A discussion broke out between Chase, JKL, and I on this. When it started, Chase and JKL agreed on the answer, and I had a dissenting opinion. When it was over, we all agreed.
Who changed who's mind, and how?
Read to find out, and add your own thoughts:
Chase: With about 4 minutes left, they kicked the FG there at the 11. I think it was a bad call, potentially a very bad call. Dierdorf strongly disagreed, as expected. Thoughts?
Doug: I am now completely programmed so that going for it is the default in my mind on 4th and 1.
But I would not have gone for it there. In 56 minutes of football, the Ravens had 160 total yards, half of which was on two fluke passes. There was every reason to expect that the Titans would get a stop with plenty of time left. And even then if they fail to win in regulation, they had about a 75% chance of winning it in overtime. Field goal was the right call.
Chase: Field goal was the safe call. The Ravens weren't going to do anything if Tennessee didn't convert, and then the Titans would have kicked a FG as time ran out. That's worst case scenario. Best case? They go in for the TD.
By kicking there, you eliminate your field position advantage, you allow the Ravens to be aggressive on offense, and you eliminate you getting the ball last.
Doug: I disagree on the last thing. Part of the reason, in my opinion, that a field goal was the right call, is that the Titans had a very good chance of having the ball last. There was more than 4:00 on the clock.
After they kicked that FG, there was about an 80% chance of the Titans getting it back with the score tied and more than two minutes and/or three timeouts.
Chase: Well I suppose that's just what happened. But I think they were even more likely to get the ball back last (and needing it) if they went for it. By kicking that FG, you force the Ravens to become aggressive. I think a 3-and-out was a lot more likely following a failed conversion than a FG. The other problem is they'll have reasonable field position, where they'll be aggressive. The ball at the 30 in a tie game is much different than the ball at the 11 up by 3.
JKL: I ran some estimates. I was trying to be conservative in favor of the field goal attempt, and also estimate based on the specific team characteristics. I still came up with a 4% difference in win probability in favor of going for it, and since I was trying to overestimate chances of certain events, I suspect the difference is actually larger.
Doug, I think you overestimate the Titans odds of getting it back with all their timeouts, unless you mean they didn't use their timeouts despite being inside the 2 minute warning and after the Ravens picked up two first downs also. I think its being conservative to say the Ravens chances of scoring by any means on the next possession was 20%, let alone saying that it was 20% that the Ravens pick up 2 or fewer first downs before the 2 minute warning.
Once the Titans decided to kick, they were a sizeable underdog in the game. This is because a) there was still some chance of a missed or blocked FG, though slight; b) primarily, the number of remaining possessions for the Ravens is either N or N+1, while the remaining possessions for the Titans is N (and the ravens will not be as rushed on time), and c) the Ravens will likely get the kickoff in positive field position.
I'll just point out also that if part of the rationale is that the Ravens are less likely to pick up first downs and the offense has only had a few big plays, that rationale should actually also mitigate in favor of going for it given the time, field position and timeout scenario. A touchdown would have been a game changer (and a FG 2 minutes later would have been much better too), while the Ravens taking over at the 9 would have required at least 2 first downs by Baltimore to clinch it, otherwise, the Titans get the ball back either a) with about 2:40 and 2 timeouts and at their own 45 on average, needing a fg to tie, or b) at about the 2 minute warning with no timeouts and at about their own 35, needing a fg to tie.
Finally, with regard to Chase's point about the decision affecting the Ravens play-calling, I probably didn't estimate enough of a difference for these scenarios, which means, again, I underestimated in favor of the FG attempt and it still says go for it. I agree with this though. The Ravens would have been more willing to trade "first down expectancy" for time/timeout burning given the differing score and field position situations.
For fairness, the Ravens decision to kick on 4th and inches with time still remaining and the Titans having a timeout was *arguably not* the optimal choice either, but this is primarily because of the length of the figgie. No coach would have gone for it there, though, unless weather conditions dictated otherwise.
Doug: Let's assume the Ravens start at the 30 after the made FG. Prior to that point, the Ravens had 10 drives in the game (not counting the one at the end of the first half). Eight of those drives would have resulted in no points for Ravens if starting from the 30. Based on how things were playing out that day, I think this is reasonable:
5% TD (Ravens win)
15% FG with not much time left (what actually happened, ravens win)
40% no points, but no time left (overtime)
40% no points, and a reasonable amount of time left for the Titans
Let's say the Titans have a 30% chance of converting in the last scenario, and that they have a 60% chance of winning in overtime. Then I get:
20% Ravens win in regulation
12% Titans win in regulation
40.8% Titans win in overtime
27.2% Ravens in overtime
So that's about 53% Titans if they make the FG. The league was 261 out of 266 on 20--29 field goals, so let's say there was a 95% chance of making it. So I figure the Titans just above 50% if they kick the field goal, and I therefore dispute that the Titans were a sizeable underdog in the game after choosing to kick.
If you have time, I'd like to see the numbers that produced that. I think you guys are underestimating how awful the Ravens' offense was all day and how flukey the actual result was.
Now, that said, I have not made the case that kicking the FG was the right call. And I'm actually somewhat starting to doubt it.
If they don't pick up the first, then we've got something like (I'm getting a little sloppy due to lack of time):
20% Ravens run out clock.
5% Titans hold, score TD, win in regulation
30% Titans hold, get FG, go to OT
45% Titans hold, but don't get FG, lose.
That's 23% win for Titans if they don't pick up the first. If they score a TD, it's over. 100% for Titans. If they get the first, and then kick a field goal later, but I think 60% (same as the overtime probability) for Titans is a reasonable guess.
So it all depends on what the probabilities of getting the first, and the subsequent TD, are. If it's 65% chance of first, 30% chance of TD, then the Titans win probability is about 60%. If it's 65% chance of first, 20% chance of TD, then the probability is about 55%. If it's 60% chance of first, 20% chance of TD, then the probability is about 53%.
I don't know. I think it's real close, but I guess going for it probably was the right call. I'm officially changing my answer.
JKL: Those first numbers (65% first down, 30% TD) are actually what I used for calculating the TD likelihood for tennessee, except for the overtime split. There is also value in kicking a later field goal, and taking away some portion of the Baltimore playbook, even if a TD isn't successful.
I think 65% is conservative given the actual distance of this 4th and 1, but that's what I used to try to justify the FG. 65% is very close to the average 4th and 1. They are at home and someone did a study once on that (4th and 1 splits at home). I suspect a QB sneak from that distance would actually be successful closer to 75%. [Editor's note: I'm not sure if the study I just linked to is the study JKL was referring to.] [JKL's note: yes.]
Now, why do the Titans have a 60% chance in OT? Do you know they are winning the toss. I used 50/50. I actually used 20% as the chances that the Ravens scored on the next drive, I thought that was conservative. I would also need a very convincing argument as to why you would assume that what the Ravens did on 10 possessions was more appropriate than what they did all year, in predicting the next drive. I have seen that rationale too many times. "We were really stopping their offense in the second half/game, I felt that punting was the better option" And what happens next, the drive for a field goal. Sorry, I've seen too many struggling teams score when given yet another chance to call it a fluke. I could just re-run every press conference from Herm Edwards if you want.
I came up with 40/60 for the Titans to win. It's based on an assumption that the Ravens chance of scoring on drive N is greater than the Titans chance of scoring on drive N (because of both time and probable field position), plus Ravens have a small chance of winning on drive N+1.
Doug: That's a fair point that using this-game-only estimates of the Ravens' offensive effectiveness might be short-sighted.
But whether you're using only data from that game or data from the season, you have to give Tennessee a more-than-50% chance in overtime. Maybe 60% is too high, but 50 is definitely too low.
JKL: Disagree on overtime probabilities. These teams are about as even as you can get. Tennessee's late TD in the regular season to come from behind is what keeps these teams from having identical records.
Chase: I won't re-run the math since you guys have done a solid job here, just some other thoughts:
1) Yes, QB sneak or running LenDale behind one of the tackles would have been a high percentage play. Let's ignore how awesome a play action roll out to the TE would have been though, since we can't expect Jeff Fisher to do that. I'll say the odds of converting that first down were about 2/3.
2) Once again, the two keys for me were: a) a later FG is much better than an earlier FG. b) Corollary of that - I just don't see the Ravens coming out of their shell following an unsuccessful 4th down attempt. Baltimore gets the ball at their own 11, we're going to see 1st and 10 run to McGahee left side for 1, 2nd and 9 run to McClain up the middle for two, 3rd and 7 draw to McGahee for four. Then Tennessee gets the ball back at their own 45. They'll need to travel a bit of yards, but they'll have enough time to go for the TD if they want. And since Bironas is very good, once they get 30 yards they're fine. They'll also have four downs to pick up every first down.
JKL's right about the teams turning it around at the ends of games, but the sidenote is they do it when they need to. Getting that early FG meant the Balt OC was now thinking "hmm, tie game, our ball, 4 minutes to go. We can play for OT, but we might lose the toss and our D is tired. This is practically overtime, anyway, except we just got the ball first. We're basically playing sudden death now -- i need a score."
3) I agree with Doug that Baltimore would be a dog in OT, but I'd say 45/55.
This entry was posted on Sunday, January 11th, 2009 at 11:06 am and is filed under General. You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed.