Well, shit, you’ve cracked the pattern. I wonder if I can get someone to make bets with me during the games. “Hey, I bet you Peyton will fuck up this next drive…” “Hey, I know he’s gonna score a TD this one…” and repeat. I’ll be rich!
Not to get into a whole new thing with you, but the desperation drive at the end of the game – where you get 4 downs instead of 3, you’re up against a tired defense, and you never stop to kick a field goal – is not at all the same thing as a drive in the middle of the 2nd Quarter. In-game TDs/drive has almost no bearing on the analysis.
Come on, friend. Surely you can concede that going for it wins the game right there *roughly *half the time, right? (Just look at what happened: the pass hit the receiver in the hands, he should’ve caught it.) And that if they fail, they’re still in it? Surely that alone is enough to make it close, even if you still think it was sub-optimal.
I don’t concede those point in the spirit you offer them, though I do concede them mathematically.
Would you think it was a borderline call if they did that while only up by 2 instead of 6? Because the points you ask me to concede would still have to be conceded.
I missed this. Very cool little thing. Everyone discussing this should be required to do it.
My result:
Results
Chances of winning if you go for it: 50%
Chances of winning if you punt: 50%
Verdict
I wasn’t on board with the call at first, but the numbers don’t lie, smart gutsy call by Belichick.
The Colts scored touchdowns on 31% of their previous possessions; I would rate their chances of scoring after a NE punt as higher than this.
Prior to their last possession, the Colts’ best starting field position was their own 25, and their average starting field position was their own 20; after a punt they would likely have had better field position than this. In their previous possessions, punting was an option; after a punt they would be playing all four downs. In their previous possessions, the NE defense had been playing reasonably well, but they appeared to be wearing down and less effective in the fourth quarter.
That field position thing surprised me, that has to be relatively unusual to not have any decent starting field position.
Cool. My personal estimations are probably a little generous to the Pats, but my numbers come out somewhere around 65-70% for going for it and 40-50% for punting.
The bolded part is very relevant. It’s the reason why changing overtime to be “both teams get a possession” is just as unfair as sudden death. (With added injury risk due to running more plays.)
So the Colts get the extra-down advantage added to 31%. To keep with the spirit of making up numbers, do we call that 40%? Three downs = 31%, Four downs = 40%. Seem reasonable?
I have a followup question because I’m having trouble opening the play-by-play pdfs from nfl.com: What were the specific results of all the Patriots punts that game? Punted from the ?, Colts took over at the ?.
Interesting question, actually. Obviously failure would be much more disastrous, but you also have more to gain by preventing the Colts from getting the ball back, because their drive from their own 30 will probably succeed (instead of probably fail) if all they need is a field goal.
Being as generous as I can possibly stomach to the idea of punting, assume that: New England converts 45% of the time, if they fail to convert they hold on to win just 5% of the time, and if they punt Indy scores just 45%. In that case, punting wins 45% of the time, and going for it wins 47.8% of the time. A virtual dead heat. Punting might make up the difference, actually, if we factor in that making the first down but giving the ball back to Indy with 25-30 seconds doesn’t necessarily win you the game if all Indy needs is 3.
Basically, though, I think going for it is still best if New England is up by 2. I’m sure you won’t like that answer. Interesting angle from which to tackle the question, though.
Since 2001 (and Brady), the Pats have coverted 76.4% of all 4th and 2 or less. Of course that’s averaged over all their attempts, including half-yard QB sneaks, but I thought it might be an interesting number.
I don’t think my personal estimation in the 60% range was unreasonable, but as you can take it as low as 40-50% and still get it to come out as roughly even or even still favorable.
Any way you slice it, it’s nowhere near “OMG WORST CALL EVER!!!” territory.
I managed to open the pdf. Here are the four NE punts:
NE 39, Ind 10
NE 20, Ind 25
NE 43, Ind 24
NE 45, Ind 18
If I had to put a number on it, I’d say if the Patriots punted the Colts would have started somewhere between their own 20 and 25.
I don’t follow your math. If Indy scores 45% after a punt, how does punting only win 45% of the time?
EDIT: And I maintain that 45% is high for Indy based on the game they played to that point. 40% seems a reasonable figure factoring in the 4-down boost to their 31% success rate.
Whoops! My bad. Inverted the results in my head. Those assumptions mean that Punting wins 55%, Going For It wins 47.8%. Still pretty close.
ETA: If I plug in estimates that I think are more accurate (Pats convert 55%, fail to convert and win anyway 8%, punt and win 45%), then going for it wins 58.6% and punting wins 45%.
If you get the same chance of winning whether you punt or go for it, why is going for it on fourth a “smart, gutsy call”?
Agreed on the OT point. Getting the ball second in college-style OT is a bigger advantage than getting the ball first in NFL OT.
40%? That’s reasonable, though I’d put it higher myself. One thing about the earlier drives was that both INTs were really uncharacteristic for the Colts. The first one was just a terrible wounded duck throw by Peyton, and the second was that miscommunication with him and Wayne. I’d say that these were rare plays that aren’t likely to happen again, so I’d raise the Colts chances of scoring. Others would probably say they show that the Colts were a bit off, and lower their chances.
But those “more accurate” numbers are not only arbitrary, they directly conflict with the evidence at hand. The Colts scored on only 31% of their drives; they had to punt like 7 times. And not all of those punts were on 4th and short; most were between 4th and 7 and 4th and 15, and many of those long 4ths followed an incomplete pass on 3rd. Assuming 3rd down performance is roughly analagous to 4th down performance, anything approaching a 50% chance to score a touchdown following a punt is totally without merit.
Those weren’t his words; he was quoting the website.
I was saying they had a 55% chance to score a field goal (or touchdown). For a needed TD, I assumed upthread it was about 40%, which (I believe) you said was reasonable.
ETA: I’m really not trying to cram down your throat the idea that going for it is clearly better (whether the margin is 2 or 6). There are all sorts of variables at play that we can’t quantify, or that might not even occur to us. I’m just saying that the bare mathematics of the most obvious facts means it would be hard for this not to close call, one way or another.
Because regardless of the actual probabilities, going for it will get you lauded or lambasted the next day while punting while get you neither (except by Gregg Easterbrook).
Patriots fans sputtering about statistics and alternate reality wins = salty caramel on the ice cream of a perfect Patriots loss.
I’d be a bit surprised if there are any Patriots fans here arguing for going for it. In general, Pats fans will be outraged by the decision.
I was rooting hard for the Colts last night.
All I know is that I’m very eager to hear what the Sports Guy has to say about all this.