to a tautology. Switch drive stats and DVOA, or even replace them with any random categories you want (last year's winning percentage, number of first downs...), and the statement will still be true. If you only look at instances where one measure is failing, then any other measure should do as well or better.

132
by Pat (not verified) :: Fri, 10/07/2005 - 4:19pm

Please reread the article. I said yds/dr is the least important of the stats and would only be used to try and predict where the redskins pts/dr might end up.

Typo. Substitute points/dr for yds/dr.

133
by james (not verified) :: Fri, 10/07/2005 - 4:26pm

Pat,
Here's an abstract. Maybe this will make my argument a little more clear.

I stated where most teams in the top 10 of last year's pts/dr defense ended up record wise. That was .500 or above.

I then stated with that type of defensive performance the offense had to make the top half of pts/dr for the team to make the playoffs. I pointed out Was has acheived a top half pts/dr rank over the last two games, which are Brunell's only two starts.

I then stated that teams who are top 10 in both reach elite status and only NE,PIT,PHI reached that plateua last year. I then stated that if the Skins could keep their top 5 status they have right now while having faced above average teamsin yds/dr defense (13.3 avg) that it wouldn't be that far of a stretch to assume that they can keep that now that the avg in yds/dr defense of their remaning competition is 18.

I then found mode,media,range of the differential(between pts/dr and yds/driv) of last year's teams' drive stats to be 3,2.96,3. Meaning that most teams were inside of 3 in differential.

Using the logic the Skins could in actuality end up with a top 10 offense in pts/dr. 5+ 3 = 8.

Key word was could. I am not at all stating what will happen. Only likely outcomes that I see.

Lambast if you must but it was clear, concise, mathematically correct way to predict stats according to the information we already have.

134
by james (not verified) :: Fri, 10/07/2005 - 4:33pm

David H,
I am speaking of instances of winning teams with low defensive DVOA and high Drive Stats rank(yds/pt).

Another example of a successful team from 2004 is the Jets 2.4% DVOA 8th in pts/dr. 11th in yds/pt. 10th in DSR.

When you see a winning team with a low defensive DVOA usually you will find much better drive stats.

135
by Ryan Carney (not verified) :: Fri, 10/07/2005 - 4:35pm

Any team that can be 0-3 just as easily as they could be 3-0 does not deserve to be talked about as a playoff team. They are not even the second best team in their division. They are going to get beat twice by a very good Philly team, and twice by a very surprising Giant team, led by, I never thought I'd say this, an exponentially improving Eli Manning. All you have to do is watch the games to see the mediocrity of the Redskins, three nailbiters against three teams that have no aspirations of contending for anything more than a wild card playoff spot, and even that by default due to divisions (Sea, Chi). This is an anomaly, and this is a football team that coming out of Kansas City and Denver in the next two weeks, will be 3-2, third in their divisionbehind the 4-1 Eagles and 4-2 or 5-1 Giants

136
by Pat (not verified) :: Fri, 10/07/2005 - 4:35pm

james:

Sorry, that still makes virtually no sense.

137
by james (not verified) :: Fri, 10/07/2005 - 4:36pm

re 132,

Actually not a typo. Sorry that you are trying to hang your hat on anything to disprove my arguement but there is statistical research behind my argument and you choose to see what you want to see instead of try and understand my point.

You're not interested in another point of view.

138
by Pat (not verified) :: Fri, 10/07/2005 - 4:36pm

Sorry, let me correct that. That still makes virtually no sense to me. I have no idea what you're trying to say there.

139
by Pat (not verified) :: Fri, 10/07/2005 - 4:38pm

#137:

My typo, not yours.

You’re not interested in another point of view.

OK, that's it. That's the third time you've gone out of your way to criticize my motives based on absolutely no information. Forget it.

140
by James Gibson (not verified) :: Fri, 10/07/2005 - 4:39pm

james - assuming that points/drive is the ultimate measure but that you want to break down what goes into the dirve, I have run the correlations for the 2003 season for a number of stats. This is what I get:

VOA to points/drive: 0.84
A statistic that is similar to OPS in baseball - that is uses first downs/down set (OBP) and yards/drive (SLG), and is adjusted for turnovers: 0.8
The same as above but uses yards/down set instead of yards/drive: 0.79
Just first downs + TDs/down set: 0.79 (I will note this was also in Pro Football Prospectus and this website as DSR, although I ran these before I knew this site came up with that stat). I should also note that when I broke these two stats down that the DSR stat is worth about 1 and the "SLG" stat is worth about 0.03. So, the yardage stats only help slightly.
First Downs / Down Set not including the TDs: 0.76
Turnover adjusted yards/drive: 0.74
Yards/Drive: 0.65
Yards/Down Set: 0.44

Based on this list, VOA by far has the best correlation to points/drive. Now, it is true that Washingon ranks well in points/drive and in DSR, which are good indicators. But I still think you are neglecting the value of turnovers where Washington's D ranks very poorly.

141
by james (not verified) :: Fri, 10/07/2005 - 4:50pm

Even simpler.

1. top 10 teams in defense drive stats pts/dr are usually above .500.
*skins are acheiving that
2. teams with top half offensive pts/dr and top 10 defense pts/dr are usually playoff teams
*In Brunell's two starts skins are acheiving that

What cant you understand?

Then in an attempt to project offensive output
1. Skins yds/dr 5th.
2. The median,mode,mean(middle school math here, when they are close it usually it correct to jump to conclusions using that data) are all 3 from last year's data.
3. The skins remaining opponents are worst in yds/dr than the opponents they have faced so far.
4. They should be able to sustain that sort of level considering where they are now and a drop in the level of competition.
5. Using the median, mode, and mean of a 3 differential a realistic projection is 8th in pts/dr offense.
6. If they don't do that well they should still make the playoffs considering the 11 was the extreme of differential last year which would put them at 16 in yds/dr. Just enough for top half.

It's a different way to look not the only way. Might as well present two sides of an argument instead of taking one as the gospel.

142
by Jeff J. (not verified) :: Fri, 10/07/2005 - 4:53pm

Wow, sorry I didn't jump into the fray as soon as Aaron posted the article. At the risk of rehashing many other thoughts, let me concur that while this is an interesting article, it fails to convince me of the "luck" Washington has had.

I'm a diehard Skins fan, and believe we have an outside shot at the playoffs based on our talent and now our record. But I'm realistic enough to note that we need to win the turnover battles and run the ball better to really be a force in the NFL.

To suggest that Brunell to Moss in Dallas and our win over Seattle was more luck than anything else is to discount all the DVO-whatever analysis. Moss burning the Cowchips once may be construed as luck; a second time--with the game on the line--requires the coach to make the play call, great o-line protection, an aware QB to make a perfect pass, Moss to have the speed/ability to get open, and then to make the catch.

Every one of those happened. Twice. That's lucky?

I would concur also with the upthread posters that Aaron likely didn't watch the Seattle game. I was there. We didn't score well, but our offense was efficient in keeping Seattle off the field. That has to count for something, and something accounts for it; and it ain't luck.

There isn't a bandwagon, by the way. Just people who love their team and suddenly have optimism.

143
by Pat (not verified) :: Fri, 10/07/2005 - 5:03pm

Every one of those happened. Twice. That’s lucky?

It is when you're facing Roy Williams. There's simply no way that would've worked had they been facing any team with a decent secondary. Moss got open because Williams blew the coverage on him twice. Virtually the same coverage, too.

It's frustrating, too. Williams should be a good SS. He really should. But he just consistently makes mistakes in coverage, and Dallas just ignores it.

144
by james (not verified) :: Fri, 10/07/2005 - 5:03pm

James Gibson,

It doesn't always work that way. The skins were 22nd in defensive take aways last year and 3rd in defense DVOA.

I'm not calling pts/dr the ultimate indicator. I'm calling it another indicator.

I'm sure I could find DVOA correlations to record as well and that would point to the skins doing poorly.

My point that everyone is missing for some reason, is that there are gonna be 1 or 2 teams every year who are winning despite poor DVOA. To point out that in 2004 one made the NFC championship and that in 2003 one made the superbowl is only to illustrate that sometimes DVOA can be extremely off.

What I'm starting to think is that DVOA is more a measure of ability to dominate.

Drive Stats are ability to stay in game.

Much rather have a good DVOA then good drive stats, but when high DVOA isn't possible good drive stats are a good substitute. Also good DVOA without good drive stats is an indicator of impending doom. It shows teams that can't win without dominating.

I hate that I'm coming off as such a homer but the skins are the best of example of potential to be one of those unexplained by dvoa winning teams this year according to other measures.

145
by B (not verified) :: Fri, 10/07/2005 - 5:08pm

"My point that everyone is missing for some reason, is that there are gonna be 1 or 2 teams every year who are winning despite poor DVOA"
Honestly, I don't thin anybody is missing this point. DVOA doesn't coorelate perfectly with wins because DVOA is opponent-adjusted, so a team who faces a weak schedule will have a record superior to thier DVOA. If you want to coorelate something with wins, VOA works much better.

146
by Trogdor (not verified) :: Fri, 10/07/2005 - 5:09pm

DavidH, I think the Cleveland comparison is actually a pretty good one. What would Washington's record be if they played Cincy, at GB, at Indy? Is there any reason to think it would be better than Cleveland's 1-2?

I've seen 2 of 3 Washington games this year, and here's my impressions:

1) Chicago. They're a good defense at home, facing a rookie QB making his first start, after being forced into service when the projected starter broke his leg a week before. I would certainly hope they'd win. And they did.

2) Dallas. Thoroughly outplayed, then the Dallas secondary made two junior high school level plays, and they steal a win. It's an important win, but a strategy built on "keep it close and hope our opponent royally *#&$s up twice" is, how do you say, not sustainable.

3) I didn't see Seattle, and honestly, who gives a crap?

So there you have it. DVOA says they're an average team that's gotten a few lucky breaks. Conventional stats say they're an average team that's gotten a few lucky breaks. Observation says they're an average team that's gotten a few lucky breaks. I find absolutely no reason to believe they're suddenly an all-world team, and have to believe that a) if their performance remains the same, things will even out and they'll go about .500 the rest of the way, or b) their performance will improve to match the record. (they could also go completely in the tank, but I doubt it) It is extremely unlikely that they can continue at their current level of play, and go 9-4 or better the rest of the way.

As for 3rd down conversions, well... there's a reason baseball doesn't hand out the batting title in April. Here's a good way to lose a fantasy league: find a guy who's a career .260 hitter, who hits .380 the first three weeks, and dump a slow-starting Manny Ramirez for him. For a month all your bloop hits can fall in for singles, and your passes on 3rd and 12 will find their targets. But don't plan your offense on it happening all year. Eventually those bloops will find their way into opposing gloves (and the .380 hitter will regress towards his normal .260), and those passes will fall incomplete or worse. The breaks will start to even out over time (although they may not get there over 16 games, of course).

Planning on offense around converting over 50% 3rd and long would be like the Eagles gameplanning in the playoffs around converting all their 4th and 26s. It's great when it happens, and count your blessings, but if you find yourself relying on it frequently, get ready for trouble.

So what do I think of the Skins? Well, those 3 wins count, and much like a good April in baseball, it puts them in a great position. If they go about .500 the rest of the way, they make the playoffs (9-7 or 10-6).

About the 03 Panthers - they've been written about a lot here. Like the Skins, they won a lot of close games while playing mediocre ball. The thing with them is, once the playoffs came they went on an incredible hot streak, and actually played like a top team. Were they playing possum, doing just enough to get in and then turning it on when it counted most? Who knows? Are the Skins doing that?

147
by james (not verified) :: Fri, 10/07/2005 - 5:11pm

Pat,

Not questioning your motives.

I presented a logical argument backed up by past data. You continue to say an argument backed up by data from last year cannot be understood.

That leaves me baffled.

I suspected your mind was made up that the skins were bad before you ever read my post.

May be egotistical of me and for that I apologize.

I'm guessing you might be a cowboy's fan after reading what you wrote about Roy's performance being frustrating. If you are a cowboys fan then I take back everything. There is no getting you to look at the skins from two sides of a story.

148
by Pat (not verified) :: Fri, 10/07/2005 - 5:20pm

If you want to coorelate something with wins, VOA works much better.

Estimated wins works even better. It takes into account statistical fluctuations (a good team that never fluctuates wins more, a bad team that never fluctuates wins far less), and a bunch of other variables that lead more directly to winning, like performance in the red zone, etc.

Estimated wins isn't adjusted for strength of schedule, though. To do that, you'd need to compute the number of excess wins that a team gets based on the distribution of their schedule. It's not good enough to average a schedule - imagine if you played a team that would go 16-0 against 0% DVOA opponents. A schedule that contains 3 30% DVOA teams, and 3 -30% DVOA teams is likely to give that team, say, ~1.5 losses due to the 3 30% DVOA teams, whereas the 3 -30% teams will only give you 3 wins you would've gotten anyway.

You'd face this same problem with VOA, though. The Titans are the classic example so far. They'd probably be a 4-12 team, except for the fact that they play Houston twice, Baltimore, and San Francisco, which gives them over 2 wins in excess of schedule normal.

So to correlate with actual wins, you'd need something like estimated wins, plus something like the SWXN I've mentioned in other threads.

149
by james (not verified) :: Fri, 10/07/2005 - 5:20pm

Troggie,
IMO, the third down conversions are not contributing to any of the wins. DSR is 14th ranked even though they are 1st in 3rd down percentage. That says that they are missing opportunities in other places and maybe less third down conversion will be made but more scoring opportunities will be converted.

About the Panthers. You are exactly right. You are just illustrating the point I am trying to make that noone seems to get.

Current DVOA does not predict future DVOA.

150
by James Gibson (not verified) :: Fri, 10/07/2005 - 5:21pm

Whether the specific plays were luck or not, the Redskins have very narrow wins, and that to me is more the main point of the article than the DVOA analysis. It's that narrow wins just don't hold up over time.

151
by Pat (not verified) :: Fri, 10/07/2005 - 5:28pm

Current DVOA does not predict future DVOA.

Actually, yes it does, on the whole. There are definitely teams that deviate, or trend upwards or downwards, but on the whole, it is predictive.

In fact, estimated wins at midseason projected to the full season is a better predictor of final record than the record itself is.

152
by DavidH (not verified) :: Fri, 10/07/2005 - 5:28pm

I know that refuting this counterexample is not exactly that important to the discussion, but I do think it illustrates part of the problem here. We are talking about one measurement that is adjusted for opponent, and another that is not. AND we are talking about their correlations to winning, and Wins are another stat that is not adjusted for opponent. Anyway...

Another example of a successful team from 2004 is the Jets 2.4% DVOA 8th in pts/dr. 11th in yds/pt. 10th in DSR.

So they were around the 10th best defensive team in drive stats, which are NOT adjusted for opponent. Let's compare it to their UNADJUSTED defensive VOA from that year (see the link in my name): 12th.

Also let's look at their offensive stats:
drive stats - 12th/9th/9th
unadjusted voa - 5th

It looks like the main difference between drive stats and VOA is that VOA says more of their success is due to offense, while drive stats gives them equal credit. Why would this be? Maybe because those drive stats are not accounting for starting field position:

ranking of avg starting field position of Jets: 22nd
ranking of avg starting field position of opponents: 9th (from the Jets point of view, i.e. 9th worst from the opponents' point of view)

So the offense had worse field position than average, and the defense had better. If you mentally adjust the drive stats to account for that, it looks like drive stats and VOA say the exact same thing about the Jets.

So maybe you should alter this statement:

When you see a winning team with a low defensive DVOA usually you will find much better drive stats.

by adding a "(since the drive stats are not corrected for opponent quality or field position)"

153
by DavidH (not verified) :: Fri, 10/07/2005 - 5:35pm

James:

I also want to say I'm interested in your research into drive stats being predictive of getting to the Super Bowl. NOT because I want something to try and argue with, but because I actually think it might be interesting. Compressing teams into one "master number" like DVOA can obscure analysis of the individual parts. It's like BA/OBP/SLG in baseball. Yeah, OPS is better than any of those alone, but looking at all three is better than compressing them into OPS.

So - do you have the research you did up on a website anywhere, or could you post a little summary of it on here, or email me or something? From your posts here today, all I really got was that defensive drive stats seem to be more important than offensive stats.

154
by DavidH (not verified) :: Fri, 10/07/2005 - 5:37pm

The link that was supposed to be in my name two posts ago is in it this time. Oops.

155
by Pat (not verified) :: Fri, 10/07/2005 - 5:38pm

The 2004 Jets defensive success is also heavily bolstered by the fact that they had a good offense and very strong special teams (punting). This doesn't show up in the drive stats, of course, because of the bias DavidH mentioned before - you don't know where the drive started from.

This pattern is very common. The Eagles last year, the Falcons the year before, etc. Most people forget to take into account special teams defense (KICK+PUNT) when thinking about team defense.

156
by B (not verified) :: Fri, 10/07/2005 - 5:40pm

IMO, the third down conversions are not contributing to any of the wins
What? What???!! What?????!!!
I'm going to assume this is a typo.

157
by Aaron (not verified) :: Fri, 10/07/2005 - 5:40pm

Well, I'm glad I've created some controversy. I just want to clarify two things.

First, I never said that the Santana Moss touchdowns were lucky. The quote from the article is: "Long-bomb touchdowns of this type, though great plays that deservedly gave Washington a one-point victory, do not usually indicate that a team will win games in the future." That's not an issue of luck.

Second, my comment about Seattle outplaying Washington for 60 minutes didn't quite come across the way I meant it. I did not mean that Seattle was better than Washington for every play during all 60 minutes of the game. What I meant was that when you break down play-by-play as a whole during the four quarters of regulation, Seattle played a better game. Washington was better than Seattle in the first half (WAS had 176 net yards, SEA 89). Seattle was better than Washington in the second half. (WAS had 116 net yards, SEA 265). The difference between second half performances was a bit larger than the difference between first half performances.

The Seattle game is also where luck had the clearest role in a win. As noted by other commentators, the missed field goals were not "lucky" because field goal kicking is random. They were "lucky" because the defense has no control over whether a field goal goes through the uprights except when the field goal is blocked. Otherwise, it depends on the snapper, the holder, and the kicker. The missed field goals were Seattle failures, but they weren't Washington successes.

Washington also won the coin flip in overtime, giving them an advantage, particularly against a team like Seattle where the offense is much better than the defense. (Of course, part of the reason why the Seattle offense didn't get the ball in overtime is that the Washington offense played well, and for that they do deserve commendation.)

The issue behind the article is not whether the three wins in the past were "deserved." Who cares. They're wins. All wins count the same. The issue is how good a team is and what we should expect from them going forward. Taking away the question of strength of schedule for the moment, if you break down play-by-play for most of the teams having unexpectedly good seasons -- TB, CIN, NYG, MIA -- the breakdown matches the results, which is a good sign that those teams will have continued success. That simply is not the case with Washington.

158
by Russ (not verified) :: Fri, 10/07/2005 - 5:51pm

#150, you have to be careful with blanket statements that narrow wins don't hold up over a season. The 1976 Raiders won their first 3 games 31-28, 24-21 and 14-13. In week 4 they got blown out 17-48. But they came back and finished 13-1 and won the superbowl. They won 3 more games in the regular season by a touchdown or less, and 1 playoff game by a fieldgoal.

Now I'm not saying the Redskins are the '76 Raiders, but sometimes narrow wins do hold up over a season. It can even hold up over a longer time. From 1975-1977, the Raiders won many more games by a touchdown or less than they lost.

159
by Pat (not verified) :: Fri, 10/07/2005 - 5:53pm

An article about a team that's had a string of losing seasons, been talked about poorly in the press, who's suddenly on short hot streak, not being as good as they might think they are?

Nah, no controversy here!

But I say again: what is it with Washington? You think after the Nationals they would've learned to not get their hopes up.

160
by Yakuza Rich (not verified) :: Fri, 10/07/2005 - 5:56pm

****Moss got open because Williams blew the coverage on him twice. Virtually the same coverage, too.

It’s frustrating, too. Williams should be a good SS. He really should. But he just consistently makes mistakes in coverage, and Dallas just ignores it.****

Parcells blamed Aaron Glenn for those 2 TD's in a press conference, stating that Roy executed his assignments properly on both plays. Behind closed doors, Parcells was reportedly fuming at Dallas D-Coordinator Mike Zimmer for making some awful defensive playcalls in a 90 second span.

Before anybody starts to say "Parcells is protecting his star player"...remember, this is Parcells here. Not only that, he's not likely to put blame on one of "his guys" Aaron Glenn unless it's completely true.

The fact is, Roy is playing very well and perhaps as good as any safety in the league this year. He's only had 2 passes completed on him that were his responsibility and they were for minimal yardage (and not those 2 TD's than S. Moss caught). He's arguably the best run stopping SS in the game and one of the better blitzers.

Dallas isn't ignoring anything. The fact is he's been great this year. Perhaps if there were some announcers and analysts that had a clue when it came to coverage and schemes, they wouldn't be so quick to misjudge Roy's play.

161
by Pat (not verified) :: Fri, 10/07/2005 - 6:49pm

Well, to be honest, that was always my view of him, but it's difficult to guess when you have blown coverage and when the other safety blows it. So, OK, I stand corrected, but that jumps Aaron Glenn up onto my "why are you still playing this guy?" list. :)

162
by thad (not verified) :: Fri, 10/07/2005 - 8:42pm

hey james gibson,
that stat that you compared to ops sounded really interesting.
but, call me stooopid, I am not quite sure how you got it, could you tell us more about it?
thanks!

163
by Yakuza Rich (not verified) :: Fri, 10/07/2005 - 8:55pm

***So, OK, I stand corrected, but that jumps Aaron Glenn up onto my “why are you still playing this guy?� list.***

1. He's the #3 CB. Good #3 CB's don't grow on trees.

2. He's been pretty good outside of the Washington game. Excellent against SD, good against SF and Oakland. Just had two horrible plays against Washington.

The real weakness in the Cowboys secondary is FS Keith Davis who has been horrible so far.

Nobody is talking about it, but the Cowboys were set to pick either FS Brodney Pool or Josh Bullocks in the 2nd round.

Pool was taken as expected. But, Bullocks was surprisingly taken by the Saints so they then had to pick LB Kevin Burnett.

Not to take anything away from Burnett, but if the Cowboys had some better FS play, they'd be a much better defense than they are right now.

164
by Bill (not verified) :: Fri, 10/07/2005 - 9:53pm

Best laugh I've had all week -- thanks. I expect the author will be eating his sour grapes along with Mssrs. Bradshaw and Long before too long.

Following week one, and the "easy" Chicago win, the commentary was the same. Following the "lucky" win over Dallas, ditto. After beating the number-two ranked offense of Seattle, bring on the chorus of "This will not last! This will not last!"

The Redskins have something this season that many of us have not seen in them for a long time: heart, spirit and hope. They are not of the "we'll win just because we exist" league of the Eagles, Pats and (snicker) Vikings and Green Bay.

Long-time Skins fans like me are used to this team's losing with little pity, and winning without respect. One more journalist saying "they can't" is just as good as many more saying "they can."

165
by james (not verified) :: Sat, 10/08/2005 - 12:55am

DavidH,
Defensive drive stats have always helped me more during the playoffs than anything. I wish I could say I had distinctive research for them, but for the most part I have always just gone to drive stats and said who is consistent among the most measurements throwing out yards.

It usually been a way to look at a surprising team. IMO, they unveil what is hidden by other stats. The best D-coordinators seem to play differently depending on the side of the 50. I think that what Drive Stats show is a succesful team philosophy that the players are buying.

I look for overacheiving teams. For instance the Pats one year were 23rd in yds/dr but top 10 in every thing else. To me that stuck out. I don't have anything concrete other than a couple of years of looking at these stats and seeing how teams perform compared. It's usually the overachievers on defense that play the best football. I wish I knew how to explain it and anyone with a good diagnosis would be m best friend. IMO, it signifies a team that has great intangibles. So far, thats what the skins are signifying, though tthat could all change. To tell the truth I have never looked at drive stats early in the season to predict later success.

I guess looking for the team that is the most underrated and overachieves the most(wins despite their measurements) are the best candidates for the SB. The world beaters usually don't pull it out in the clutch.

I'm long winded but again thanks for reading and comments welcomed.

166
by james (not verified) :: Sat, 10/08/2005 - 1:15am

B,

about third down conversions

I'm looking at the down success rate on football outsiders and seeing the skins as only 14th ranked. To me that signifies that maybe the skins aren't converting things to level people are seeing just based on 3rd down conversions. DSR measures td success, and fg success some kind of way and though the skins 3rd down rate 52% they are still only ranked 14th in overall conversion success rate. Possibly this is the measurment we are all looking for.

To me this a number that more fairly assesses the skins offense as middle of the road when it comes to important conversions. Aaron's help would be great on this subject.

Again I am stating that when using DSR instead of just plain old third down conversions the skins are middle of the road. Maybe we are all making too much out of their third down rate and not looking at the bigger picture which DSR might be better at doing.

I don't know, what does everyone else think?

I hope everyone understands that I am looking for more of a discussion than just a blanket who's right and who's wrong. I got a little overly frustrated because I felt that noone was bringing anything new to the discussion besides whats already in the press or "common thinking". By saying that I mean, I want to read more than just the skins suck followed by insults, or something thats already been stated 100 times already.

I'm not stuck on one way of thinking. B, you can atest that in previous arguments I can be shown a better way of thinking and change my mind and be gracious in giving credit to those with good points.

I apoligize to anyone who I offended because my words were probably a bit to offensive. I only want more of a discussion and less blanket statements. I hope thats what we are all hear for. To learn more from other people and not completely be unwilling to change what we came to board thinking.

I think I am coming off to much as someone who wants to change other's minds. That is far from the truth and I wish I had presented myself differently. Instead I am looking for a better way of thinking besides what I can come up with myself or what has already been said.

This is an intersting subject and we can all get more out of this discussion than insults and statements that say little more than "now way post no. 87...you're a moron".

Some posters have been addding to the discussion(davidh and james gibson come to mind off the top of my head). I sometimes have and sometimes haven't and would rather do the former rather than the latter.

thanks for reading all and may everyone's favorite team win except, cowboys, sd, stl,atl,sf, and denver(boys' cuz i hate em and eveyrone else cuz i bet against them)

167
by EorrFU (not verified) :: Sat, 10/08/2005 - 2:30am

The drive stats show that the redskins offense has been proficient at getting first downs. The inability to score is immaterial to the ability of the offense to move the ball down the field.

The defense inability to create turnovers is the primary reason the skins offense has one of the worst starting points on their offensive drives.

It would seem that this lack of defensive ability makes it more difficult to score because the offense must string together more plays on average than otherwise to increase their chance of scoring.

This lack of defensive turnovers may in fact be as important to the Redskins lack of scoring as the offense.

while this may not mean that the team is better than their DVOA, it does seem to suggest the offense is not as anemic as the scoring stats.

In this area, how many yds gained for every point may be an indicator of luck to a certain extent like Balls BAtted In Play (BABIP) in play. Both are probably highly afffected by the defenses quality and the defenses ability to create turnovers deep within an apponents territory.

This also makes me wonder if where a defense creates a turnover is more random and how does this overrate an offense by making it easier to score points. DVOA controls this pretty well but I still wonder how randomly distributed are turnovers.

168
by EorrFU (not verified) :: Sat, 10/08/2005 - 2:31am

Wow, 167 psts so far are amazing. This is great for the site.

169
by james (not verified) :: Sat, 10/08/2005 - 3:00am

Maybe need to rethink my theory of skins defense is good enough to neglect causing turnovers.

Just found out there is a 17% swing(most teams) in DVOA ratings between home games and away games.

What might be interesting is a week by week voa report after every game. Instead of seeing the aggregate we can see what each team accomplished in each game as well the aggregate.

That might move the skins argument along a little. If would be nice to know what the skins voa was against each of their opponents this season and what their opponents voa was agaisnt them.

This would be a great tool to measure should of and could of wins.

Aaron, make it so.!

Just kidding, who could ask for anything more? This site keeps me up most nights.

I know.... I need to get a life

170
by James Gibson (not verified) :: Sat, 10/08/2005 - 11:11am

Thad (#162) - what I did was think about what would be equivalent to OBP and SLG first - which I decided was First Downs + TDs divided by down set for OBP and Yards/Down Set for SLG. I calculated these for all of 2003 and then compared them to points/drive. Using the equation A = Bx + Cy, where A is points/drive, x is DSR, and y is yards/down set, I used the matrix function in Excel to determine what B and C should be. I did the same with yards/drive instead of yards/down set. For yards/drive, I got C = 0.02 and B = 6.1. Then I just ratioed both of them to 6.1 to get the 1 and 0.003 I mentioned above. To tell you the truth, I was surprised how little the yardage stats helped. DSR by itself seems to be the best "single" stat to look at to determine points/drive, without going over play by play logs.

171
by B (not verified) :: Sat, 10/08/2005 - 11:14am

The Redskins have converted 25 of 52 third downs (49%). The Skins have a DSR of .690, which seems like around where it should be with a 49% third down conversion rate. I think what is significant here is they are converting thier third and 10+ at a slightly higher rate: 12 for 19 (52%). This indicates they have been better in third and long then they have in other third down situations. So the question is which is the real skins offense, the one that's very good at converting third and long, or mediocre at converting all third downs. All things being equal, I'd take the larger sample size, so my conclusion is the skins offense is average at converting third downs, but have gotten lucky on a few plays. Of course both sample sizes are pretty small, so it could turn out that the rest of the offense catches up to the third and long performance. However, it's usually a bad football strategy to get into third and long and hope you convert.

172
by Andrew (not verified) :: Sat, 10/08/2005 - 1:19pm

Aaron (#157) has it right. The question is not whether the Skins were a good team over the past three games. It's whether they'll be a good team over the next thirteen.

Saying that a 50% third-down conversion rate is a "fluctuation" is meaningless. Of course that rate won't last through 16 games. But so what? That's not a useful way to describe how good a team is. But it may be predictive of how good the team will be in the remaining games. (Or not.)

We've got lots of stats being thrown around here (DVOA, VOA, DSR, yds/drive, potential yards gained), and they may all be useful, but which of them are going to be the best indicators of future performance? Will performance regress to their current DVOA numbers, or will DVOA catch up to actual performance that the formula is missing?

Is yds/drive a sign that the offense really is good and will score more when they get more turnovers and better field position? Or is their turnover ratio a sign that the offense really is bad and will score less when they're not able to get as many yards per drive?

In other words, which stats are going to regress to the mean, and which stats are sustainable at current levels?

There may be 170+ posts here, but I haven't seen many that try to answer the real questions.

173
by james (not verified) :: Sat, 10/08/2005 - 1:29pm

james gibson,

One thing that might make DSR correlate better with pts/dr is that td and fgs count as successes in the formula being used.

Thats why I have been using yds/dr instead.

For the equations (yds/dr rank)-(pts/dr rank) I got 3,2.96, and 3 as the mode,mean, and median. Am I dillusional for thinking that makes a pretty strong correlation.

I realize yds and points in real life have little to do with each other but for this data set with mode, mean and median being so close that is why I thought that it was o.k to make a jump between the two ranks.

I figure since I am only using rankings(performacne related to rest of league) and not using an actual formula to correlate yds/dr to pts/dr that it makes sense.

A little help from someone with a better statistical background than me would be great.

Thanks

174
by karl nichols (not verified) :: Sat, 10/08/2005 - 2:48pm

Biggest boob owner?? Do you not know who Bill Bidwell is??

175
by Jeff F (not verified) :: Sat, 10/08/2005 - 4:25pm

Russ - With a sample size of six, in cases where the end result of the game is similar to a coin toss, likely favoring the better team slightly (As in, the better team will win somewhat more than half of their very close games), your one example fails to prove anything. You also didn't describe the circumstances the scores were achieved: maybe in three of the games, the opposing teams were 9 points down, and scored a TD with under 3 minutes to go, making the score, and game, look much closer than it was. Cases like that would have to be accounted for when making such an argument.

Given the Raiders record for the year, I wouldn't be surprised at all if none of those close scores, or maybe just one, were the Raiders edging out a close win, and if half or more of them were the result of the losing team scoring near the end of the game, while having a very small chance at winning.

So, if we were to consider 2 of those six games cases where the Raiders were playing from ahead, and the other four being very close games, then we will assign the Raiders a 90% chance of winning two of those games, and a 55% chance of winning the other close games: .9*.9*.55*.55*.55*.55 = .074 = 7.4% - so, there'd be a roughly 7.4% chance that the outcomes that all of those games would result in wins. While it wouldn't be expected, it would be within 2-3 standard deviations of expected wins, which puts it squarely in the "doesn't happen all that often, but is very plausible" camp.

Had the Raiders come back with 1:30 left on the clock in two of those games for a tight win, then, yes, there would be a much lower chance of winning all six of those close games. However, given the data you have given me, it would be much more likely for them to be the team that was ahead.

James - should you want your opinions to carry any weight, you need to show more than one year's results. The results of one year will fluctuate quite a bit, with some things, and not nearly as much with others. If you could sample, say, 10 years, and then use that to make your assertations, only then could they be taken seriously.

There have been PLENTY of fluky teams in the past (My beloved Pats in 01, as a primary example). With a schedule of all of 16 games, and luck being a significant factor, you will see occurances where the results don't "match up" with DVOA, because the statistics haven't caught up with the teams, yet. If you maintain X level of play, and win 50% more games than you expect to, if you continue the same level of play, your chances of winning future games is not indicated by your record, but by your level of play.

A totally *average* team could *easily* go 11-5 for a season. See, 2001 Pats, 2004 Falcons.

Right now, this article essentially says that if the Redskins continue playing at the level they are playing at, they will not win many games against opponents that play at a level above average. This even includes the fluky 3rd down numbers, of which 3rd down conversions for 10+ yards count VERY well for DVOA, but also, the preceeding two plays to get to 3rd and long counted against them. As has been mentioned before, the Skins have been converting 3rd and 10s at an unsustainable rate, not unlike rolling one die ten times and getting 5 6s (Chance of occurance is 1 in 200), - it's not that big of a deal as the sample size is too small.

176
by James Gibson (not verified) :: Sat, 10/08/2005 - 5:13pm

james - when I did this for 2003, I didn't count FGs as part of the way I calculated DSR (or as I dubbed it DSSR - down set success rate). If you look at my #s above, I also have one where I removed TDs and it's still better than yards/drive: 0.76 vs. 0.65. Turnovers adjusted yards/drive was 0.74. I tried combining turnover adjusted yards/dirve and DSSR without TDs or FGs and DSSR still wins out, although I don't know the exact #s because Excel crashed when I was doing that correlation and I haven't recalculated it yet.

177
by james (not verified) :: Sun, 10/09/2005 - 1:21am

james gibson,
That seems like a much better method. I forgot that you had changed the dsr equations. Thanks for your work.

Im going to disappear to the lab and try and fiddle with some more ideas.

178
by james (not verified) :: Sun, 10/09/2005 - 1:34am

jeff F,

Great points all around. I am looking for how magic seasons happen more than a measure of how an nfl team would play over 100 games of data.

To me, that doesn't capture how 16 game NFL seasons play out and that's what I am looking for. What am I searching for is teams that overachieve and what data shows an overachieving team vs. DVOA rank. 2001 Pats are an example of an overachieving team because of what the drive stats told us about that team, and are the biggest reason I started looking at drive stats.

THis is for a different forum but I have methods for choosing which games will be very high scoring and which will be very high scoring and you would be surprised with some of the results that have come out correct. Gambling wise I chose 2 games out of 14 wrong over/under wise. Maybe it can't be explained but drive stats show plenty about what can be expected out of teams in a short season run.

I don't want to know how good the team x would be if they kept playing that way for 100 games. I want to know if they can overcome their flaws for a playoff run, or a season run.

I would never argue that DVOA isn't the best measure for longer periods of time measurements.

Maybe I have neglected to realize what the arguement between myself and my f.o poster peers has actually been. ooops.

179
by skippyx (not verified) :: Sun, 10/09/2005 - 3:39am

A few problems with the Panthers of 2003.

1) They had no Eagles in their division.
2) They won games with missed extra points, blocked field goals, a coach not trying to win in the playoffs and a hurt Westbrook.

Good for them but no way can you find any logic supporting 2005 Redskins = 2003 Panthers.
Teams like the 2003 Panthers and 2001 Bears get destroyed the next year. Most similar teams never get the 11+ win season since they are not that good.

You can't gameplan on getting a 4th round rookie QB at home in game one when it is game 5 against the Chiefs in KC. You can't ask other teams to take a -1 run to set up a 47 yard field goal instead of picking up 7 and kicking a 40 yarder to beat you.

Relying on luck builds real nice casinos.

180
by DavidH (not verified) :: Sun, 10/09/2005 - 5:14am

James -

I will have some statistics to show you in a couple days, or at least some anecdotal evidance from the last 5 years. In the meantime, what are your over under bets for this week

I have Saints-Packers over 41.5 and Colts-49ers under 46.5.

181
by james (not verified) :: Sun, 10/09/2005 - 11:43am

skippyx,

I agree totally.

What are we both pointing out(though with opposite view points) are teams that can take a few lucky breaks and turn that into confidence and better season than what would have happened otherwise.

Teams like this begin to believe in themselves and the more everyone says they can't, the more they believe in themselves. Tha backbone of that is a stingier defense than most mediocre teams.

We both have the same point.

182
by james (not verified) :: Sun, 10/09/2005 - 11:45am

aaron,
taking me a while to have them finalized with all the injuries but right now I have

sea/st louis over 49 and indy-49ers under 47

183
by ThinkQuick (not verified) :: Sun, 10/09/2005 - 12:00pm

180+ posts and I'm not reading them all.

Dallas did NOT pass the ball well against Washington. Other than that flea flicker, they didn't do anything outstanding.

Seattle did NOT pass the ball well against Washington. They had one big drive the entire game and that was it. Washington also happened to be missing starters that game and that drive.

Washington's defense is better than DVOA is reflecting right now, although probably not at the level it was last year. It was ranked 9th last week, so those flukish couple drives Seattle made is definitely skewing their rating.

The fact that Washington's offense is 16th is pretty good. A 16th ranked offense and 10th ranked defense should be just enough for a wild card spot.

184
by james (not verified) :: Sun, 10/09/2005 - 12:27pm

Aaron,
gb/n.o- no play
tenn/hou- under
Car/az-under
pitt/sd-over
cinci/jax-under

Phi/Dal-just taking Phi

I do little more than take field position expected plus expected yardage gained per drive.

My numbers don't represent actual scores predicted. I have a system that evaluates scoring ability with 20 being the average.

What's considered average for each team changes according to talent but that is already factored into the total.

So basically my system makes every total 40. If both teams score over 20 I consider that an over play and vice versa. If one team scores over 20 while another scores under 20 then I just take the better team against the spread, and neglect the total.

I'm making an acception for the Indy under because I don't think San Francisco can score a TD today.

N.O/GB- 20/20 no play
Tenn/Hou-15.3/13.7
Car/Atl-18.7/18.7
Phi/Dal-27/18.7
Pitt/SD-22/22
Cinci/Jax-18/15.8
sea/stl-22/21.26
indy/sf-27/9.2

185
by DK (not verified) :: Sun, 10/09/2005 - 7:19pm

The Skins are down 14-10 at the half in Denver.

Total Yards: 249 Skins, 150 Broncos.
1st Downs: 16 Skins, 6 Broncos.
TOP: 19:07 Skins, 10:53 Broncos.

Skins are 4-7 on third downs.

The difference in this game is the -1 turnover diff.

There's still another half to go, but in the early going, the Skins haven't crashed and burned yet against a tough team on the road.

186
by james (not verified) :: Sun, 10/09/2005 - 10:18pm

Here's something that will never happen.

Dominate time of posession in the first half yet go into halftime losing. Teams don't win when they don't take advantage of opportunities ont the road.

Skins make too many mistakes. All the things that went their way in the first three games went against them.

Encouraging that for the third straight game they were driving as time was running down with the game solely in their hands.

Wish the conditions would have been different so we could have gotten a better idea of their 3rd down capabilities. They were 50% before the ball became impossible to handle.

187
by Joey (not verified) :: Mon, 10/10/2005 - 12:28pm

Coming in, I didn't think much of the 'Skins. Now, I actually believe they might be going someplace. Yes, they lost, but Denver's a tough place to play and they kept it tight throughout. They also dominated the statistics, which results in a win the vast majority of the time. I'm not jumping on the bandwagon, but now I'm at least going to aknowledge somebody out there might be building one.

188
by JonL (not verified) :: Mon, 10/10/2005 - 2:11pm

I know there's really no good way to quantify "coaching ability," but could red-zone success be a possible window into this? A team could have all the talent in the world, but if the coaches keep calling bad plays (like running into the line against a good run defense, for instance), they're not going to have much success.

Although if I'm going to be literal about it, a team with all the talent in the world would never be in the red zone, because they would score from their own 1 yard line.

189
by DavidH (not verified) :: Mon, 10/10/2005 - 2:41pm

I will have some statistics to show you in a couple days, or at least some anecdotal evidance from the last 5 years.

Reading this now, I'm not sure what I was talking about, but it really doesn't matter, as I'm going to be way too busy with work for the next week to do any kind of time-consuming football study.

The Redskins acquitted themselves well yesterday...

190
by cowboycrusher (not verified) :: Tue, 10/11/2005 - 6:34pm

I'm proud to stand by my team after their loss @ Denver.

They dominated one of the leagues "best" teams.

Once the redzone is figured out, Gibbs will have another ring.

-Hail Bitches

191
by Sid (not verified) :: Fri, 10/14/2005 - 11:39am

I was surprised by how well Washington did in Denver, I admit. But if you want something that points to Washington crashing and burning, you just have to look at the Seattle game. They should have lost to Seattle. They were hosting Seattle. Enough said.

192
by Sid (not verified) :: Fri, 10/14/2005 - 12:03pm

RE: 85

Two long bombs on Dallas in the 4th quarter. Yeah that got lucky against that all pro safety two times in a row. That Moss-Brunell combo is no good.

You have GOT to be kidding me. I thought intelligent people realized by now how overhyped Roy Williams is. He can't cover. Period. Especially on deep balls. He screwed up his responsibility on both plays. Williams lays hard hits on people, but is not always a sure tackler, and is terrible in coverage. He needs to be lined up near the LOS to be effective.
Washington got lucky twice to win that Seattle game. First the FG, and then winning the coin toss.

193
by Sid (not verified) :: Fri, 10/14/2005 - 12:21pm

james:

You want to talk about an undefeated team that is fraudelent. You need to look in Florida.

Tampa is easily better than Washington. Tampa is going to make the playoffs (at least I think so. In fact, I picked them as a wild card before the season). Not sure if I can say the same about the Snyderskins.

194
by Sid (not verified) :: Fri, 10/14/2005 - 2:25pm

RE: 119

I also thought the Vikings would win, and they still have a shot. If they stop renting boats and strippers and start focusing on football...

The Packers have no shot. They are almost a lock for last place, in my book. O-line banged up, defense sucks, running backs injured, Javon Walker done, etc.

195
by Mykhael (not verified) :: Sun, 10/16/2005 - 9:53am

I'm not gonna front, I've haven't really looked over what you base your DVOA ratings on and therefore can't say I understand it. But, I gotta believe the high ranking Giants and Bucs DVOA ratings must be off the charts. Since the Giants wins have come against 26th ranked Ari, 29th ranked StLou and 30th ranked NO.

And the Bucs loss coming against the 23rd ranked Jets who's QB had all of 2 weeks to prepare and a gift wrapped win courtesy of the replay officials overturning the correct call made by a referee against 27th ranked Det.

And since you believe the Skins were more lucky then good in defeating the Boys and Hawks I understand why you have the Skins lagging behind both these teams. But, I saw the Cowboys 33-10 ass whuppin they put down on the Eagles. I wouldn't attribute any of it to luck so why do the Cowboys lag 6 rankings behind them is a conumdrum to me as well.

I do like that you factor in ST in your ratings and truth be told. The Skins ST scare me too death. Our FG kicking has been crappy since 99, our coverage teams haven't been anything to be proud of, cept for a year here, a year there for the last 10 years. Nothing to write home about as far as our return teams and just to add some spice to the mix this year. Our punting has went AWOL with Tupa's injury as well. On an upside Chris Clemons did block a Den punt, which I guess kinda gets trumped by us having 2 FGs blocked.

Any Skins fans who claims not to be concerned with our lack of sacks, 3 for the year I believe and TOs, 1 INT, 1 FUM REC I'd have to say is straight trippin. But, in the most important aspect of defense, giving up points, I think our def has held up well surrendering 58 total points so far. The run def has shown some kinks the last 2 weeks, but I attribute that more to 3 plays, Alexsander's 34 yard run and Bell's 34 and 55 yard runs then to an overall breakdown.

Now I guess a case could be made that Dallas outplayed us. And they did have a missed FG. They had more plays then us, 1 play. More total yards, 5 yards. Time of poss. More 3rd down conv. factor in 4th down conv and we converted more and total 1st downs were a tie. They had more net passing yards due to their 5 sacks for -49 yards, but, Brunell passed for 291 yards to Bledsoe's 262. We had more rush yards. And here where's my homerism comes into play I believe. I think Brunell missed 2, possibly 3 TD plays, 1 to Moss from a Dal 29 yd line on our 1st possession that Brunell underthrew. 1 to Patten from Dal 38 yd line we were flagged for delay of game on the play. It was the play immediately after Moss's 41 yard reception. And a possible TD, a for sure huge gain on our 3rd possession from Dallas 35 yd line Brunell underthrowing Patten at around the Dallas 40, he had a couple steps on Henry.

As for Sea, yeah their FG kicker may have missed 2 FGs, but our rookie FG goal kicker was with the Skins a whole week before that game and our regular FG kicker, John Hall came up hurt after Chi. So you'll have to forgive me if I don't shed any tears over any FG kicking woes of our opponents.

And while they did have more total yards then us, by 2 yards. And more passing yards. We had more 1st downs, rushing yards, time of poss, total plays, 1st downs and way more 3rd dn conv.

I'd also like to add, unlike bad officiating calls that are out of the control of players and coaches and unfairly help or hurt a team. Missed FGs, however one wants to look at them fall in the same category as underthrown or dropped passes, missed blocks etc etc. No fan wants to see these things happen to their team, but, the law of averages and human nature have no concern about fans of sports team.

Instead of factoring missed FGs against teams and listing them as lucky. Maybe they should count against the teams that commit them in a lack or failure to execute category.

Still, though. It ain't no thang really. If people wanna call the Skins lucky, so be it. I personally don't believe in luck. I do believe that Gibbs is instilling within the Redskins the intangibles of never giving up and to continue to compete till the end. And building among Skins players a belief that they can be and
will be competitive with any team in the NFL.

These things can't be labeled as a stat or put in a computer for a percentage of times it will happen.
These things have been missing from the Redskins for a long time.

And whether it's because the Skins have become better or suddenly become lucky. Or the Eagles have become worse or suddenly snake bitten. Makes no difference to me. When other teams step on the field with the Skins now. They better be careful of puttin their trust in DVOAs or how it was last year.

Times they are a changin