Are defensive-trained head coaches more conservative than offensive-trained ones?

One of the fascinating things to consider with head coaches is how they got to where they are, and what impact their background has had on them. While head coaches are responsible in the final sum for both their offence and their defence, most have come up primarily from one side of the ball or the other. A few, like current Hamilton special-teams coordinator and former Winnipeg head coach Jeff Reinebold, have worked on both the offensive and defensive sides of the ball, but they're exceptions. The current list of CFL head coaches has four guys who have primarily worked on the offensive side of the ball (John Hufnagel, Scott Milanovich, Kent Austin and Jim Popp; Popp's mostly been in personnel, but when he has coached, he's served mostly in offensive roles) and four who have mostly worked on defence (Mike Benevides, Kavis Reed, Corey Chamblin and Tim Burke), so it's not that either side of the ball necessarily prepares you for a head-coaching job better. However, those different backgrounds can have a substantial influence on how you coach.

The impact of those backgrounds is felt in many ways, as head coaches often tend to work more closely with the unit they came up through, supervising the coordinator there in a more hands-on fashion or sometimes even acting as their own coordinator and/or calling plays. An interesting element that's not normally mentioned is the differing philosophies that tend to exist on offence and defence, though. NBC Sports' Joe Posnanski has an intriguing take on that in a piece on Carolina Panthers' coach Ron Rivera (a former star linebacker who came up through the coaching ranks on defence), and it's one with CFL implications. Here's the key part of what Posnanski wrote:

Obviously, it's oversimplification to say that all offensive coaches are aggressive by nature and all defensive coaches are conservative. In fact, you can make a case that the blitz-happy styles of coaches like the NFL's Rex and Rob Ryan or the CFL's Casey Creehan are more aggressive than the West Coast offence approach favoured by people like former CFL and current NFL head coach Marc Trestman. However, by and large, it would seem reasonable to argue that coaches with an offensive background are more willing to gamble on things going right for the offence, while coaches with a defensive background are more concerned that their offence will make mistakes. Does the data support that, though? Well, thanks to friend of the blog Rob Pettapiece, we have some very interesting answers.

Perhaps the easiest way to judge how aggressive a coach is is by looking at their decisions on third down. Previous research from Pettapiece suggests CFL teams should go for the first down more often instead of punting or attempting field goals (similar studies have been done south of the border) , but what we're really interested in here isn't whether going for it is the right decision (or if it even worked in a given case), but simply how often various coaches elect to go for it and if their background has any impact on that. To investigate that, Pettapiece pulled up data from third and one, third and two and third and three situations in 2012 (the most recent complete season of data) and analyzed how many times each coach went for it. (Obviously, there are third and inches and third and longer-than-three situations to consider too, but they're excluded here: the call on third and inches is almost always to go for it, and the call on third and longer-than-three is usually to punt, unless it's late in the game and you're losing.) Here are the results. Let's start with the five coaches who were head coaches for the whole season last year and are still the head coach of their team:

In that analysis, the offensive coaches went for it on 36 per cent of their total opportunities, while the defensive ones went for it 30 per cent of the time. This isn't a huge sample size, and time/field position/score effects aren't considered here (for example, even the most conservative coach will likely go for it on third and three when down by seven with less than a minute left, while even the most aggressive coach may opt to punt or kick a field goal on third and one if winning by 20 late; those time/field position score effects may affect why defensive coaches here were actually more likely to go for it on third and three than offensive ones), so it's far from a conclusive result. It is interesting, though, and it would seem to support Posnanski's argument about coaches' backgrounds influencing their decisions.

What about the two guys who were CFL head coaches for a full season last year and aren't now, George Cortez (offensive background, Hamilton, fired after the season, now the Riders' offensive coordinator) and Marc Trestman (offensive background, Montreal, left for the NFL's Chicago Bears)? Well, data on their 2012 decisions follows this pattern too. Pettapiece combined the 3rd and 1, 2 or 3 results for them and found they went for it 27 times in 73 opportunities, or 37 per cent of the time, higher than even Hufnagel and Milanovich. Paul LaPolice, another guy with an offensive background, was fired by Winnipeg after eight games, but went for the first down seven times in 21 chances during those games (33 per cent).

Oddly enough, the main outlier here is the guy who replaced LaPolice for the final 10 games last season, Tim Burke, who has a defensive background. Burke went for it 17 times in 27 opportunities last season, 63 per cent of the time. However, it's notable that Burke's Bombers suffered several horrific blowouts, including a 52-0 loss to Saskatchewan, a 44-3 defeat against Calgary and a 29-10 pummelling from Toronto. Being on the wrong end of games like that tend to increase the percentage a coach goes for it, as the risks of failure are lower (you're already getting clobbered) and it suggests you're still trying. (He also has a smaller sample size than everyone but LaPolice, which may swing things.) It certainly wouldn't seem that Burke's all that aggressive of a coach in general, particularly when you consider moves of his like taking a knee Friday.

Overall, more extensive research, a larger sample size and a way to consider time/field position/score effects is probably needed to see if offensive coaches do tend to be more aggressive than defensive ones on the whole. It's an intriguing theory from Posnanski, though, and one that makes sense intuitively. A brief glance at data from the CFL in 2012 suggests it might not be far from the truth.

(Thanks to Rob Pettapiece for the research help here. You can find more of his Canadian football research in the CFL category on his blog.)