To Turn the Bills Around Bring Moneyball to Buffalo Before It Catches Fire

As of this writing--it appears that baring a complete and unprecedented football miracle the Buffalo Bills will once again miss the playoffs, despite lofty preseason hopes. It also appears to be the end of another leadership regime in Buffalo. Chan Gailey is most certainly on the hot seat, and it's difficult to envision him being given another year (nor does he deserve one--as this article will hopefully expose). Buddy Nix appears better suited to survive the coming changes in the off-season but should he?

What direction should the Bills look to as they decide where to go from here? Doesn't it feel like we've tried everything?

The answer is no. We have not tried everything. In fact no NFL team has gone where the Bills should go, but it is only a matter of time before they do.

I submit to you that the best way for the Bills to rebuild is to immediately embrace the use of advanced statistics as it comes to coaching decisions and as it comes to player evaluations. In essence, go the moneyball route.


Get a coach who goes for it on Fourth Down Aggressively!

In 1995, a UCal study persuasively demonstrated that NFL teams do not go for it on fourth down enough. Teams kick field goals in the middle of impressive, potentially game altering drives, rather than take the favorable odds of picking up a fourth and short. Teams punt at mid-field, achieving very little marginal value, rather than go for it on fourth and short. You can read the study for yourself. The mathematical evidence is fairly persuasive.

Virtually all deniers of Romer's study argue that the math doesn't take into account the momentum shift of going for it on fourth down and coming up short. This argument is preposterous and steeped in a loser's mentality. First, it totally ignores the sacrificed "momentum" of kicking a field goal or punting. Second, as demonstrated in Oregon, a culture of going for it on fourth consistently gives an offense an additional swagger. They finish out drives more, and nothing builds confidence more than touchdowns.

As an example take the Packer's game last night. It was 7-7 and the Packers had al the momentum coming off of a huge deep TD pass. What's more is they had the ball and were marching deep into Giant's territory. However, the drive stalled out on the 36 yard line as they came up about 2 yards short on a third down play. The math clearly says that in this situation you go for it. Instead, the Packers kicked a long field goal, missed and handed the Giants all the momentum in the world. Then at the end of the half, the Packers repeated this mistake, declining to go for a fourth and short on the ten, in favor of a momentum killing field goal. This was the last time they would even get a sniff of the Giants.

Another potential retort someone could come up to the mathematical evidence is that, if the evidence were so strong, why hasn't someone utilized it yet. To this I say two things. First, Chip Kelly. Second, there is a certain level of groupthink that is promoted by the fact that NFL coaches are so closely scrutinized. A coach can better withstand middling results while conforming to league coaching norms than he can while being a non-conformist. And so coaches are extremely reluctant to embrace going for it on 4th down because doing so would put heightened scrutiny on their results.

But what if as an organization, the Bills made it clear that they were embracing advanced statistics? What if they got buy in from the fan base?

This year we've gone for it on 4th down 4 times and we already lost seven games. Ask yourself if that makes a lick of sense.

Advanced Statistics in Player Evaluation

Football scouts, and I include Buddy Nix in this group, tend to be old school. They tend to value their subjective assessments of who can do what, and how well they can coach up guys to put together a winning team. So let's take a look at the 3 biggest decisions that Chan and Buddy made this year:

1) To keep Fitz as their starter.

2) To limit CJ Spiller's touches to around 15 per game.

3) To sign Mario Williams

In all three categories, we see a unique pitfall of Player evaluation and how an advanced stat regime might behave differently.

The Fitz Debacle

Frankly, from the start this was doomed. Ryan Fitzpatrick has a career 76.7 Quarterback rating. This is marginal at best. I cannot find a statistic for the average QB rating of playoff teams in the last decade, but I'd wager it is higher than 76.7. (That's certainly a stat I wish the front office would have considered in deciding to keep fitz). On a team that had one of the league's worst defenses last year, it seems relatively clear that you are a long shot to make the playoffs with that level of QB play.

But the Bills, poo poo'ing Fitz entire body of work, with a large sample size, insisted that Fitz could become a playoff QB. And to their credit they have coached him up to one of his best seasons ever. He is posting an 85 QBR. However a QBR of 85 is still, I would imagine, well on the low end of what typical playoff teams get out of their QB. And so you are thinning your margin of error elsewhere--you're running game, defense and special teams all better be elite. Which brings us to our next point.

Give Spiller the BALL

Everyone agrees here. There is no need to beat a dead horse other than to say that a statistics driven regime would not have made the same mistakes CHIX has made with Spiller. They would have rejected their arbitrary decision that anything other than 15 carries per game risked injury as being both baseless in medical and marginal value analysis.

And they never would have allowed a runner gaining 6.6 YPC to play second fiddle to someone averaging nearly half that for so long.

Mario Williams and the Small Market

I'm not aware of any statistic on QB hurries per passing play rep. I would like to see such a stat. I'd venture that over the course of his career, Mario does not rank in the top 10 in the league with at least a certain number of reps (but of course I'm just going on my observations of his game--which seem to be typified by a few brilliant rushes per game and many quiet reps in between).

The point here is that while I'm not sure what Mario Williams value is, a stats driven team would have looked at his QB hurries per passing play rep, and not focused on his sacks or his pro-bowl appearances (which are basically a function of sacks) or how dominating he can look at times. They would have looked at how often he can disrupt a passing play, on average, and how valuable that was to increasing their win percentage relative to other moves they could have made.

Especially before dolling out 100 million to him.


1. Fire CHIX.
2. Hire a regime committed to advanced stats.
3. Assuming my assumptions about Mario's marginal lineup value are correct, trade him for something with more marginal lineup value.
4. End the Fitzpatrick experiment.
5. Assuming Spiller continues to be an effective yard per touch person--build your offense around getting him the most amount of touches you can.

It's really only a matter of time before this kind of analysis becomes popular like it already has in the NHL, NBA and MLB. The Bills happen to have a golden opportunity this off season to be the first to try. I doubt Chip Kelly is available--but if he is--I'd lock him up in a second.

Just another great fan opinion shared on the pages of

Log In Sign Up

Log In Sign Up

Forgot password?

We'll email you a reset link.

If you signed up using a 3rd party account like Facebook or Twitter, please login with it instead.

Forgot password?

Try another email?

Almost done,

By becoming a registered user, you are also agreeing to our Terms and confirming that you have read our Privacy Policy.

Join Buffalo Rumblings

You must be a member of Buffalo Rumblings to participate.

We have our own Community Guidelines at Buffalo Rumblings. You should read them.

Join Buffalo Rumblings

You must be a member of Buffalo Rumblings to participate.

We have our own Community Guidelines at Buffalo Rumblings. You should read them.




Choose an available username to complete sign up.

In order to provide our users with a better overall experience, we ask for more information from Facebook when using it to login so that we can learn more about our audience and provide you with the best possible experience. We do not store specific user data and the sharing of it is not required to login with Facebook.