Writers of Pro Football Prospectus 2008

Most Recent FO Features

ReplayIns14.jpg

» Scramble for the Ball: Getting it Right?

Instant replay review is one of the cornerstones of the modern NFL. The process and its myriad special rules have been internalized and constantly debated. Mike Kurtz wonders: is it worth it?

20 Jan 2006

2006 Championship Previews

by Aaron Schatz

For the final three games of the season, we're putting the championship previews on FOXSports.com. Don't worry, they are the same complex FO previews you know and love. Even the colorful week-to-week DVOA charts are there.

We've created one thread to discuss the actual previews themselves, or you can discuss the games in the separate threads we've set up for each conference championship:

While I've got this space, however, I wanted to answer a couple of mailbag questions, one related to each game.

I understand the reversal on the Broncos pass defense from a DVOA approach but am a little confused as to how the run defense is also reversed. While I concur that the NFL statistical rankings based on yardage are infantile I still can't get past the rushing defense as "middling" being accurate for the Broncos. If weighted adjustment allows for the longer runs being carried out against "prevent type" defense settings where the Broncos have held comfortable leads as opposed to say the desired goal of acheiving a first down does this impact the analysis?

Rob Visser - Port of Spain, Trinidad

I thought this would provide an interesting look at how DVOA differs from the straight NFL ratings, so let's work through this gradually.

  • The Broncos allowed 1363 yards this year on the ground. Rank: #2.
  • But they also faced just 344 rushing attempts on the season, the lowest in the league by a hefty margin. This is what happens when you are 13-3. (Kansas City was second, facing 383 rushing attempts.)
  • So the Broncos allowed an average of 4.0 yards per carry. Rank: #15.
  • Actually, this includes some plays that I don't count as runs in my numbers, including quarterback kneels and aborted snaps. Take those out, and Denver allows slightly more yards per carry, but the Broncos move ahead of the Jets and 49ers. Rank: #13.
  • However, when the Broncos fixed their red zone defense, they didn't fix the defense against the run. Denver gave up 3.6 yards per red zone carry, the worst figure in the league.
  • And they fixed their third-down pass defense, but not their third-down run defense. Denver allowed conversions on 67 percent of opposing third-down runs, the worst figure in the league.
  • Denver also caused just three fumbles on running plays.
  • When those three things are considered, Denver's run defense VOA spits out as -0.1%. Rank: #20.
  • Of course, Denver plays in the same division with the teams ranked second (San Diego), fifth (Kansas City), and tenth (Oakland) in rush offense DVOA. So when you add in the opponent adjustments, you get -3.3% DVOA. Rank: #17.

The other question was asked in a discussion thread two weeks ago, and I can't remember who asked it or even which thread it was. Anyway, somebody wanted to know how the late-season surge by the 2005 Carolina Panthers compares to the late-season surge by the 2003 Carolina Panthers. That chart below has both seasons, game by game.

If you are interested in the NFC Championship game, you also might want to check out this long, extensive interview I gave to Seahawks.net.

Posted by: Aaron Schatz on 20 Jan 2006

26 comments, Last at 22 Jan 2006, 4:05pm by charles

Comments

1
by PerlStalker (not verified) :: Fri, 01/20/2006 - 10:59pm

Ahhh! Intelligent, witty comentary. How long I have waited for thee. Long may FO reign. :-)

I'm so very much looking forward to the AFCCG. It'll be football the way it should be. Unless it's not. I think that game will come down to the adjustments both coaches make.

Here's to a good game. /me raises his glass

2
by Crushinator (not verified) :: Fri, 01/20/2006 - 11:10pm

An interesting note that you made about Denver was that they're good at shutting down the opposing tight end but not terrific on the number 1 WR. How much of this has to do with playing in a division with Antonio Gates and Tony Gonzalez? If you remove those games, does Denver's defense against #1 receivers noticably increase?

3
by Return of the son of the FO Message Board Curse (not verified) :: Fri, 01/20/2006 - 11:22pm

Crushinator, maybe it would be more interesting if they changed the games (4/16) where they faced Gonzo/Gates to make the TE the #1 WR. I don't know if the Broncos shift responsibility for the TE when they face those two, though I remember someone who drafted Gonzo really high being pissed that Champ Bailey was all over his ass.

Here's to a good game...I have no glass to raise as I drink exclusively from funnels or keg stands.

4
by shonk (not verified) :: Fri, 01/20/2006 - 11:27pm

Fourteen years later, it turns out that Denver did get some value out of their 1992 draft after all.

Brilliant.

5
by Yaxley (not verified) :: Fri, 01/20/2006 - 11:40pm

Wow, these previews may be numbers-heavy, but they sure do reward not skimming with some witty lines. And the football analysis isn't shabby either.

6
by ernie cohen (not verified) :: Fri, 01/20/2006 - 11:49pm

Okay, maybe I'm misreading this, but I'd swear from the CHI and PIT DVOA graphs that you have CHI with a higher DVOA than PIT for their week 14 game that PIT won 21-9. (I'm noticing this just now because of the similar phenomenon in the PIT-IND playoff game.) Now I can understand coming to this conclusion for the IND game, but for the CHI game, this is just nuts. PIT dominated statistically, and I don't think anybody thought that this game was close. Is there something funny going on here?

Another DVOA question: Suppose a good team and a bad team play, and execute identical plays on O and D, resulting in a tie. Because of adjustments for the quality of opposition, won't the bad team finish the game with a somewhat higher DVOA? This adjustment would seem to indicate that DVOA says that IND played better than PIT by an even wider margin.

7
by calig23 (not verified) :: Fri, 01/20/2006 - 11:51pm

Young fans in Pittsburgh are all wild for that latest literary classic, Choose Your Own Adventure: Antwaan Randle El Returns a Punt.

Hah!

8
by BOOMAN (not verified) :: Sat, 01/21/2006 - 12:38am

Finally, the Broncos will lose big time to the unstoppable COLTS in their scary HOMEDOME! How exciting for COLTSHOMERS! Just like the COLTSHOMERS have been discussing since, oh, week eight or so! BRING IT ON, BRONCOS!!!!!
We'll see you Broncos in Indianapolis, MWaahahahahahahahahahahahahahahaahahah!!!!!!

9
by Pat (not verified) :: Sat, 01/21/2006 - 1:00am

Another DVOA question: Suppose a good team and a bad team play, and execute identical plays on O and D, resulting in a tie. Because of adjustments for the quality of opposition, won’t the bad team finish the game with a somewhat higher DVOA?

Depends. DVOA doesn't lower all numbers for a team playing a weak team. If you've got a team which has a good run defense and a terrible pass offense, and you pass like mad against them - yah, DVOA will say "please, my mother could do that." But if you run, run, run, your numbers might actually be higher. And it's actually more situational than that - if you read the DVOA description, you'll see that the adjustment is very situational ("2nd and 10 passes to a tight end when down by more than a touchdown when in the BACK zone").

10
by MAW (not verified) :: Sat, 01/21/2006 - 1:01am

Re #4:

Aaron's final line sent me off a-Googling. I had no idea Maddox has been in the league that long.

11
by Pat (not verified) :: Sat, 01/21/2006 - 1:03am

Oh, and I would like to point out that those two Carolina graphs are freakishly similar. I mean, really, really, similar. I mean, you could probably write a computer program to recognize "2003 Carolina" when fed the 2005 graph. Buildup as the season begins, inconsistency as injuries build up, then a ridiculous surge at the end of the season.

12
by Jake S. (not verified) :: Sat, 01/21/2006 - 1:05am

What I don't understand is why you can't convert your predictions to a most likely score, even if it has decimals.

13
by Kibbles (not verified) :: Sat, 01/21/2006 - 2:01am

Note that Denver, for some reason, never, ever passes to tight end Putzier in the red zone.

It's because passing to him on the sidelines won't do Denver's offense much good. Putzier is the first person who gets pulled off the field when Denver hits the red zone, usually replaced with Kyle Johnson. It's a blocking thing.

14
by Sergio (not verified) :: Sat, 01/21/2006 - 3:16am

Re:11

Sure, but '03 Carolina never demolished a team quite like the '05 did...

Still, Aaron's right. It's going to be hard for them to outscore Seattle straight on...

15
by Jake S. (not verified) :: Sat, 01/21/2006 - 3:32am

Oh, and I just remembered that both teams have played many close games against very bad teams. Carolina has even lost some of these games? New Orleans?? Detroit! San Fran!!

Weird stuff, just sayin'.

16
by Will Allen (not verified) :: Sat, 01/21/2006 - 4:06am

I've been looking for the black and blue report, because the injury picture on Seahawks/Panthers has me straddling.

17
by RowdyRoddyPiper (not verified) :: Sat, 01/21/2006 - 5:21am

The freakish thing about the graph, if it is to be believed, is that Carolina appears not to have had a bye in 2003...maybe they insisted on playing 17 games to make themselves tougher?

18
by shonk (not verified) :: Sat, 01/21/2006 - 6:41am

Re: #10

Well, he hasn't, entirely. He didn't play a down in the NFL from 1996-2000, and didn't even attend any training camps in '98, '99 or '00. Apparently he ran an insurance company for a while before going to the Arena League in 2000, which led to the XFL in 2001, and I think we all know the story from there. That's why he's only credited with 9 years NFL experience despite having been drafted 14 years ago.

19
by admin :: Sat, 01/21/2006 - 10:46am

Sorry, no Black and Blue this week. It's impossible to get info about Peppers and the Alexander report is in the preview.

I must have forgotten to empty the cell with the 2003 bye so it came out 0%. One of those 0% weeks is the bye week.

20
by Born a Bronco Fan/Die a Bronco Fan (not verified) :: Sat, 01/21/2006 - 10:53am

Aaron,

Awesome game preview for the AFC! It is great to read in depth comparisons that make sense and are not just full of personal opinions. Hope it is as good a game as the numbers say it is going to be.

One question, do you have the stats on how Denver's offense does vs. the 3-4 scheme? I know they stuggled against Dallas but Dallas flipped in and out of it often. Does Pittsburg, ever give the 4-3 look or is that something you don't break down--only performance not schemes?

I just thought that might be a tell-tell for this game because their are many differences between the two.

Go Broncos!
By the way, who is going to win in the AFC?
I believe the Seahawks will win but I will stay off the AFC game-don't want to jinx my team.

21
by jeff t (not verified) :: Sat, 01/21/2006 - 11:41am

Bronco Fan,

To answer your question, no the Steelers don't use the 4-3 look. However, they will use 4 down lineman in some of their nickel and dime formations. Usually its LBs Porter and Haggans lining up at DE. DVOA, as far as I know, doesn't take into account formations and personel packages. So, it would be impossible to calculate Denver's numbers based on 4-3 vs 3-4, etc. But, one could look at the overall trend vs 3-4 and 4-3 teams. It looks like the Broncos have played pretty well against other 3-4 teams.

As an aside, a Steeler fan (biased much?) on another web site looked at the Broncos conventional statistics vs. physical teams (subjective choices)and found that Denver did not fair well, especially in the running game. Obviously, there are a host of problems with this type of analysis (conventional stats, no context, etc,) but I though it was interesting anyway.

This should be a great game with two very similar and evenly matched teams.

Here we go Steelers, here we go.......

22
by Catfish (not verified) :: Sat, 01/21/2006 - 11:55am

Re: 12

I don't think Aaron can't, so much as he doesn't want to. He has stated before that he doesn't want this to be a gambling site, and an "Expected Score Differential" number would bring the gamblers in droves.

23
by luz (not verified) :: Sat, 01/21/2006 - 12:49pm

#4

agreed. it took me a second but when i finally figured it out, i couldn't stop laughing. i think that was my favorite FO joke of the year.

24
by Sid (not verified) :: Sun, 01/22/2006 - 2:10am

Too bad Foxsports doesn't have nice printable versions like FO. They include other trash on the page.

Nice work as always. I heard you on Friday morning (streaming online, I think it was the Zone).

25
by Sid (not verified) :: Sun, 01/22/2006 - 2:31am

It seems many ESPN.com voters (overall) can see past the hype and realize the Denver D is actually stronger against the pass. Click my name.

26
by charles (not verified) :: Sun, 01/22/2006 - 4:05pm

Re: 22

Nope. I believe (and this is just speculation) that Aaron does in fact calculate expected score differential or similar and uses it to wager on games. While one non-ecommerce website and its advertising might not be enough to support the Schatzes, the revenue from being able to predict even 65% of games against the spread absolutely would.