May 9, 2013

Do hot streaks exist?

The enlightened conventional wisdom is that hot streaks in sports don't exist. But, as I pointed out in February, that is probably simplistic. 
Are ‘Hot Hands’ in Sports a Real Thing? 
Winning streaks in sports may be more than just magical thinking, several new studies suggest. 
Whether you call them winning streaks, “hot hands” or being “in the zone,” most sports fans believe that players, and teams, tend to go on tears. ... 
But our faith in hot hands is challenged by a rich and well-regarded body of science over the past 30 years, much of it focused on basketball, that tells us our belief is mostly fallacious. ... 
Now, however, some new studies that use huge, previously unavailable data sets are suggesting that, in some instances, hands can ignite, and the success of one play can indeed affect the outcome of the next. 
In the most wide-ranging of the new studies, Gur Yaari, a computational biologist at Yale, and his colleagues gathered enormous amounts of data about an entire season’s worth of free throw shooting in the N.B.A. and 50,000 games bowled in the Professional Bowlers Association. Subjecting these numbers to extensive (and, to the layperson, inscrutable) statistical analysis, they tried to determine whether the success or failure of a free throw or a bowling frame depended on what had just happened in the competitor’s last attempt. ...
In these big sets of data, which were far larger than those used in, for instance, the 1985 basketball study, success did slightly increase the chances of subsequent success — though generally over a longer time frame than the next shot. Basketball players experienced statistically significant and recognizable hot periods over an entire game or two, during which they would hit more free throws than random chance would suggest. But they would not necessarily hit one free throw immediately after the last. ...
But if winning streaks have some rational basis, then by inference so would losing streaks, which makes the latest of the new studies, of basketball game play, particularly noteworthy. In that analysis, published last month in the journal Psychological Science, Yigal Attali, who holds a doctorate in cognitive psychology, scrutinized all available shooting statistics from the 2010-11 N.B.A. season. 
He found that a player who drained one shot was more likely than chance would suggest to take the team’s next shot — and also more likely than chance would suggest to miss it. ... 
The reason for this phenomenon might be both psychological and practical, Dr. Attali wrote; players seemed to take their second shots from farther out than their first ones, perhaps because they felt buoyed by that last success. They also were likely to be defended more vigorously after a successful shot, since defenders are as influenced by a belief in hot hands as anyone else. 
But what the findings underscore, more subtly, is that patterns do exist within the results. The players were more likely to miss after a successful shot. And this anti-hot hand phenomenon, said Dr. Yaari, who is familiar with the study, was itself a pattern. “It is not completely random and independent” of past results, he said. 
These new studies do not undermine the validity of the magisterial past research on hot hands, but expand and augment it, Dr. Yaari and the other authors say, adding even more human complexity. Yes, we probably imagine and desire patterns where they do not exist. But it may be that we also are capable of sensing and responding to some cues within games and activities that are almost too subtle for most collections of numbers to capture. 
“I think that our minds evolved to be sensitive to these kinds of patterns,” Dr. Yaari said, “since they occur frequently in nature.”

Cold streaks certainly exist, so maybe hot streaks are just the absence of cold streaks. 

Consider Pedro Martinez at the zenith of his pitching career, in the hitters park of Fenway during the steroid era of 1999-2003. Over those five years, he went 82-21 with a 2.10 ERA, 2.28 times better than the average pitcher in the league, which is astounding. In some ways, that's about as good as any pitcher has ever been. 

Yet, Martinez had numerous cold streaks during those five years where he didn't have even a chance of winning: during those five years, he missed about 40 starts, over 20% of his opportunities. 

Martinez is not very big for a pitcher (5'-11" and 170 pounds) and baseball teams have gotten more enlightened about not overstressing their aces. When Martinez would get slightly injured, the team would sit him down, maybe put him on the disabled list. In the past, the manager probably would have told him to man up and pitch through the pain. So, the more modern pattern is to have a Martinez pitch only when he was hot, and to yank him for a reliever or a substitute starter as soon as something goes off-kilter. 

In the past, Martinez would have had more cold streaks and hot streaks while he struggled through his troubles. Modern managing does a better job of getting pitchers out of the game entirely, so their cold streaks can only be inferred from starts missed. Now, it looks like Martinez was hot all the time for five straight years, but in reality he missed the equivalent of a full season over that half decade.


Anonymous said...

What about the related concept of "runs" in basketball i.e. stretches of time where one team scores a bunch of points while limiting the other team to zero or few points?

sunbeam said...

Do these guys have any idea of the nature of the sports they are supposed to be investigating?

In basketball it's pretty common for a guy to be a "streak shooter" (as opposed to a garden variety shooter). Of course even guys who aren't streak shooters can get the "hot hand."

It sounds kind of funny writing about it, but whether the guys that use these terms have Phd's in statistics or not, they have been involved with the game a long time and are trying to win.

Larry Bird had a lot of streak shooter in him with his jumper. The all time poster child for streak shooters was Vinnie Johnson, the "Microwave" of the Pistons. Watch some footage of the Pistons. They would stick him in, if he was cold they would take him out unless someone had foul trouble. They would stick him back in, sooner or later he would get hot and no one could really stop him. You can almost always get a jumper off if you are willing to accept a suboptimal shot, and when Vinnie Johnson was hot it was going in.

A lot of us that are middle aged now remember playing lots of pickup basketball and horse. We all know what it is like for the ball to start coming off your hand, and you get that sweet feeling. Remember how it felt to shoot one of those running skyhook shoots from half court and know it's going in? And then do it all over again?

Sometimes our impressions lie to us, but it seems to me this phenomena is a fact that stat guys haven't come up with a good way to measure yet.

I guess this is a case where you have to decide to believe these guys or your lying eyes. I'm particularly curious as to why one guy decided to investigate free throws. That really isn't something that seems to be a hot hand phenom, because it is split up over the game. The hot hand is something that is short term and lasts a couple of minutes, though you might get several in a game.

Calvin Johnson had that ungodly consecutive free throw shooting streak in the 80's, but that really wasn't a hot hand thing, I mean it took weeks of play to do it.

I don't know much about baseball, but when a guy starts really popping the ball off his bat it's pretty obvious. Guys whose jobs are determined by this, will put a guy in the lineup because he is hitting well lately.

Luke Lea said...

If these streaks exist why can't people make money by betting on them?

Anonymous said...

Anyone who every played baseball at a level beyond Little League is well aware of the reality of hot streaks. When you're "hot", it didn't matter if the opposing pitcher was the ace or the scrub long reliever, the ball actually looked bigger and easier to hit.

The opposite is just as true, when you were "cold", the ball literally disappeared on you.

MKP said...

A very interesting topic. I've read a few studies on it (not sure how I missed your piece in February, Steve).

Anyone who has played basketball on a serious level - to the point where you're practicing every with your team, or at least playing consistently several times a week against good competition - knows that hot shooting streaks are real. If you're a jump-shooter who plays basketball a lot, you know the feeling when you've hit a few in a row, including some where you were off balance or at a tough angle, and you can tell you're on.

No idea what the numbers say, the science might disagree. But if you pulled me out onto a court on any given day and let me shoot 3-pointers in a game environment, I'd probably hit 35 to 40 percent of them. But I can distinctly remember days when I hit 5 or 6 in a row, and every time you're on offense, you're just thinking "if this guy gives me an inch, I'm going to stick another one in his face."

DWBudd said...


I remember reading the 1985 paper when I was in graduate school. I read it in context of some research I was doing with respect to defining true 'randomness' using, inter alia complexity and compression suggested by Kolmogoroff-Chaitin.

As a former baseball player and current (poor) golfer, the idea of a "hot hand" is one that I find compelling, even if the mathematical evidence is weak.

Skill sports involves at the centre a complex, physical action that is heavily repetitious. No doubt, you've heard of golf instructors drilling into their charges the importance of a consistent swing. Breaking down the backswing, weight transfer, follow through into simpler parts, and then practicing them so that you do the same thing every time.

I believe that success on a pitch, or a free throw, or a golf swing can provide a certain reinforcement and comfort, and the player can have an apparent "hot streak" when that repeated movement gets more or less automatic.

The converse is also true.

But for team "hot streaks" of the sort that the author of the article suggests (in this case, the Chicago Bulls), I am much more sceptical. Runs (in basketball) are in the similar vein.

Ex Submarine Officer said...

If anyone has been on a hot streak lately, it is Steve Sailer.

Anonymous said...

you gotta know when to hold 'em, know when to fold 'em.

Pat Boyle said...

True randomness is a mathematical concept. In the physical world be get something like that when quantum effects bubble up in to our reality. We also get a pretty good approximation from instruments that detect Brownian movements.

All of these are very low level events. There are also pseudo random events that are part of a chaos process. Chaotic events are perceived of by us as random even though they are completely determined. Simple equations go into chaos after some point and look completely unpredictable. But the next point is always uniquely determined.

Human performance is not like any of these. Our events look random only because they are complex composites of factors that we cannot compute.

Our long term free throw average seems to be composed of individually random shots that follow that long term probability. But humans actions are random only in a cumulative statistical fashion. Like in chaotic processes they should be analyzable but in practice they are far too complex.

For example, what is the body temperature of those who are on a 'streak"? How many hours of sleep? Their last sexual intercourse? There are thousands of these individual factors that might be relevant. Most of the time those that improve performance are balanced by those that hinder performance. They tend to balance out but obviously the average performance is a complex composite of many, many factors.

So our expectation should not be that an athlete's performance is steady state and 'hot streaks' are outliers. Rather we should expect that everything is a streak but most of them are too small to notice most of the time.


Mr Lomez said...

Hot streaks are absolutely a thing. The "enlightened conventional wisdom" is...misguided.

Of course, a 40% shooter sinking 8 of 10 threes in a game or something will happen according to normal variance, just as, over enough seasons, a .280 hitter will hit .600 over the course of a 4 game series.

The difference is, in sports, unlike say blackjack, the athlete is totally conscious of his improved performance while it's happening. It feels different. Everything happens in slow motion. The physical movements become entirely automatic. It's hard to describe the experience without resorting to metaphysics.

In any case, this study is pretty lame. Using free-throws is a terrible way to judge how "hot" someone is, since free-throws, by their nature, happen outside the normal flow of the game. You'll in fact often see coaches try to cool off a hot hand by putting a shooter on the line.

JeremiahJohnbalaya said...

I like the idea of concentrating on exactly what happens after a given made basket(eg. shot percentage decrease, the next shot is typically longer,etc). That seems much more reasonable, controlled, and likely to be more meaningful than the "big data mining" analysis.

The psychological experience of hot-streaks is truly mystifying. I grew up playing basketball constantly, but in college I was playing for hours every single day (rain, snow, exams, hungover, whatever, it didn't matter) against good competition (NCAA scholarship players, first round NFL draft picks, etc) and maybe experienced something like the Gladwell 10,000 hour thing. And there is no doubt that there were times when I didn't even have to look at the rim. And it wasn't just shooting that seemed to be affected; but also dribbling, timing blocked shots, etc. And I really think it was different than the psychological conditioning of fearlessness of your opponent, which I think was more like a prerequisite.

I've been practicing guitar a whole lot recently, really for the first time in 15+ years of noodling, and I am just starting to sense something similar might be possible there.

WRT to someone else's comments, I always felt like the free-throw line was the place to go if I wasn't feeling my shot. I think it was a matter of being able to go there and recover the consistent form from all the years of practice.

Anonymous said...

If these streaks exist why can't people make money by betting on them?

Cause odds makers see them too.

Cail Corishev said...

And there is no doubt that there were times when I didn't even have to look at the rim. And it wasn't just shooting that seemed to be affected; but also dribbling, timing blocked shots, etc.

Yes, there are days when everything just goes in. Not only do you make every shot, but it seems effortless -- you're rising higher on jump shots and rebounds, your shots seem to take no effort at all, and everything just works.

Then there are days -- often the very next day -- when you're bricking everything and tripping over the ball.

So I have no doubt that there are hot games. I can't say that it extends to the next game or practice with any predictability, though. Other than basketball, billiards is the only other skill game I've played enough to get in the zone, and it's the same way: one night I can't miss, but tomorrow is another day.

If a hot streak doesn't extend beyond the current game, it's not very useful for predicting anything. It's no help for gambling. It might help the coach decide who to have holding the ball on the last play, but it won't help him plan his roster for the next game, let alone the next season.

TontoBubbaGoldstein said...

If these streaks exist why can't people make money by betting on them?

Who's saying that they don't?
--The Ghost of Lefty Rosenthal

Anonymous said...

the mental phenomenon is real.

the streak sounds like T boost once a male wins, thus increasing his chances to win again. and again. and again.

ben tillman said...

Calvin Johnson had that ungodly consecutive free throw shooting streak in the 80's....

That was Calvin Murphy who set the record (since broken) of 78 straight, although Kevin Johnson did hit 57 straight in 1989.