Since i can't watch the game today I had some free time. So for the last 4 years here is a scatter plot for winning percentage vs yards per penalty. The idea is that if mental error penalties (false starts, illegal formations) had an effect on winning then we would see teams with lower penalty yards per penalty have a losing record.
However, there seems to be no pattern.
Just for completeness, here is winning percentage vs number of penalties per game
That huge outlier for 2021 is Toledo. Tennessee this year is the undefeated outlier. It could be said that if you want to be undefeated you need to have less penalties overall, but what those penalties are doesn't really matter.
Forums:
DISCLAIMER: Forum topics may not have been written or edited by The Key Play staff.
Comments
I once saw a story about predictors of winning and losing in football. It was very eye opening to what was more significant. For instances turnovers, ball control, and rushing attempts were more precise predictors of Win / Loss than penalties. Great work...Your science checks out....
some of these things are chicken-egg... if youre already winning big, you probably start running the ball and chewing up clock. Teams that rely on running a rushing-based ball control offense have such a thin margin for error, need to be really highly efficient on red zone possessions (both offense and defense) and avoid turnovers
Yep. The 5 factors Bill C identified were:
Success Rate is far and away the most important indicator (statistically speaking). Here's a piece on it.
FTFY
Obligatory (or at least it should be):
Now I have to go draw constellation tonight!
IMO, penalty data is useless unless it's segmented by type. You need to at minimum separate procedural penalties (False starts, illegal motions, too many men on the field, etc) from other penalties. My guess is that procedural penalties are correlated with bad teams, and other penalties have no correlation.
Bill Connelly was asked about the relationship between penalties and success on a recent episode of the Solid Verbal (at about the 43:00 mark):
*Above transcript was lightly edited
Procedural penalties are 5 yards. So typically teams that have the lowest yards per penalty are committing those types of penalties more often, or they are committing penalties in the redzone/ behind the 20.
Ahh I see. Creative idea!
The other issue with your data is that you're using winning percentage as the output. Tough to compare 8-1 Tulane to 8-1 Tennessee (who has played UGA, Bama, LSU, etc). Something like SP+ rating would be better since it accounts for difficulty in schedule.
Serious question, does difficulty in schedule really matter? I think it's another data point to look at but the whole "do the types of penalties affect winning?" doesn't matter for how hard the wins are.
Tulane and Tennessee are on opposite spectrums of committing penalties, but Tulane is 8.4 ypp and Tenn is 8.5 ypp. So arguable that Tulane commits similar types of penalties, they just do it at less than half the rate that Tennessee does.
It sounds like you are trying to test the hypothesis that procedural penalties result in worse outcomes. However, there's a lot of things that impact an 'outcome'; so you want to try to control for other factors (so you're comparing apples-to-apples). There's A LOT of things that factor into 'does a team win' - One is how 'good' a team is, and another is how 'good' their opponent(s) are.
One thing you could do is use a metric 'normalizes' outcomes, like SP+. Another thing you could do is segment the data so you're only comparing 'apples to apples' (eg; instead of mapping all of college football, just map the SEC, because all teams play similar teams).
Let me put it this way... If Tennessee played Tulane's schedule, they could probably get a lot of penalties, and still have 8 or more wins (because they have so much talent on their team). Conversely, Tulane could not win 8 games with Tennessee's schedule, no matter how few penalties they have.
Because strength of opponent will always have a larger impact on if you win a game than number of penalties, you need to adjust for that somehow.