Sort of OT: Use of statistics in football
As someone who is more of a casual fan of football, I see posts on this board that rely heavily on statistical analysis of the games. I was just wondering if coaches use the same types of analyses when preparing for games, or evaluating players, or is it something that this board has perfected because there are a lot of smart statisticians and engineer-types who just enjoy these kinds of exercises? If coaches do use these types of schematics, how do they help?
September 22nd, 2010 at 9:31 PM ^
The famous "4th down card" probably gives some insight to coaches using statistics. I doubt they use super intricate things, but there's probably grad assistants tracking at least crude tendencies.
September 22nd, 2010 at 10:26 PM ^
the "Going for 1 or 2" chart? That one seems to be used by everybody.
September 22nd, 2010 at 9:35 PM ^
But I think that there are a number of "math nerds" like me on this board (I'm working on a Ph.D in physics...reading between the lines, I'm guessing that we have some computer science guys and engineering types around too. Perhaps even a few mathematicians.)
Whenever I try to attack a problem, I try some sort of mathematical model. And as we have "lots" of games, statistics is the sensible approach. Most problems like this have been (mostly) solved elsewhere, so it isn't like we are reinventing the wheel here. But the amount of computing power we have at our fingertips and the "easy" accessibility of game data makes this kind of activity something that we can do in our spare time.
Now as far as your question about what coaches do...let's take an example from baseball: Sabermetrics (Wikipedia). Some managers/GMs probably do this (I think that Theo Epstein of the RedSux uses Sabermetrics) but others don't. I saw a 60 minutes interview where the creator of Sabermetrics said "There is no such thing as a clutch hitter," (according to his math). Yankee fans might disagree when trying to forget what David Ortiz did to them...Back to football. Yes, there are people who do this stuff, at least for fun or for books they are writing (Romer). Probably there are coaches who actually pay attention to this numbers stuff, at least to a limited extent. Simple math is quite prevalent...for example, you've just scored a TD to go up by 1 pt. with 1:30 left in the game. You decide to go for 2 because being up by 2 means that a FG wins the game. But, especially in the game of football, sometimes there is nothing that can replace instinct. I think that MSU vs. ND is a prime example of this. On 4th and 14 from the ND 29, MSU should have kicked the g**d*** field goal because while failure=loss in both cases, the probability of success was higher for a field goal. But Dantonio followed his instincts, and the rest, as they say, is history.
September 23rd, 2010 at 12:39 AM ^
Just an added comment on sabermetrics.
Despite what Joe Morgan would have you believe, Moneyball was not written by Billy Beane's computer but by Michael Lewis (who also wrote Blindside, a much better book than movie). Moneyball focused on Billy Beane, GM of the Oakland Athetics, and how he exploited every opportunity to help his team win. For example, at the time of the book's writing, on-base percentage was largely ignored and Beane was able to create division-winning teams on the cheap. After the enormous success of the book, statistical analysis began to have a much bigger and more noticable effect on baseball. Nate Silver and friends ran Baseball Prospectus while The Hardball Times and Fangraphs increased in popularity. (You might recognize Silver from FiveThirtyEight.com, his political blog that predicts races using his statistical model. (He picked every single state correctly in 2008)) However, as Bill James was hired by the Red Sox, Keith Woolner was hired by the Indians, Paul Depodesta was hired by the Padres, Neil Huntington was hired by the Pirates and Josh Byrnes by the Diamondbacks (among others) the market adapted to the value of OBP. You'll notice some low OBP guys on the current A's team (Mark Ellis, Rajai Davis) and that is because Beane has shifted to a more defensive focused approach. As the market has caught up to OBP, Beane has identified defensive as a market inefficency.
Reader's digest version:
Statistics can be used to take advantage of an organization's resources to increase its chances of winning. In baseball, this can mean not bunting in the 3rd inning. In football, it can mean going for it on 4th down more often.
A further comment on why we don't see coaches going for it on 4th down as often as statistics suggest they should;
In a 4th and 3 situation on the 50 yard line, you have three possible outcomes.
First, going for it and making it. We can agree that this is a positive result.
Second, going for it and not making it. The resulting meltdown and second-guessing from ESPN and your local columnists is severe.
Third, not going for it. The safe choice.
As we saw last year when Belichick went for it in Indianoplis and was stopped, virtually every pundit denounced it as a terrible decision even though Smart Football and Advanced NFL Stats provided evidence that it was the right call. Belichick had the right combination of job security and not giving a fuck what you think that enabled him to make that call. Most coaches don't, which is why you will see the safe call more often than the right call.
September 23rd, 2010 at 2:30 AM ^
The thing that makes advanced stats in baseball so nice is that so much of the game is isolated to one on one situations. Unfortunately it's tough to isolate how much credit the QB should get for completing a pass in relation to the receiver and the defense.
That said, advanced stats are definitely making big splashes in just about every sport out there at this point. Basketball has seen a big increase in the use of advanced stats recently, with a lot of interesting work being done based on different 5-man units. Brian Scalabrine, 5 years ago, got a 15 million dollar 5 year contract because Daryl Morey (at the time an assistant with the Celtics, now GM of the Rockets) noticed that when he was on the floor with Jason Kidd/Vince Carter/Richard Jefferson all three of those guys shot a significantly higher percentage.
There has been a good deal of statistical analysis with football too. Many here have seen the analysis suggesting that teams should go for it on 4th down in most situations. I also recently read an article about how most teams pass a less than optimal amount. There was recently article about how when Michael Vick runs the ball he significantly increases his team's chances of winning, while when he throws the ball, well...not so much.
September 22nd, 2010 at 9:41 PM ^
A lot of coaches do utilize statistics, and probabilities. Although I personally believe in having balance. If I was a guy who believed heavily in stats and probabilities, then I'd want my coordinators to be the Don Zimmers to my Joe Torre.
September 22nd, 2010 at 10:29 PM ^
I would guess many coaches are aware of, for example, the predicted advantage of going for it on 4th more often. But if they take an unconventional approach and lose, they are blamed (more than usual) for the loss. Even if they would be better off in the long run as predicted by the statistics, they might not be around to take advantage of the long run if too many "risks" don't go their way.
Also, the coaches that might be most willing to take perceived chances are the coaches with job security, but coaches with job security have good teams, and good teams are justifiably risk averse (see: Tresselball).
September 22nd, 2010 at 10:51 PM ^
Bill Belichick going for it against the Colts at around the Pats 20 yd line last year is a good example of your 1st statement. He was torn apart in the national media for it.
ESPN Mag had an article entitiled "The Analytically Correct Guide to Football" in their NFL preview this year. It touched on a few different situations (did not have much depth, but still interesting). I can't find the link, Insider or not, on the site.
Edit: wow, google. Unfortunately Insider.
http://insider.espn.go.com/nfl/insider/news/story?id=5496577