Home / General / Why are so many NFL coaches so terrible at probabilistic decision making?

Why are so many NFL coaches so terrible at probabilistic decision making?


I’m something of an aficionado of terrible punts, as in punting under circumstances where, statistically, punting is an awful idea. Sunday’s Carolina-Chicago game featured no less than three such calls in the fourth quarter alone by Carolina coach Ron Rivera. Let’s go to the tape:

(1) Score: 17-3 Chicago. Fourth and 13 at the CHICAGO 38, 13:28 remaining in the game. This was Carolina’s eighth possession of the game, with those possessions resulting to this point in slightly less than one touchdown. The Panthers needed at least two touchdowns in what at this point was likely to be three more possessions at most. As extra added bonus, the Bears’ offense had been almost completely short-fingered impotent all afternoon, and would remain so. Chicago would finish the game with 153 yards in total offense — a remarkable total for the winning team in an NFL game played in something other than a hurricane or a blizzard. So turning the ball over inside the Chicago 40 would not exactly qualify as a disaster (That punts themselves are not classified as turnovers, although they obviously are, helps explain why coaches are so eager to employ them).

(2) Same score, fourth and two at the Carolina 32, 9:52 remaining. Here’s a thought experiment: if you’re Chicago, what do you want Carolina to do in this situation? Exactly.

(3) Same score, fourth and 12 at the Carolina nine, 3:44 remaining. Did Rivera have the under? I didn’t watch the game on the teevee, but if anybody in the LGMverse did, did the announcers give him any grief about this? I mean at long last have you left no sense of decency Coach Rivera?

And the thing is — these aren’t unusual decisions for an NFL coach. It’s not like Rivera is some sort of outlier here. The outlier is somebody like Bill Belichick, who doesn’t act like there’s some sort of rule that you are “supposed” to punt on fourth down, except in situations that are so outrageously obvious that even the average NFL head coach (average salary: $4.5 million per year) can’t get it wrong.

Possible explanations:

(A) Math is tough! I.E., these guys just can’t do even the simplest probability analysis, so they stick with extremely crude heuristics, in the form of the traditional conventional wisdom (you punt on fourth down in football. Or maybe third down if you’re in Canada).

(B) The psychological tendency to choose to delay the complete inevitability of defeat for as long as possible, even when this involves making decisions that minimize the probability of winning.

(C) People hate losing more than they like winning, that is, they are risk averse when seeking what are framed as gains and risk seeking when avoiding what are framed as losses. Going for the first down and not making it is an obvious “loss,” while punting isn’t, given the traditional framing of coaching decisions in these circumstances (see A).

(D) Coaches are punished both for losing and for defying convention. They are particularly vulnerable to criticism/firing when they lose while defying convention. If you’re going to lose, do so while not offending the delicate sensibilities of “old school” sportscasters etc.

(E) A related math failure: poor decisions in the context of what would remain a likely loss in any event tend to get something of a pass. Example: Suppose not being an idiot by punting three times while trailing by 17-3 in the fourth quarter would improve your odds of winning from 3% to 6%. A good poker player sees this as a 100% improvement in his odds, which is why he will go all in on a shaky hand now if the alternative is the high likelihood of getting blinded out a few hands down the road. But many people are prone to rationalize these situations (“what difference does it make? The real problem was being down 17-3 in the fourth to a team with no offense” etc.).

(F) The great improvement among NFL punters as a class. This is like handing NFL coaches some high grade hydroponic weed instead of 1970s-style schwag, and then expecting them not to keep getting high on punting. What these guys don’t realize is that the other team’s punter is also a lot better now than he would have been Back in the Day.

For instance, Rivera no doubt thought he would “pin Chicago deep” with an expert pooch punt when he kicked from the Bears’ 38, and the Panthers’ punter more or less did, putting them at their own ten. The Monsters of the Midway proceeded to gain a net of zero yards on the next three plays, leaving Carolina with great field position when — oops, Chicago’s punter rockets a 66-yarder with no return, leaving Carolina 76 yards away from pay dirt.

(G) Bogus ad hoc psychological explanations of the type “turning the ball over on downs is much more demoralizing than punting.”


  • Facebook
  • Twitter
  • Google+
  • Linkedin
  • Pinterest
It is main inner container footer text