John Harbaugh Decided to Not Look Desperate, Rather than Make it a One Score Game Late

Here was the situation. Baltimore got to a first and goal situation at the Tennessee 9, trailing 23-10 and with all three timeouts remaining. Then the Ravens had a delay of game penalty on 3rd and goal followed by an incompletion, setting up 4th and goal from the 11. Harbaugh chose to kick the field goal to make it a 23-13 game with 7 minutes left, and kicked off to Tennessee.

As it turned out, the decision didn’t matter in this particular game, because Tennessee maintained possession for most of the remainder of the game and kicked a field goal to make the final score 26-13 with less than a minute remaining. However, I still like to look at these situations because you never know when the decision will impact the outcome.

Here were Harbaugh’s thoughts:

"It was fourth-and-what from the what? The percentages aren’t really good. So I figured, let’s get the three [points] right here. We’ve got about seven minutes left, we have all our timeouts. [We can] kick it down there, try to get a stop, and we’ve got enough time. If it had been fourth-and-six at the goal, then you’re going for it. My thought was to go for it, but it got to be fourth and 11. That was a little too desperate. You can probably make a case the other way, but I don’t know how better a case it would be."

Cue Michael Lombardi of nfl.com, who comes to Harbaugh’s defense in arguing that “Harbaugh’s field goal call was not the head scratcher it seemed.”

"Conventional thinking would have you believe that if the Ravens convert the 4th down and eventually score they are just one possession away from tying or possibly winning the game. However if the Ravens failed to convert, essentially the game is over, in part because the Ravens will need two touchdowns to win, which would require recovering an onside kick — really long odds."

I think both demonstrate issues with probabilistic thinking and properly weighting the different outcomes. We’ll start with Lombardi. I’m not sure why he thinks that the “game is over” if they fail and turn the ball over on downs at the Tennessee 11 down by 13, and that scenario would require an onside kick, but needing two scores by way of a touchdown and field goal to tie  would not.

In actuality, the difference between being down 13 (after a failure) or down 10 is minimal, while the difference between down 10 and down 6 is fairly large. If we assume a touchdown on the first drive when down by two scores, then being down by 6 isn’t really that much worse than down 3. Coaches tend to fall into the lure of getting in field goal range, and as Brian Burke of Advanced NFL Stats pointed out using game data, coaches needing only a field goal to win actually had a lower win expectancy than those needing a touchdown, when needing to travel a roughly similar distance to get into field goal range versus score a touchdown.

It’s an interesting contrast to last year, in fact, when Harbaugh’s fourth down decision at the goal line against Pittsburgh laid the stage for the dramatic win at Heinz Field. Here, Harbaugh’s decision boils down to this: hold the Titans twice and get two possessions resulting in a TD and FG with 7 minutes left, then win in overtime; or score, and need a TD on one more possession to win outright; or fail to score, and hold the Titans twice and get two touchdowns to win outright in regulation.

Harbaugh states that the difference of moving back to the 11 influenced his decision, and he would have gone for it from the 6. He correctly surmises that the chances of a TD went down, but I don’t think he accurately assessed the reward versus risk. Advanced NFL Stats game chart shows Baltimore with an 11% chance of winning when they got to 1st and goal at the 9. Four plays later, when Cundiff’s kick sailed through the uprights and they were still down by 10, those chances dropped to 4%. That’s quite a drop based on the promise of scoring 7 to the certainty of scoring 3.

Harbaugh wants to know how much better a case? Well, in either case, the chances were slim, but my estimate is that he cost his team about a 2% chance of winning, which sounds small but he cut his team’s chances almost in half. But he should have gone for it if he had at least a 10% chance of getting that touchdown on 4th and 11, because of the massive difference in reward between 6 vs. 10 vs. 13. Lombardi mentions that the conversion rates on 4th and long is about 33% (not adjusted to position on field). Since 2002, teams that go for it between 4th and 9 and 4th and 13 near the opponent’s goal converted 17 of 52 times (32.6%). There were very few exact same distance plays on 4th down, but teams on 3rd or 4th from the 10 scored a touchdown 18 of 96 times (18.6%). That last number probably understates the Ravens’ chances, since some of those teams weren’t desperate to score a touchdown on the play and settled for a shorter throw. I’m willing to guesstimate that the Ravens had about a 25% chance of scoring a touchdown there.

They should have gone. I don’t care how desperate it looks. You are down 13 with 7 minutes left. You should be desperate. And if you want to think of it in poker terms, the Ravens had more “outs” if they went for it, and either scored to make it a one score game to win, than by needing two possessions and a TD/FG combo to merely tie. They were in almost as much trouble down 10 as they would have been down 13, even if Harbaugh didn’t feel quite as desperate.

[photo via Getty]