## Checkdowns: Lane Kiffin’s 4th down decision

McKayla Maroney is not impressed.

Lane Kiffin is not a very good coach, and that’s putting it mildly. He’s one of the most hated men in college football and he’s the face of a USC team that has had the worst season of any preseason favorite since at least 1964. With a 7-5 record, no one is defending Lane Kiffin. And given the various ways he mismanaged the clock against Notre Dame, nobody can defend his performance in that game.

But let’s puts everything aside — Kiffin for being Kiffin, the fact that USC had 1st and goal from the Notre Dame 2 with over 5 minutes remaining and then proceeded to waste over 150 seconds of clock — and look at one particular decision. Facing 4th and goal from the Notre Dame 1-yard line, trailing by 9 points with 2:33 left, Kiffin decided to go for it.

This clearly defies conventional wisdom, and when the move failed, it opened him up to even more criticism. But was it the right call? According to Brian Burke, if this had been an NFL game, the correct call would have been to kick the field goal.

That may not surprise traditionalists, but readers of this blog and Advanced NFL Stats may be surprised to find that, according to the 4th Down Calculator, when trailing by 9 with 2:33 remaining, you need an 87% chance of converting to make going for it the correct call. (I will note that if you are trailing by 10, things change dramatically and going for it is the correct play.)

But this was not an NFL game. Burke’s model is based on two assumptions that are relevant here: one, the team has an average number of timeouts remaining, and two, that the clock will stop with 2:00 to go. USC had one timeout left (which is probably below average for this situation) and there is no two-minute warning in college football. So it’s likely that using 2:33 is not the correct number to use the 4th down calculator for college.

If you use 2:33 remaining, you need an 87% chance of converting to make going for it the correct call.

But, according to the same model, if you use 2:03, it drops to 64%.

If you use 1:33, it drops to 13%.

Obviously figuring out which input to use is very important. However, let’s think about it in a different context.

If USC scores a touchdown and does not onside kick (which I don’t think they do), ND gets the ball at roughly the 25-yard line with 2:25 left. On 1st and 10, they run, USC calls timeout, and there is 2:20 left. On second down, Notre Dame runs, 40 seconds tick off, and there is 1:35 left. Rinse, repeat, and Notre Dame punts with 50 seconds left. This means USC gets the ball with roughly 40 seconds left at say, their own 45.

Here college football’s rules benefit the Trojans because the clock stops momentarily on first downs. At this point, it comes down to this:

1) What are USC’s odds of scoring a touchdown with no timeouts and 55 yards away with 40 seconds left? Call this T.

2) What are USC’s odds of scoring a field goal with no timeouts and 55 yards from the other team’s end zone with 40 seconds remaining? Call this F.

Let’s assume, conservatively, that USC has a 50% chance of scoring a touchdown from the 1 (in the NFL the conversion rate would be 68%, but Notre Dame’s run defense is excellent. In reality, USC called a good play but poorly executed the pass.) and that USC has a 100% chance of success if they kick the field goal from the 1.

In that case, the only remaining question is whether F is twice as large as T or not. My guess is yes — it’s significantly easier to drive for a field goal than to score a touchdown. And if you think USC has a greater than 50% chance of scoring a touchdown from the 1, then F can be even lower.

No hard math here, just my gut. And let’s not forget that if you fail from the 1, you have at least a chance of forcing a safety and getting the ball back with a chance to tie. The odds aren’t great, but that possibility does enter the calculus. Lane Kiffin is not a good coach and he’s made a zillion mistakes. This might have been one of them — I don’t have a definitive answer either way — but my opinion is he made the correct call. Scoring a touchdown from 55 yards away with only 40 seconds left is really, really hard. The difference between that and scoring a field goal in that situation has to be larger than the difference between the two scoring options from the one-yard-line.

• MikeN

You have to account for the fact that field goals in college aren’t as automatic, so you have to get pretty close to the goal line to kick one.

• Mark

Did you notice the contradiction in your own observations? You say no two minute warning and only one timeout makes it like there is less time left ….. but the clock stopping for first downs makes it like there is more time left, which again makes the time remaining calculation close.

The other thing you need to adjust for in college, as MikeN said, is much worse kickers. The chip shot is pretty safe. But where a 43 yarder might be an 80% proposition or better, not so in college (I think I read USC kickers were 2/6 from beyond 40).

I also think assuming no onside kick in both scenarios is a mistake. I don’t have the math on it and it can be argued that you don’t take that 25% shot if you get the TD, but taking the field goal you likely DO onside kick.

And the college clock stopping plays into this, too. In the pros, with the clock rolling, you just can’t drive a long field with under a minute and no timeouts. The D can cover the sidelines, and anything deep down the middle takes up too much time to run down and clock it. But in college the defense has to defend the whole field, and you can eat up big chunks of yards without using much clock. So taking the gamble on the onside kick makes more sense than in the NFL, for the field position, and to minimize the risk of losing to an opposing first down.

• Chase Stuart

It’s not a contradiction; the NFL rules help the offense but hurt the defense. If this was the NFL, after a score, and if you have two timeouts left, you would get the ball back with about 1:50 to go. Now once you’re on offense, the college rules are more favorable, but you lose a full minute. That’s very significant in this situation — 1:50 from the 45 isn’t so hard.

No doubt kickers are much worse in college – that does make kicking the chip shot a more desirable option. But that’s included in the calculus.

You also need to keep in mind that onside kicks are extremely difficult to recover in college football. There’s no way it’s 25%, especially with the new rules which basically prevent you from doing the one-hop onside kick. Considering the fact that the kickers aren’t very good and the rules are not favorable, the odds of recovering the kick would be very low. (It’s also worth remembering that if you choose to onside kick only in the FG option, well that has to go into the calculus of whether it’s smart to kick or go for it.)

• Mark

A full minute between college and pros rules is about two mid length passes over the middle in difference. As for the timeouts, I don’t think you should assume 2 is the average for his calculation without asking. In any way, my point is that you can’t discount the amount of time left based on timeouts and 2 minute warning, without also factoring in that that smaller remaining time would go farther in college.

Yes, working an onside kick into the odds would have to be figured in. And all of this exercise is predicated on the rather suspect notion that you can determine anywhere near reliable odds for any of these things occurring and calculate them in a meaningful way. I was also unaware of a difference in college onside kick rules.

Anyway, the main takeaway is that by far the biggest mistake was the prior playcalling and clock management, and I agree with that. But I still have yet to be convinced that going for it on 4th makes any sense. Tell Notre Dame that they get to have one play on defense from the 2 to absolutely lock in their win (and still a good shot of winning if they fail), or they can face fielding a kickoff, trying for a first down on the ground only, risking turnovers, covering a punt, and then defending the field when one big play can cost them the game, and they’ll take option one every time.

• Nate

> …And all of this exercise is predicated on the rather suspect notion that you can determine anywhere near reliable odds for any of these
> things occurring and calculate them in a meaningful way. … But I still have yet to be convinced that going for it on 4th makes any sense…
> trying for a first down on the ground only.

If I read things correctly, at the point of decision, USC had had eight posessions. From those, it got one touchdown, into sure field goal range twice, and scored two field goals. So let’s say, for the sake of discussion, that USC is four times as likely to be able to score a field goal or touchdown in a drive as they are to be able to score a touchdown. Then going for it makes sense if they have a one in four (or better) chance to get the touchdown.

Those numbers are not crazy. Naturally the chances to convert the fourth down, and to score on the ensuing drive are speculation, and the difference is marginal.