If you watched yesterday's game between the New England Patriots and Indianapolis Colts, you undoubtedly saw Bill Belichick's failed decision to go for the first down on fourth and two from his own 28, up six points, and with 2:08 remaining in the fourth quarter. Well, unsurprisingly, there has been a lot of backlash against Belichick for making that call, and i'm here to say that if you think Belichick's decision was wrong, you're probably dumb.If that sounds elitist it's because it sort of is. Sorry, but there's no way for me to sugarcoat it.
Brian Burke of Advanced NFL Statistics (outstanding site, by the way) explains the situation best.
New England coach Bill Belichick is taking a lot of heat for his decision to attempt a 4th down conversion late in the game against the Colts. Indianapolis came back to win in dramatic fashion. Was the decision a good one?Basically, even not accounting for the context of the situation, Belichick's decision resulted in a 9% increase in the chances that the Patriots would win the game. Throw in the context that the Patriots have Tom Brady, the Colts have Peyton Manning, and that the two teams combined for 884 yards, and there really is no statistical argument against going for it in that situation.
With 2:00 left and the Colts with only one timeout, a successful conversion wins the game for all practical purposes. A 4th and 2 conversion would be successful 60% of the time. Historically, in a situation with 2:00 left and needing a TD to either win or tie, teams get the TD 53% of the time from that field position. The total WP for the 4th down conversion attempt would therefore be:
(0.60 * 1) + (0.40 * (1-0.53)) = 0.79 WP
A punt from the 28 typically nets 38 yards, starting the Colts at their own 34. Teams historically get the TD 30% of the time in that situation. So the punt gives the Pats about a 0.70 WP.
Statistically, the better decision would be to go for it, and by a good amount. However, these numbers are baselines for the league as a whole. You'd have to expect the Colts had a better than a 30% chance of scoring from their 34, and an accordingly higher chance to score from the Pats' 28. But any adjustment in their likelihood of scoring from either field position increases the advantage of going for it. You can play with the numbers any way you like, but it's pretty hard to come up with a realistic combination of numbers that make punting the better option. At best, you could make it a wash.
Of course, there will always be those fans who can't understand statistics, or don't care to take them seriously, but even if you are one of those people, the decision still made complete sense.
By choosing to punt, the Colts would get the ball around their 30, resulting in a 40 yard gain in field position. That's not much of an advantage to get when you consider that the Colts have an elite offense and Peyton Manning running it, especially when the Colts essentially run a two minute drill as their standard mode of operation.
Perhaps more importantly though, with the game on the line, don't you want to put the ball in the hands of your own elite offense and Hall Of Fame quarterback? Why wouldn't you go to Tom Brady when the chips are down and you need a big play? How is that not better than simply praying Peyton Manning screws up? Furthermore, even if you don't get the two yards you need, your defense still has a shot at preventing the Colts from scoring a touchdown. It's not that hard, really.
On a side note, it sort of disgusts me that I have to defend the Patriots and Belichick.
So how does this discussion correlate to baseball at all? It doesn't really, but the negative fan reaction to Belichick's decision makes a point about analysis and logic. Point being, the common fan hates both of those things.At times, I get the feeling that certain fans would justify pinch hitting Eric Hinske for Albert Pujols if it ended up working. Yet, I would bet those fans would be the ones complaining the loudest when the same thought process fails time and time again in the long run.
In that same vein, I suppose this situation is a great opportunity for me to drive home how stupid results based reactions are.
If the Patriots pick up the first down, Belichick gets labeled a genius by the media. Since the Patriots ended up getting stopped, Trent Dilfer almost has a brain aneurysm on live television and everybody thinks Belichick is an idiot.
Why does this disparity in perception exist? It doesn't make sense to me. Why can't something be completely correct and just happen to not work that one time? Stuff happens. Ugh.
Sorry for this off-topic rant, but I had to get it off my chest in a public forum. The Winter League updates are coming soon.
37 Comments:
the other thing missing from the fans' thoughts is that he was an inch or a favorable spot away from getting the first down.
Even without this great statistical breakdown I was surprised at the uproar over that call. They saw the chance to put the game away and they took it. Not a totally unreasonable call and probably something they've done before.
i agree with the post and the comments, but shouldnt the wasting of one or two timeouts be involved in the discussion of whether belichick managed those last two minutes correctly or not? it's hard to quantify the value of a timeout given the time, field position, offenses, etc. but i suspect that overall, the win probability added of going for it (on the 25) over punting is less than +9%...
Oh no don't start with stats on football too...
Well I guess I have to disagree with the strength of your convictions in the same way I disagree with the strength of the fans' convictions as demonstrated by the answers to the ESPN poll (which is flawed anyway since it is biased by the fact the outcome is known). I don't think the "right" answer is quite so obvious.
If you took the view of a statistician, you would need to consider such factors as the variability of the underlying data, size of the sample base, etc. to determine the answer suggested by statisitics. My guess is that those factors would push your "answer" into a grey area where none of the projected outcomes is statistically clear-cut. Since we are discussing a situation that is impacted by many factors, I really don't think it can be reduced to the simple formula that was used and categorically say there is a 9% better chance of success with the option chosen by the Pats. Certainly, using league-wide data is a significant weakness in the projection.
I see this as too close to call one way or the other. Belichick had to make at least in part a judgement call, which he did and it didn't work out. I don't think it was as bad a call as some of the pundits are suggesting nor was it as clearly the right call as this blog is suggesting.
McCheap-I know, right? Imagine wanting to have logic behind your convictions and not just throwing out crap because it's an exercise in risk aversion.
Luke-Most of your concerns about where the data is coming from and the sample size are dealt with on his site already. He states his facts so generally because he assumes he already has a baseline of people that read his site that understand where he's coming from.
Using league wide data isn't a significant weakness because including context to account for that specific situation only furthers my point.
You have Tom Brady and an elite offense. You have had a ton of offensive success against this specific team on this day. You are historically successful on 4th downs. You are historically successful on 4th downs within your own territory. You are historically successful on 4th and short. You're taking the Colts defense by surprise. The receiving team has Peyton Manning. The recieving team has had a ton of offensive success against your team on this day. He just led two long drives down the field for touchdowns against your defense in around two minutes in the quarter. Your defense is exhausted, as evidenced by the lack of depth on the defensive line coming into the game and the drive that just concluded leading to the scenario itself. Etc.
Like I said, statistically, using league average numbers, it's a very sound decision. Including context just makes it seem all the more intelligent.
It's a no brainer to go for it to me. The only reason more people don't see it that way seems to stem from traditionalist views of what a coach is supposed to do. Belichick isn't successful because he sticks to the book, after all.
A.L.-Belichick's managing of timeouts was a much bigger mistake. Completely preventable.
Kensai - Dead on here, and you don't even need the percentages to work it out. Belichick might be a piece of shit, but he made the right call there. Indy has players in the secondary hurting and your offense has been able to pretty much dictate to them for most of the game. Meanwhile, the Patriots' defense is pretty obviously both gassed and figured out by Peyton Manning, so it is a matter of when, rather than if the Colts will score. In fact, if the Colts were to score on a quick strike from the position they were in, which is always a possibility, the Pats might have come out of it with a chance to run a few plays.
4th and 2 is a clear call in that situation. You don't want the ball in Peyton Manning's hands, with the 2 minute warning, a time out and the Pey-Gun (anyone old enough to get that reference?) offense likely to score on the next play.
The time out mismanagement was clearly a killer in that game, as was the smart time out usage by the Colts. It also helps to have the better quarterback, which Peyton Manning clearly is.
Yeah, I forgot about the fact that the Colts secondary is decimated.
At least I made a ton of money on the over. :o
Kensai - It would have almost been the better play to throw a bit deeper pass to take advantage of the secondary. Oh well. Good prevailed over evil.
Incidentally, a ton of the talking heads are absolutely shitting on Belichick for the decision.
One more reason to hate me:
I'm a Colts fan!
However, Billy-Boy knew that the Colts would score and he had a 70% chance of getting a first down. It's a no-brainer. I hate "Belicheat", but he made the right decision - it just didn't work. If they had gotten a first down, everyone would be talking about what a genius he is.
BTW,
I live in Indy!
Honest question: Does almost every football fan beside Pats fans hate the Patriots simply because they win more games/titles than any other franchise? Kinda sad, IMO...
If you were not so "dumb" you would have stated that the errors resulting from those statistical calculations would have far exceeded the 9% you demonstrate as proof that going for it was the "correct" decision.
You throw these numbers around like they are absolutes when in reality there is so much error built into those percentages its just silly to treat those like the gospel.
Please don't be so quick to judge lest ye be judged and perhaps break out the statistics textbook.
LDY4L: yup. You're a yanks and laker fan. same thing with them.
i hate the Pats because i just hate all thing Boston. Celtics, pats, sox, the people there(lol even though i've never been there).
just dont like boston. they hate L.A., and we hate them
LDY4L - Screw the Pats, they haven't won more than my beloved Steelers!
Ed-Oh? The burden of proof is not on me though.
The guy who runs Advanced NFL Stats has his numbers based in empirical evidence, so the burden of proof is on you to show that the margin of error is greater than 9 percent. Even then, that's only what it would take to even this decision out for league average teams, much less within the context of the situation.
Feel free to breakout the statistics textbook and solve this problem. Then, when you solve it, please do me a favor and forward it to all the mathematicians and show them why you're right and they're wrong.
Personally, I have yet to see any decent argument against Belichick's decision that didn't have a bunch of rhetoric about how "football is supposed to be played", how "statistics are just numbers" and the people who use them are "nerds", or some childish but vague jab at the person presenting the numbers.
Your critique falls in the third category. I'm sure you won't be back to respond to this but since you are obviously smarter than this guy and the guy who runs Advanced NFL Statistics, please send them your calculations as well, so they can correct themselves. Thanks.
Alireza-His mistake was that he knew he was going for it on 4th down, so:
1) He made a mistake by not informing the rest of the team clearly enough that he wasn't going to punt and had to burn a timeout in the confusion.
2) He probably should have ran the ball, which would have likely made it a 4th and 1 or 4th and inches, thus making it necessary for the Colts to account for the possibility of a sneak.
LDY4L-I hate BB because he is an arrogant smug dick with no class. I don't see how that's an arguable point. Then again, he deserves to be after all he has accomplished and it helps him be successful. Do I think he cares what people think of him? No, that's why he's so great.
Oops, proving it just got harder. :o
Chad - I think the debate over this call is an outstanding thing (for all sports) and, much as I dislike Belichick, applaud him for the intellect and cojones he continually exhibits in going about his job. Situational statistics are under-used in the sports arsenal, generally because decision-makers do not understand how to apply them.
As important as applying statistics, though, is understanding their limitations. As I noted in my post yesterday, I don't think this is as clear-cut as you state it is. Assessing the context in a football setting is complicated because of the number of potential factors affecting the outcome. Statistics naturally are more reliable in controlled settings with a limited number of discrete outcomes and lower numbers of interdependent events.
I grant you that certain factors in the context of the Pats-Colts game argued for trusting the game to the Pats offense. But I could argue that the statistics do not consider, in that situation, the highly unlikely event that the Pats will go downfield on the play, tipping some of the percentage differential you quote to the defense. You state, accurately, that Manning was having a big offensive evening so turning the ball back to him was dangerous. I could argue that Manning had endured his first multi-pick game of the year, which is probably not captured in the statistics, and that should turn some advantage to the Pat defense assuming the Pats punted.
The point not being to argue with you on the call but to make a point that it is not so cut-and-dried as jumping onto statistical formulas as "proving" anything. The sites you reference (excellent by the way) do make note of this in referencing limitations of sample size, the need to judgmentally adjust numbers for difficult-to-measure factors, etc.
Anyway, we should appreciate the value of the debate and recognize that nothing is really served by the pundits calling Belichick an idiot for the call he made, nor by you calling everyone "dumb" who disagrees with what is by no means a no-brainer decision.
http://en.wikipedia.org/wiki/Standard_deviation
LDY4L - It wasn't that long ago that the Pats were mediocre, with occasional flashes of brilliance (Super Bowl under Parcells) and a whole lot of untapped potential. I actually cheered for them against the Rams, because I have hated the Rams since 1995. My issue with the Pats is that Bellichick is a piece of shit, not really anything else.
Kensai - I agree and I'll add that Indy had a tough time dealing with the Pats' ground game. They may well have gotten the first down by just running.
bayareadfan, very true. I never watched FB in NY and my best friend who introduced me to the NFL is a Steelers fan. They are one of the two teams I root for, as the Jets are the others(I root for the Jets because they are from NY, although I never really rooted for them...)
I honestly have nothing against the Pats, other than I dont like Boston. However, if you go to my school saying you're a Pats fan, someone will try to jump you(lol). Maybe its because I dont have 'a team', but I really dont give a s*it about what the Pats do or dont do...
BTW, I hate the Giants. I dont know why, but I dont like them :)
http://www.mlbtraderumors.com/2009/11/several-teams-interested-in-smoltz.html
Dodgers interested in Smoltz...
Alireza, that may be true, but when you have a HOF QB(and yes, Brady is a HOF, no matter if you do or dont like the Pats), you put it in his hands...
If you got the best punter in the league like the chargers do in a dome....you punt that bitch
I agree with Tom Brady here. I'm paraphrasing, but he said that if you told him you could guarantee a win if they could gain two yards over two plays, he would take it all year long. It didn't work obviously, but the intent was likely in the right place.
Luke-It's quite the no brainer. I don't see how you can state that the statistical models don't prove anything when they clearly do prove that certain events are more likely to occur.
Again, your context argument stems from a lot of specific events and scenarios, which is fine, I suppose, but for every unlikely event somebody could throw out as a negative to going for it, I could throw out one that is a positive. That continues ad nauseum until somebody gives up, but that's what the statistics part is for. It provides hard data needed that accounts for all those crazy situations. It's built into the system by definition.
You seem to be caught up with the semantics of the term "no brainer". Well, obviously you would need to think about it, but the point is that it wasn't a hard decision for a guy who relies on the numbers like Belichick does. The hard part is getting called out by dumb people if it fails, and that's what's happening now, and that's why the NFL is such a case of risk aversion.
There's little value in this debate besides a bunch of mouthbreathers on television ranting about some appeal to authority fallacy and bashing unconventional thought. If anything it's illicited negative reaction to forward thinking, as evidenced by that article linked on fangraphs.
Ed-What is standard deviation supposed to prove?
I'm asking the two of you to prove it if you're so sure that it's not an easy decision.
I can't prove 100% that Barry Bonds is better than Juan Pierre either. 99.9% sure, but perhaps over 100,000 more plate apperances something would have changed.
I just go with what would most likely result in a positive outcome. Therefore, it wouldn't be a difficult decision in a vaccum. What makes it difficult is the reaction occuring now.
Point being, if you have numbers and evidence on your side, it's not your responsibility to continually have to prove something is true.
I don't have to prove Juan Pierre is not a good baseball player. Whoever disagrees with me has to show me evidence to the contrary.
Similarly, I haven't seen anything yet that resembles evidence based on probability or logic that leads me to believe it was anything but the correct decision.
Nothing posted here changes that either. All the arguments I have seen made thus far have already been made in the Advanced NFL Stats comments.
mgrumsb-Yes, that would logically sway the numbers in the opposite direction of the league average numbers. Therefore it would require further evaluation.
Unfortunately for the dissenters, in this situation, the context and situational statistics do nothing but suggest more (or at least equal) than the stated percentage of the initial study.
Chad - OK, I give! I am not trying to prove whether Belichick did or did not make the right call, or whether you are right or wrong for agreeing with him. Hell, even after all of this debate, I am not sure what call I would have made in that situation (which I guess makes me an idiot in your book, but I suspect that is a thick book).
I am only making the observation that there are many situational contexts in a sports contest, many more than just that the two offenses are strong and were playing well that you point to as your situational context. Many of the other situational contexts are impossible to measure, are not reflected in the statistics, and render the 79-70 margin very close for making conclusive decisions.
If you prefer to believe that the Advanced NFL Statisitcs are absolute truths, so be it. I would though ask you to explain to me, looking at his graph of the actual touchdown success rates from various field positions with two minutes to play, why teams are 31% successful when they have 81-85 yards to go and only 22% successful when they have 61-65 yards to go. He has "wiped away" this differential by turning this into a linear graph (which in itself is debatable since there is no guarantee that, with a larger population, there will be the reversion to a mean that this implies - there could be a legitimate reason for this difference). Interestingly, if you use the actual experience rate for success of 22% with 61-65 yards to go instead of the 30% he used in the calculation in his article (his "straight line" value), the decision to punt vs. go for it is an arithmetic push.
"Statistics are like women - they can be the mirror of purest virtue and truth, or like whores to use as one pleases" -Theodor Billroth
It's not about giving or taking, i'm just stating my case, and I have yet to see anything that would convince me that Belichick's move isn't at 51% and up than 49% and down as far as winning goes
Again, I wouldn't get caught up in the verbiage, as I knew full well that calling a group of people dumb would garner attention, even from the otherwise uninterested. It always seems to work.
Of course, that's not to say I did it purely for the reaction, but I think it draws people in. Hopefully at least a couple of those people actually critically think about it for once instead of just referencing conventional wisdom.
From your posts I hardly gleen that you are one of those that I am referring to, since you are obviously not the mouthbreathers I see on public board. :o
I knew people would disagree with me, intelligently (you) or unintelligently (McCheap), and that's fine, but if one side has empirical evidence, I just expect some kind of equal counterpoint in proof is all.
Same goes for baseball discussions, I think, which was the overriding point.
I don't understand why a regression line is hard to understand as a concept either. I hope people realize why he uses a regression line.
Ironically, your quote at the end applies to the dissenters more than me or the guy who wrote the cited piece.
We aren't the ones attempting to manipulate the data. I have no vested interest in this decision, I wanted the Patriots to lose, and I hate Brady and Belichick. The people who could not deal with the initial result are the ones attempting to manipulate through extravagent (and sometimes hilarious) methods. We're not the ones desperately trying to manipulate the standard model that he has used for years to prove it's wrong.
Perhaps the dissenters should read that very quote and think twice. Statistics and context. I know i've got my context, dunno about them.
I am fully on-board to the need for regression lines - I was just pointing out that in his chart he had some odd patterns, probably resulting from a small sample size, and logic would tell you they would disappear with a larger sample size.
The worst aspect of this whole controversy is now I see it being referred to as "Bill-gate". Shoot me, please!
One dimension of this not discussed is the impact of his decison on his defense's confidence heading into the remainder of the season (he can talk percentages all he wants but most of his defense will see it as a "no-confidence" vote). I believe confidence is somewhere around #1 on the prerequisites for success in sports list. I am also pretty sure Belichick could give a rats ass about denting their confidence.
How much fun do you think Belichick would be at a cocktail party? Interesting, yes; fun, no.
I am done with this topic, back to baseball! Interesting discussion today on Grienke.
I think that's where having three rings and being known as a genius better than everybody else comes into the territory. Also having a leader known as one of the best QBs in history supporting him.
Hard to say anything against those two.
Post a Comment