Saturday, December 21, 2013

Put Me In Coach

I beat up on coaches a lot. More in conversation than in this blog in fact. I've done my share of armchair, from the bleachers, and Monday morning quarterbacking. Allow me to defend coaching a little and relate some economic concepts to the coaching profession.  I want to focus on college football coaches and to use Oklahoma's Bob Stoops in 2013 as a specific example, but this applies in large part to coaches at all levels and in all sports.

Flat out, coaches have a tough job. Yes, many are very, VERY well paid to do this job. Of course, many more are not. The job is tough because it is high-profile performance judged by a vast sea of people who have much less information and skills and who tend to approach the issue from an emotional standpoint. (Not me, of course; when I am yelling at my TV, it is because of my passion for reason and logical decision making.)

A coach has to balance between running an on-going training program while producing output that meets consumers' high demands. The training program is comprised of the gamut from relative beginners to high-value-producing experts (I'd call them professionals, but this isn't that blog post)--all of them thrown into the same "classroom".

Let me use the 2013 OU football team as an example of how coaches face issues involving asymmetric information, decision-making under uncertainty, skewed risk-reward payoffs, and management of public and intra-firm relations.

Throughout Bob Stoops' very successful 14 seasons as Oklahoma's head football coach he has either had a high-profile, all-star quarterback or an inexperienced newcomer who struggled not just when compared to his high-profile predecessor but also in absolute terms. 2013 was of the latter variety.

As Stoops sought to replace the 4-year record holder Landry Jones, he was evaluating the options with many backseat onlookers. The obvious choice to many was Blake Bell, the two-year backup. But in late August Stoops awarded the 2013 starting job to freshman Trevor Knight. When Knight stumbled some in early games, the natives including me grew restless for Bell to be given a shot. A combination of a bad first half and a slight injury gave the natives what they wanted in the West Virginia game, and Bell performed well. But then a few games into his starting role, Bell too fell into a malaise. The offense stumbled contributing greatly to OU's losses to Texas and Baylor. A little in and out substitution between Knight and Bell over a couple of games ended with Knight regaining the starting job for the Kansas State game (a victory) only to exit the role at half-time against OSU due to injury. Bell came in and played well if not better than Knight. Oh, and the formerly third-string sophomore Kendall Thompson was inserted before Bell replaced him in the OSU game.

To say this wasn't according to script is an understatement. But the script isn't actually written by fan dreams. It is an emergent process governed by both luck and coaching decisions. The coaching decisions are governed by a couple of underappreciated forces--uncertainty and asymmetric information. Coaches know a lot, and I mean A LOT, more than the rest of us. They see these players in practice and in games and in replayed videos of both. They interact with them. They also have a game plan and a complex strategy of plays to accomplish that plan. We don't know the plays, the formations, the game plan theories, or how well or poorly the players fit into them all. Add to that the complexity that combinations of players will imply different outcomes. Oh, and players are living lives all this time meaning they simply aren't the same in Spring of sophomore year as they are in December of senior year. Oh, and coaches are humans with biases and informational blind spots. They are operating in a cloud of uncertainty. We are in a fog orders of magnitude more dense than coaches are due to the asymmetric information.

And yet we judge them and will call for their heads if too many of their decisions turn out "wrong". Was Knight the right choice for Stoops to make in August? In September it seemed like the answer was no. In October and November it seemed more and more like the answer was probably yes. In one half of one game in December (n = .5 for statisticians out there) the best we could say was, "Looks like it was a toss up either way". It took us as onlookers an entire season to finally say what we should have been saying all along. To wit, "The coaches probably are making the best choice available, and that choice is still a guess".

"Coaches are paid the big bucks to make those calls and get them right!" you say. Well, yes and no. What is "get them right"? Right as judged by critics--media, fans, detractors, players, administrators, donors, parents, etc. Coaches have many masters. Effectively managing the intra-firm (i.e., players, assistant coaches, administrators, donors, some fans) relations along with the public (i.e., media, some fans, detractors, other team's coaches) relations implies they have interests that may conflict with simply maximizing the probability of long-term winning. Their risk-reward payoff matrix is skewed to a degree that is hard to appreciate. Balancing this well is an art.

Reflecting on the Sooners' 2013 season has humbled me and caused me to appreciate the coach's job(s). I don't think it is just because I view the season as a success with hindsight knowledge (it would be judged a failure from an ex ante point of view). Trying to put aside how a last-minute comeback victory over Oklahoma State makes me feel, I think I would feel that Stoops did a great job in 2013 win or lose that game.