Eagles, Packers, Ravens, and Saints all in for worse sea (1 Viewer)

[benson_will_lead_the_way]

19 of the 25 teams with a turnover ratio of at least +10 in the past 5 years had at least 3 fewer wins in the following season.

Just an interesting statistic.

 

[Chase Stuart]

An interesting statistic, but not surprising to those who think turnovers are largely random.

Here's a look at all teams with a +10 differential in turnovers in any year since 2000, along with their winning percentage in that season, their turnover differential in that season, their winning percentage the next year (N+1), their win differential in the next year, and their number of turnovers in the next season:

year tm win% turn N+1 win diff N+1 turn2009 GNB 0.688 24 ? ? ?2009 PHI 0.688 15 ? ? ?2009 NOR 0.813 11 ? ? ?2009 BAL 0.563 10 ? ? ? 2008 MIA 0.688 17 0.438 -4 -82008 TEN 0.813 14 0.500 -5 -42008 BAL 0.688 13 0.563 -2 102007 SDG 0.688 24 0.500 -3 42007 IND 0.813 18 0.750 -1 92007 NWE 1.000 16 0.688 -5 12007 TAM 0.563 15 0.563 0 42007 SEA 0.625 10 0.250 -6 -72006 BAL 0.813 17 0.313 -8 -172006 STL 0.500 14 0.188 -5 -102006 SDG 0.875 13 0.688 -3 242005 CIN 0.688 24 0.500 -3 72005 DEN 0.813 20 0.563 -4 02005 CAR 0.688 16 0.500 -3 -52005 IND 0.875 12 0.750 -2 72005 NYG 0.688 12 0.500 -3 02005 JAX 0.750 11 0.500 -4 12005 SEA 0.813 10 0.563 -4 -82004 IND 0.750 19 0.875 2 122004 NYJ 0.625 17 0.250 -6 -62004 SDG 0.750 15 0.563 -3 -82004 CAR 0.438 12 0.688 4 162004 BAL 0.563 11 0.375 -3 -102004 PIT 0.938 11 0.688 -4 72004 BUF 0.563 10 0.313 -4 42003 KAN 0.813 19 0.438 -6 -62003 NWE 0.875 17 0.875 0 92003 TEN 0.750 13 0.313 -7 -12003 SFO 0.438 12 0.125 -5 -192003 MIN 0.563 11 0.500 -1 12003 IND 0.750 10 0.750 0 192002 TAM 0.750 17 0.438 -5 22002 GNB 0.750 17 0.625 -2 02002 KAN 0.500 16 0.813 5 192002 PHI 0.750 14 0.750 0 42002 OAK 0.688 12 0.250 -7 -12002 ATL 0.594 12 0.313 -4.5 02002 JAX 0.375 12 0.313 -1 -42002 SFO 0.625 10 0.438 -3 122001 NYJ 0.625 18 0.563 -1 42001 TAM 0.563 17 0.750 3 172001 SFO 0.750 15 0.625 -2 102001 CHI 0.813 13 0.250 -9 -72001 GNB 0.750 12 0.750 0 172001 DEN 0.500 10 0.563 1 -52000 BAL 0.750 23 0.625 -2 -82000 DEN 0.688 19 0.500 -3 102000 OAK 0.750 17 0.625 -2 -12000 TAM 0.625 17 0.563 -1 172000 MIA 0.688 15 0.688 0 -102000 PIT 0.563 14 0.813 4 72000 DET 0.563 11 0.125 -7 -16 0.689 14.7 0.528 -2.6 1.8
The last line there shows the team averages of the teams from '00 to '08; they went from a .689 winning percentage (essentially an 11-5 team) to a .528 winning percentage (8.5 wins), and a large reason for that is dropping from a +14.7 turnover margin to a +1.8 turnover margin.

We can also look at the results in reverse, examining teams there were -10 in turnover margin:

year tm win% turn N+1 win diff N+1 turn2009 WAS 0.250 -11 ? ? ?2009 CLE 0.313 -12 ? ? ?2009 OAK 0.313 -13 ? ? ?2009 STL 0.063 -13 ? ? ?2009 DET 0.125 -18 ? ? ? 2008 HOU 0.500 -10 0.563 1 -12008 DAL 0.563 -11 0.688 2 22008 DEN 0.500 -17 0.500 0 72008 SFO 0.438 -17 0.500 1 92007 STL 0.188 -10 0.125 -1 -52007 KAN 0.250 -11 0.125 -2 52007 OAK 0.250 -11 0.313 1 12007 SFO 0.313 -12 0.438 2 -172007 HOU 0.500 -13 0.500 0 -102007 BAL 0.313 -17 0.688 6 132006 TAM 0.250 -12 0.563 5 152006 CLE 0.250 -15 0.625 6 -22006 OAK 0.125 -23 0.250 2 -112005 STL 0.375 -10 0.500 2 142005 BAL 0.375 -10 0.813 7 172005 ARI 0.313 -11 0.313 0 32005 GNB 0.250 -24 0.500 4 02005 NOR 0.188 -24 0.625 7 -42004 CLE 0.250 -12 0.375 2 -72004 GNB 0.625 -14 0.250 -6 -242004 DAL 0.375 -15 0.563 3 -52004 OAK 0.313 -17 0.250 -1 -42004 MIA 0.250 -17 0.563 5 12004 SFO 0.125 -19 0.250 2 -92004 STL 0.500 -24 0.375 -2 -102003 CLE 0.313 -11 0.250 -1 -122003 SDG 0.250 -11 0.750 8 152003 ARI 0.250 -13 0.375 2 12003 BUF 0.375 -16 0.563 3 102003 NYG 0.250 -16 0.375 2 42002 ARI 0.313 -10 0.250 -1 -132002 BUF 0.500 -12 0.375 -2 -162002 WAS 0.438 -14 0.313 -2 22002 CIN 0.125 -15 0.500 6 22002 MIN 0.375 -18 0.563 3 112002 STL 0.438 -19 0.750 5 72001 STL 0.875 -10 0.438 -7 -192001 MIA 0.688 -10 0.563 -2 02001 IND 0.375 -13 0.625 4 -52001 BUF 0.188 -14 0.500 5 -122001 DET 0.125 -16 0.188 1 -72001 MIN 0.313 -21 0.375 1 -182000 MIN 0.688 -10 0.313 -6 -212000 STL 0.625 -10 0.875 4 -102000 DAL 0.313 -14 0.313 0 -92000 CIN 0.250 -14 0.375 2 -92000 ARI 0.188 -24 0.438 4 -32000 SDG 0.063 -28 0.313 4 2 0.348 -14.9 0.451 1.6 -2.5
These teams showed similar turnover differential regressions, going from -15 in turnover margin to -2.5; the teams weren't as "bad" as the other teams were "good", so the regression isn't as strong when it comes to records. They went from a 0.348 winning percentage (5.6 wins) to 0.451 (7.2 wins). The five teams last year were all bad teams with ugly turnover differentials (i.e., 5 wins or fewer). If you look only at the teams on the above list with 5 or fewer wins, they had an average winning percentage of .234 (3.7 wins) with a turnover differential of -15.8; the next year, they improved to a .393 winning percentage (6.3 wins), or increased by about 2.5 wins. They also improved their TO margin to -2.9.

 

[Mustang Man]

That is interesting. I would venture to say 2 of them 4 teams will not make the playoffs.

 

[Dinsy Ejotuz]

Chase, are the relationships stronger if you just look at the TO margin for fumbles lost and won? I think I've read that pass defense and QB quality are a lot more consistent than the result of fumbles (which IIRC is essentially random after the ball is on the ground). I'm wondering if that shows up here?

 

[Chase Stuart]

Chase, are the relationships stronger if you just look at the TO margin for fumbles lost and won? I think I've read that pass defense and QB quality are a lot more consistent than the result of fumbles (which IIRC is essentially random after the ball is on the ground). I'm wondering if that shows up here?

I've looked at the issue, and concluded that QB INTs are almost entirely random, and I think by extension, DEF INTS are pretty much entirely random.http://subscribers.footballguys.com/2009/0..._qbintrates.php

I think fumbles are pretty close to random, too, while fumble recoveries are about as random as anything in football. That said, there is some evidence that fumbles aren't entirely random, and that they're correlated with good players like Adrian Peterson: http://www.pro-football-reference.com/blog/?p=997

 

[baconisgood]

Chase, are the relationships stronger if you just look at the TO margin for fumbles lost and won? I think I've read that pass defense and QB quality are a lot more consistent than the result of fumbles (which IIRC is essentially random after the ball is on the ground). I'm wondering if that shows up here?

I've looked at the issue, and concluded that QB INTs are almost entirely random, and I think by extension, DEF INTS are pretty much entirely random.http://subscribers.footballguys.com/2009/0..._qbintrates.php

I think fumbles are pretty close to random, too, while fumble recoveries are about as random as anything in football. That said, there is some evidence that fumbles aren't entirely random, and that they're correlated with good players like Adrian Peterson: http://www.pro-football-reference.com/blog/?p=997

A thought for you QB INT piece that could strengthen or weaken your conclusion. If defensive INTs aren't random that could correlate to making QBs interception rates appear random thanks to changing Schedules each year. To explore this could be to simply run a regression using only #s from divisional games.

 

[Michael Fox Fan]

Great posts Chase, not surprising but interesting nonetheless.

 

[Dinsy Ejotuz]

Thanks for the reply Chase - very interesting stuff. I'd missed that article.