Don't Overlook the Effect of Penalties

One of the more overlooked aspects of team performance is the tendency for being penalized. Penalty rate, defined as penalty yards per snap is one of the more reliably stable team stats. Compared to things like running or passing efficiency, the absolute size of penalty rate's effect is small, but because it tends to be consistent, it can be fairly predictive.

This Sunday's conference championship games feature teams on opposite sides of the penalty spectrum. ATL has by far the league's lowest team penalty rate at 0.21 penalty yards per snap. For context, the league average is 0.41, and the next best team is NYG at 0.29. ATL is 3.9 standard deviations better than the 2012 mean! SF is third worst in the league 0.46 penalty yds per snap.

On the other side of things is BAL. They're averaging 0.53 penalty yards per snap, the league's worst rate. That's 1.9 standard deviations worse than the mean. NE is tied for 6th best, at 0.39 penalty yds per snap.

How important a consideration is penalty rate on the outcome of a game? As it stands, the game probability model has the SF-ATL game at 55/45 in favor of SF and the BAL-NE game at 33/67. If I set all four team's penalty rates to the league average and rerun the probabilities, we can directly see how they're affecting the prospects for each team.

The SF-ATL game becomes a 61/39 contest, representing a swing of 6 percentage points. The BAL-NE game becomes a 36/64 match up, representing a smaller swing of 3 percentage points. The swing is larger in the NFC game partially because it begins closer to a 50/50 game, where smaller changes have bigger leverage.

Things are a little less extreme when we look at the actual impact of penalties on each team. The Win Probability Added (WPA) due to penalties for each team this season paints a slightly different picture. WPA is highly senstive to game context, so a 10 yard holding penalty on 1st and 10 from the 50 in the 1st quarter won't have nearly the same effect as the same penalty in the 4th quarter of a tight game. And in a blowout, many penalties would have zero effect on WP.

BAL, with its very high penalty yardage rate, have cost itself 0.50 WPA, (about half a win) in team penalties in the regular season, putting them squarely average in the league at 16th. But BAL benefited from opponent penalties slightly more, to the tune of 0.61 WPA (T 6th), for a net penalty effect of 0.11 WPA (10th).

NE had less of an impact due to own team penalties than BAL, but benefited less from opponent penalties. NE's own penalties cost just 0.19 WPA (best in the league), and benefited by 0.29 WPA (29th), for a net effect of 0.10 WPA (11th), right behind BAL.

SF cost itself 0.36 WPA (8th), which is relatively little considering it's high penalty yardage rate. SF benefited by 0.61 WPA, (T 6th with BAL). All together, that's a net of 0.24 WPA,

ATL, with its amazingly low penalty yardage rate, cost itself 0.36 WPA (7th), and benefited by 0.23 WPA (31st in the league, and not even as much as SF's net WPA). That's a net of -0.13 WPA for ATL, tied for 26th in the league. Tragically, despite playing highly disciplined football all season, ATL has suffered a net negative in terms of penalty effects on wins and losses.

But keep in mind that WPA is not terribly predictive. Teams really can't control other team's proclivities to be penalized. You'd think that might be the case to some degree. Fast WRs can draw pass interference calls, or monster DEs can draw holding penalties. But that doesn't seem to be the case. Opponent penalty stats don't correlate from week to week like own-team penalty stats do.

10 Responses to “Don't Overlook the Effect of Penalties”

1. Ian says:

I wonder if maybe the refs that typically work the games in Atlanta just don't call as many penalties as other ref teams? That could explain the low WPA effects that Atlanta has experienced.

2. Anonymous says:

Ian, refs travel every week, and often don't ref at the same city twice in a season.

3. Anonymous says:

Did you consider that there may be an actual optimal rate of penalties.we know that multiple penalties occur in each game that are not noted. Never committing an actual penalty may be worse than be occasionally penalized for some while getting away with others.

4. Anonymous says:

If you never commit a penalty you should likely be playing more aggressively. It would help to break down the penalties. False start is likely never advantageous, however, pass interference and holding may have an optimal rate of occurrence.

5. Anonymous says:

If they are playing too passively with regard to penalties, that should show up in lower passing/rushing/sack numbers. Given a specific rate of performance, fewer penalties is always better.

6. Anonymous says:

fair points. Some penalties are nothing but bad (like false start), but some penalties are preferable (pass interference as opposed to a touchdown, holding as opposed to a sack, intentional grounding, etc).

one confusing thing, the article states how the penalty rate makes a very big difference in the Game Win Probability, but then goes on to say how the various teams benefit from net penalties.

How does GWP not take that into account?

SF/ATL has a 6 point swing because ATL penalty rate is low and SF is high (ATL has a 6 pt advantage). But, SF has a net penalty advantange of +.24 WPA, while ATL has a net penalty disadvantage -0.13 WPA. So how does SF gain 6% when penalties are removed?

7. Brian Burke says:

Anon above,

The difference between what WPA is sayinga nd what the penalty rate in GWP is saying is due to context. WPA is very context-dependent We can't predict when or in what context penalties will occur, only that there will be certain tendencies of each team to have a lot of penalty yards or not.

8. Anonymous says:

thanks Brian,
that makes sense. And it also highlights the difference between 'looking back' and 'projecting forward'.

9. Wizard says:

Brian, do you have the year to year correlations of penaltry rates (you mentioned they were fairly predictable).
thanks

10. Anonymous says:

What has the Ravens GWP stats been in the postseason and how are they different from the regular season stats (or are they winning more on luck)