Belichick’s Coaching Blunder Bigger than Carroll’s

Bill Belichick is the luckiest man in America this week.  While Pete Carroll is being pilloried in the press and at the water cooler following a questionable play call at the end of the game, it should have been Belichick whose coaching reputation took a throat punch on Sunday night.

But wait, you say: Everyone, and I mean everyone, contends that the worst coaching decision in Super Bowl (Football??) history was indisputably Pete Carroll’s decision to throw the ball on second and goal from the one-yard line with twenty-five seconds to go in a game his team was trailing 28-24, especially when he had one timeout and the beastly Marshawn Lynch in the backfield.  At first glance, this view seems rational.  This was likely the wrong call.  In fact, I will agree that it was probably a bad call, but not as bad as Belichick’s call forty seconds earlier.  Not nearly so.

Do I have your attention?

Rational football fans (and economists who aren’t football fans) can agree that whether or not a play call is a good one or a bad one must be judged from the standpoint of the decision maker at the time the decision was made – before the outcome of the call is known.  It is simply disingenuous to evaluate a call after considering what actually happened.  The only honest and fair analysis involves evaluating percentages.

Furthermore, the most accurate measure of just how bad a coaching decision was can best be determined by measuring the difference between the probability of success of the course taken and the probability of success of the course that would have maximized the probability of success.

So, for example, determining whether Carroll’s decision to run a passing play (rather than kicking a field goal) from the 10-yard line with 0:06 left on the clock at the end of the first half while trailing 14-7 was or was not a good call cannot take into account what actually happened.  On that play, a variety of outcomes could have occurred – incomplete pass in less than six seconds (leaving time for a field goal), incomplete pass that ran out the clock, a completed pass short of the goal line than ran out the clock, a sack that ran out the clock, an interception, etc.  I am fairly confident that Nate Silver would tell us that the rational (efficient) move in that situation would have been to kick a field goal.  Take the points and the momentum into halftime. But Carroll rolled the dice . . . and scored a touchdown.  It was a “gutsy” call, many would claim, and shows his coaching prowess.  But it was only a gutsy call because the 15-25% probability event occurred.  But that is simply not a rational way of evaluating the play call.

Likewise, had Russell Wilson completed the pass for a touchdown at the end of the game, Carroll’s call would have been labeled “inspired” because everyone on the field, in the stadium, and in the television audience knew the rational decision would be to give the ball to the Beast.  But the call that was made, even had it resulted in a touchdown, would have been the same call.  Again, we need to judge the wisdom of the call by evaluating the relative percentages of the possible outcomes.

Let’s break it down.  Here is what we knew the moment Marshawn Lynch’s knee hit the ground with 1:02 left in the game after being tackled at the one-yard line after a four-yard gain on first down:

Seattle had one time out left and the clock was running;

Marshawn Lynch is in the backfield, and he had gained four yards in the prior carry;

Russell Wilson is a smart quarterback (who won last year’s Super Bowl) and isn’t prone to boneheaded mistakes; and

New England is expecting a run.

But here are two other things Carroll knew, at least generally understood:

Marshawn Lynch had carried the ball five times from the one-yard line this season and had scored only once.

Over 100 passes had been thrown in the NFL from the one-yard line this season, and none had been intercepted.

Simplifying things just a bit (by removing a run/pass option that could have been called for Wilson), Seattle had two arguably rational courses of conduct at this point.  First, it could hand the ball to Lynch on second down (by this point, there were 25 seconds to go – no shame in huddling up here, especially since the clock was running).  If he scored, great.  If he didn’t, Seattle could either call a time out or spike the ball.  If they spiked the ball, they would get only one more play because the next play would be fourth down.  If they called a timeout, they would have 20 seconds left and it would be third down, but they would have no timeouts left.  At that point, they could either pass twice, pass once then run once, or run once.  It is unlikely they would have had time to run twice more if Lynch was stopped the first time on third down.  So if they ran on second down from the one, they would very likely get either two running plays, or one running play and two passing plays.  It is highly improbable that they would get two running plays and a passing play or three running plays.  (This could only happen if one of the running plays ended with the runner getting out of bounds).

The second rational course of conduct would be to choose to throw on second down.  If this happened, they could: score, the pass could be incomplete (thus stopping the clock), Wilson could be sacked, or it could be intercepted.  Let’s presume Wilson is smart enough that he wouldn’t have allowed a sack.  If the pass was complete, it would very likely be a touchdown and Carroll would be applauded for his courage and cunning.  If it was incomplete, then Seattle still had two runs from the one with Lynch, calling a timeout after the first one if he didn’t score.  So under this scenario – throwing on first down – the odds were very, very high that Seattle would get two running plays and one passing play from the one yard line.  The only risk was a miniscule chance of an interception.  Incidentally, the odds of Lynch fumbling were likely as high — or higher — than of Wilson throwing an interception.

Give any NFL, College, High School, or Pee-Wee Coach the following multiple choice question, and you will get the same answer:

If you are coaching the Seattle Seahawks in the Super Bowl and have three plays to score from the one-yard line, which combination of plays would you choose?

1. One run by Lynch and two passing plays;
2. Two runs by Lynch;
3. Two runs by Lynch and one passing play.

This is not a hard question.  “C” is the correct answer.

Once again, I am not arguing the Coach Carroll made the right call.  He probably should have run the ball with Lynch.

But his decision is not nearly as irrational as it seems, given the benefit of hindsight.

So, pray tell, what call by Belichick was worse?

As noted earlier, when Lynch was tackled at the one on first down, there was exactly 1:02 left in the game.  New England had two timeouts and Belichick had a decision to make.  Once again, it is simply not acceptable to take an end view of what actually happened (interception on the goal line) and then judge Belichick’s decision not to call a timeout as wise.  If Belichick knew Carroll was going to call a pass play that New England would intercept, Belichick is a much better cheater than we even knew.  Instead, we must judge his decision from the standpoint of what we knew at the time.  We must examine the probabilities as they then existed.

The probability was that Seattle would score a touchdown.  And it would probably happen on second down.  (Lynch had just run for four yards on the prior play).  Belichick is a smart guy, so he knew that.  He knew that if Seattle scored, his team would trail 31-28.  He also knew that he had arguably the best quarterback to play the game standing behind him, and he had one of the best field-goal kickers in the NFL standing next to him.

Here is another easy multiple choice question.  If you are trailing by three points in the Super Bowl, would you rather have:

1. Two timeouts and 20 seconds to go 45 yards to kick a field goal to tie the game;
2. One timeout and 55 seconds to go 45 yards to kick a field goal to tie the game.

This is not a hard question.  Ten out of ten times, you call a timeout.

Play this out.  Say Belichick called a time out with 1:00 to go.  If Seattle scored on the next play, he would have no less than 55 seconds left and one timeout.  If Seattle didn’t score, then he would burn his final timeout, and then Seattle would have likely scored on third down.  If so, then Belichick would have no less than 50 seconds left and no timeouts.  If Seattle didn’t score on third down, then it would likely have taken the clock down to three seconds and then called a timeout.  After which the one remaining play would determine the season.  But the odds of that happening were vanishing small.  The best bet would have been that Seattle would score on second or third down.  In such an event, Belichick would have had at least 50 seconds, needing only a field goal to tie.  And remember, Brady and Gostkowski are stellar in the clutch.

Instead, Belichick chooses to let the clock run, and the ball is snapped with 25 seconds remaining.  This means that Seattle can run two or three plays (see discussion above) to take a three-point lead.  It also means that New England will have, at best, 20-22 seconds remaining to kick a field goal, assuming Seattle scored on second down.  If Wilson’s pass had been incomplete, then the best-case scenario for New England would have been 16-18 seconds to score if Seattle scored on third down.

In conclusion, I am confident that Nate Silver (or other number crunchers) could mathematically demonstrate that Belichick would have maximized New England’s chances of winning by calling a timeout with one minute to go.  I would also bet my left you-choose-the-body-part that Belichick’s decision deviated further from the optimal coaching decision than did Carroll’s decision to throw on second down.

Prove me wrong.