Are ‘Hot Hands’ in Sports a Real Thing?

[ElNono]

Statistics is a mathematical system. Epistemlogically speaking, math and science are about as far apart as two knowledge systems can get.

Science uses statistics, but statistics themselves are not science.

Statistics (science)

[Chinook]

It depends on what people mean by hot hand? If they mean the statistics independent of any other factors, then no, it doesn't exist. Thinking so pretty much causes the gambler's ruin scenario.

Every statistical event is independent, so the probability of something occurring again is not related to the what happened in the past. For example, the probability of getting a heads on a fair coin is .5. That's true even if the last 20000 flips before it were heads. If humans were like coins, then believing the hot hand would be ludicrous.

But applying statistical theories at such a shallow level of analysis is fool-hardy. A player is not an automaton who has a fixed probability distribution every time he shoots. His success rate depends on a lot of situations, such as confidence, as mentioned above. Also, things such as the physical state of the player, crowd noise, the temperature of the ball etc. can come into play. If we could quantify all of those factors, we would indeed find certain situations in which a player has a higher probability of making shots. The hot-hand idea is just a non-statistical way to describe that phenomenon.

[phyzik]

Having made several shots in a row gives you confidence, and confidence improves your performance in sports. As someone mentioned, golf is a great example. You have less distracting thoughts, you simply don't think/question yourself, your motions are more natural, your game flows. Same is true for jump shooters, I know I experience it. If this weren't the case, Tiger Woods would still be winning every single major he plays in. Maybe not every single one, but he would have won a major by now. He won 14 majors in 14 years, but hasn't won one since 2008.

I think you can believe and understand math/statistics while also seeing how confidence affects your play. Call it whatever you want, hot hands, confidence, w/e it is, it definitely is a factor in sports.

This...

"Hot Hands" is a real thing in any sport. I used to shoot rifles compe ively and every once in a while I would get in the zone, my best "in the zone" moment was at state compe ion... I cant even remember hearing the other rifles going off, all I could see was the targets. I still get a rush thinking about that day... hit a 298 out of a possible 300, that's 10 shots through each of the 3 positions, standing, kneeling and prone.... meaning I hit dead center on 28 of my 30 shots, the other 2 where judged 1 ring out of center. Even hit dead-center on my 2 practice shots before I started. I also finished a good 3 minutes before anyone else.

Granted, it was held in-doors so I didnt have to dope or use my mil dots as my scope was dialed in perfectly, but it was still an awesome feeling to walk away with that score.

[ElNono]

That makes no sense. Facepalm for passing along an article that apparently some people feel worth discussing?

smh

[Chinook]

Statistics (science)

It's a common mistake that many people make. I'm disappointed in Britannica for doing it, too. If you read the article, they aren't even describing a science; they're describing a mathematical system. Science is a term being applied too lazily today.

I guess it's fine to refer to statistics as a science, but then it loses its a priori power. Meaning, it becomes open fallacy for no good reason. If people accepted if as math and not science, then it would be a stronger discipline.

[jag]

You can tell which people in this thread have and have not played basketball compe ively, or any sports in general.

[HarlemHeat37]

In basketball arguments, there's usually:

A) Advanced stats haters that don't understand the concept of informative numbers and data..they generally use terms like "killer instinct" and use double standards..they usually dislike advanced stats because the numbers don't support their favorite player..

Ex. Kobe fans

B) Advanced stat nerds that ignore variables in basketball and stubbornly support their flawed formulas..

Ex. Wages of Wins supporters


I fall in the logical group that uses numbers, in addition to actually playing basketball and understanding cir stances, tbh..

Studies of this nature are generally supported by people that don't play sports, tbh..

[Spur Bank]

tesseractive, my comments are in bold blue

I have a concrete, falsifiable hypothesis regarding players feeling "in the zone:"

1. There is a significant set of players who can tell the difference between when their shot feels right and when it doesn't. --True

2. If you have the ability to record those players' EKGs and high-speed detailed mechanics of their shots without impairing their ability to play, there will be observable differences that corrrespond to those feelings. --True

3. There will be observable differences in those players' success rates at hitting shots of equivalent difficulty that correspond to those different mental states. --True

Physiologically, people can be anxious, tighten up, and exhibit poor mechanics because of their mental state. The opposite can be true. They might also be more or less injured, more or less fatigued, or be affected by various other conditions. --True

None of these effects show up in aggregate statistics, --Ah, the problem!!! Why don't they?

but they all can have an effect on how likely a given player is to make a particular shot at a particular moment. --True

All shots by the same guy in the same area of a shot chart are not equivalent to operations of a statistical random variable in the same way that radioactive decay is. This should be obvious, even to scientists.--True

The main issue I have with your post is that you're not arguing the hot hands theory. You're saying there are many factors that can influence a basketball shot... and yes, there are tons. But that's not what we're examining here. We're looking at whether a player on a hot streak is more likely to keep making shots.

You said "none of these effects show up in aggregate statistics." What do you mean by that? That's the big argument on my side, and it's a pretty massive argument because it's the very definition of the question. To put it simply, the argument is: Statistically, a player on a hot streak is NOT more likely to keep making shots. This can be verified by examining the data. Therefore, the hot hands argument is a fallacy.

Everything you said is fine and true, but it doesn't refute that argument at all.

[Spur Bank]

You can tell which people in this thread have and have not played basketball compe ively, or any sports in general.

I was thinking you can also tell who got a decent education.

But oddly, I somehow fall in both your category and mine.

[Obstructed_View]

I watch basketball. I've just obviously never played it.

You're welcome.

[jag]

It depends on what people mean by hot hand? If they mean the statistics independent of any other factors, then no, it doesn't exist. Thinking so pretty much causes the gambler's ruin scenario.

Every statistical event is independent, so the probability of something occurring again is not related to the what happened in the past. For example, the probability of getting a heads on a fair coin is .5. That's true even if the last 20000 flips before it were heads. If humans were like coins, then believing the hot hand would be ludicrous.

But applying statistical theories at such a shallow level of analysis is fool-hardy. A player is not an automaton who has a fixed probability distribution every time he shoots. His success rate depends on a lot of situations, such as confidence, as mentioned above. Also, things such as the physical state of the player, crowd noise, the temperature of the ball etc. can come into play. If we could quantify all of those factors, we would indeed find certain situations in which a player has a higher probability of making shots. The hot-hand idea is just a non-statistical way to describe that phenomenon.

Great post.

There are so many factors involved that cannot be quantified, so analysts are left to assume the players are robots and each attempt occurs in a vacuum. They view previous FG attempts as no different than previous coin flips.

[Spur Bank]

I don't get it. And not that it matters, but I played basketball in high school (no, not college) and still play in rec leagues.

[GSH]

A lot of people know about statistics - not many understand how to apply it. A lot of people know about sports - not many understand what goes on at a high level. And almost none of the people who try to apply statistics to sports know what the they're talking about. It's that simple.

Anyone who's played compe ive golf knows that there are days when the cup looks as big as a dinner plate. You feel like you can make every putt, and you DO make far more than normal. Baseball players see the ball larger and slower than normal. And basketball players talk about the hoop looking like a washtub. Maybe it's a matter of physical and mental abilities being at their peak at the same time. A lot of people have tried to quantify it by talking about "biorhythms". Nobody knows. But on certain days, the game is just easier than normal. And on those days, more shots go down. No "maybe", no illusion, no fallacy. The shots fall. Call it a "hot hand" or whatever you want.

As for the math? One guy shoots 3's at a 30% clip for an entire season - another shoots them at a 40% rate. What's the difference? Could be any of dozens of things. Most likely several small things. But if you watch a random guy put up a 3-pointer, do you know whether the ball had a 30% chance, or a 40% chance of going in? Nope. You can only tell that by looking at his season statistics. Observing, you can only see that it went in, or didn't go in. Now... suppose that the 30% shooter is "feeling it" physically and mentally, and shoots the ball EXACTLY like the guy who shoots at 40%. When he puts the ball up, is there any way you can tell that his shot is 33% more likely than normal to fall? (A 10% increase is 33% of his normal 30%.) Nope. The shot either falls or it doesn't. And at the end of the night, if he goes 4-10 instead of his usual 3-10, could you tell that it's because his stroke, and his confidence, were both better that night? Nope. Guys who shoot 30% have 4-10 nights, just like they have 2-10, or 1-10 nights.

If you've ever played sports at a moderately high level, you know when you see a guy who is in the zone. The bat looks like a toothpick in his hands. When he does miss a putt, it catches the lip and looks like it's going in. His shooting stroke looks smooth as silk, and maybe he's elevating a little more than normal without seeming to try. But even guys who are in the zone get tired - see: Curry, Steph. And even guys with a "hot hand" don't make them all.

Athletes get in the zone. Period, end of story. No mis-applied, non-sequitir math formula can change that. And the pencil-necked, no-sports-playing mathematicians can kiss my ass.

[jag]

I was thinking you can also tell who got a decent education.

But oddly, I somehow fall in both your category and mine.

"Decent" is a relative term. I enjoy and understand statistics, but as an engineer I also understand why statisticians are rarely allowed to make critial decisions in the business world.

[Spur Bank]

It depends on what people mean by hot hand? If they mean the statistics independent of any other factors, then no, it doesn't exist. Thinking so pretty much causes the gambler's ruin scenario.

Every statistical event is independent, so the probability of something occurring again is not related to the what happened in the past. For example, the probability of getting a heads on a fair coin is .5. That's true even if the last 20000 flips before it were heads. If humans were like coins, then believing the hot hand would be ludicrous.

But applying statistical theories at such a shallow level of analysis is fool-hardy. A player is not an automaton who has a fixed probability distribution every time he shoots. His success rate depends on a lot of situations, such as confidence, as mentioned above. Also, things such as the physical state of the player, crowd noise, the temperature of the ball etc. can come into play. If we could quantify all of those factors, we would indeed find certain situations in which a player has a higher probability of making shots. The hot-hand idea is just a non-statistical way to describe that phenomenon.

You say if we could quantify factors such as the physical state of the player, crowd noise, the temperature of the ball, etc., we would indeed find certain situations in which a player has a higher probability of making shots.

Then why don't we? If what you say is true, shots would follow a poisson binomial distribution (similar to a binomial distribution that has changing parameters), and the hot hand theory would be statistically true. Analysts could estimate the p parameter of prior shots and use that to determine future expected outcomes. But we don't see that. The only possible explanation for that is because it doesn't exist.

[Spur Bank]

A lot of people know about statistics - not many understand how to apply it. A lot of people know about sports - not many understand what goes on at a high level. And almost none of the people who try to apply statistics to sports know what the they're talking about. It's that simple.

Anyone who's played compe ive golf knows that there are days when the cup looks as big as a dinner plate. You feel like you can make every putt, and you DO make far more than normal. Baseball players see the ball larger and slower than normal. And basketball players talk about the hoop looking like a washtub. Maybe it's a matter of physical and mental abilities being at their peak at the same time. A lot of people have tried to quantify it by talking about "biorhythms". Nobody knows. But on certain days, the game is just easier than normal. And on those days, more shots go down. No "maybe", no illusion, no fallacy. The shots fall. Call it a "hot hand" or whatever you want.

As for the math? One guy shoots 3's at a 30% clip for an entire season - another shoots them at a 40% rate. What's the difference? Could be any of dozens of things. Most likely several small things. But if you watch a random guy put up a 3-pointer, do you know whether the ball had a 30% chance, or a 40% chance of going in? Nope. You can only tell that by looking at his season statistics. Observing, you can only see that it went in, or didn't go in. Now... suppose that the 30% shooter is "feeling it" physically and mentally, and shoots the ball EXACTLY like the guy who shoots at 40%. When he puts the ball up, is there any way you can tell that his shot is 33% more likely than normal to fall? (A 10% increase is 33% of his normal 30%.) Nope. The shot either falls or it doesn't. And at the end of the night, if he goes 4-10 instead of his usual 3-10, could you tell that it's because his stroke, and his confidence, were both better that night? Nope. Guys who shoot 30% have 4-10 nights, just like they have 2-10, or 1-10 nights.

If you've ever played sports at a moderately high level, you know when you see a guy who is in the zone. The bat looks like a toothpick in his hands. When he does miss a putt, it catches the lip and looks like it's going in. His shooting stroke looks smooth as silk, and maybe he's elevating a little more than normal without seeming to try. But even guys who are in the zone get tired - see: Curry, Steph. And even guys with a "hot hand" don't make them all.

Athletes get in the zone. Period, end of story. No mis-applied, non-sequitir math formula can change that. And the pencil-necked, no-sports-playing mathematicians can kiss my ass.

There are so many things wrong with this post, I don't know where to begin.

You feel like you can make every putt, and you DO make far more than normal. -- Read this: http://en.wikipedia.org/wiki/Clustering_illusion
And on those days, more shots go down. No "maybe", no illusion, no fallacy. The shots fall. -- Wrong word at the end. The shots fell. They won't continue to fall.
But even guys who are in the zone get tired - see: Curry, Steph -- LOLed at you using a counterexample of hot hand as something that supports your argument.
Athletes get in the zone. Period, end of story. -- Yes, they get in zones. It just doesn't affect their ability to make future shots. The math behind it is very simple.

[tesseractive]

tesseractive, my comments are in bold blue

The main issue I have with your post is that you're not arguing the hot hands theory. You're saying there are many factors that can influence a basketball shot... and yes, there are tons. But that's not what we're examining here. We're looking at whether a player on a hot streak is more likely to keep making shots.

You said "none of these effects show up in aggregate statistics." What do you mean by that? That's the big argument on my side, and it's a pretty massive argument because it's the very definition of the question. To put it simply, the argument is: Statistically, a player on a hot streak is NOT more likely to keep making shots. This can be verified by examining the data. Therefore, the hot hands argument is a fallacy.

Everything you said is fine and true, but it doesn't refute that argument at all.

Good response. To put it more precisely:

1. The statistical variable "last shot went in" is not a good representation of what players mean by feeling in the zone, so the fact that it is a poor predictor of the outcome of the next shot doesn't have any particular bearing on whether players can be right when they feel they have a hot hand.

2. There are a great many mental and physical factors that a player's mind can use in recognizing a pattern the player can recognize as "feeling hot" or "feeling cold." Human brains are really good at assimilating lots of data and recognizing a pattern. To develop a good statistical variable to reflect this feeling, you would either need to record comparable data and perform similar pattern recognition, or use EKGs to identify an "in the zone" brain heat map and use that as your variable.

3. I hypothesize that it is possible, given sufficient data collection, to successfully perform (2) and produce a variable that is a statistically significant predictor of shot success and that reflects a player's feeling that he has a hot hand.

[Chinook]

You say if we could quantify factors such as the physical state of the player, crowd noise, the temperature of the ball, etc., we would indeed find certain situations in which a player has a higher probability of making shots.

Then why don't we? If what you say is true, shots would follow a poisson binomial distribution (similar to a binomial distribution that has changing parameters), and the hot hand theory would be statistically true. Analysts could estimate the p parameter of prior shots and use that to determine future expected outcomes. But we don't see that. The only possible explanation for that is because it doesn't exist.

The problem comes from the fact that analysts don't know HOW to quantify those factors. If you made a basketball simulation, then you could set those equations and make a pretty realistic game. But they can't get it exact, because there're too many details that they don't know. We didn't make our universe, so we didn't make up the math behind it. That doesn't make those factors impossible to ever know, though, or at least to approximate well enough for our tastes.

[Spur Bank]

The problem comes from the fact that analysts don't know HOW to quantify those factors. If you made a basketball simulation, then you could set those equations and make a pretty realistic game. But they can't get it exact, because there're too many details that they don't know. We didn't make our universe, so we didn't make up the math behind it. That doesn't make those factors impossible to ever know, though, or at least to approximate well enough for our tastes.

Good response. To put it more precisely:

1. The statistical variable "last shot went in" is not a good representation of what players mean by feeling in the zone, so the fact that it is a poor predictor of the outcome of the next shot doesn't have any particular bearing on whether players can be right when they feel they have a hot hand.

2. There are a great many mental and physical factors that a player's mind can use in recognizing a pattern the player can recognize as "feeling hot" or "feeling cold." Human brains are really good at assimilating lots of data and recognizing a pattern. To develop a good statistical variable to reflect this feeling, you would either need to record comparable data and perform similar pattern recognition, or use EKGs to identify an "in the zone" brain heat map and use that as your variable.

3. I hypothesize that it is possible, given sufficient data collection, to successfully perform (2) and produce a variable that is a statistically significant predictor of shot success and that reflects a player's feeling that he has a hot hand.

Thanks for the clarifications.

I'm on board with these ideas. Point #3 above interesting, and I wouldn't be at all surprised if you were correct. To rephrase what you said ... just because a player has made a bunch of shots in a row doesn't mean he has a hot hand or is in the zone, hence, why that statistical correlation doesn't exist; but if we could better measure what truly should be considered in the zone, then it likely would. Very interesting idea, it certainly makes sense.

[ALVAREZ6]

Klay Thompson is ing hot right now, shooting 7/8 from 3 and hitting a wide variety of different types of 3s...spot up, off the dribble, slightly fading, you name it. He is ing hot and no one can refute this.

[hater]

at this thread and it's timing

where is longtimespurfan when you need him

[phyzik]

I think it would be really interesting to develop a compact brain mapping system that could fit, in say, an NFL players helmet... Set this up in a lot of players helmets across the league and check their brain activity to cross reference with actual performance. unfortunately, I dont think there is yet a way to miniaturize it to that scale just yet where it wouldnt hamper a players performance.

This type of study COULD work however for compe ive sports where the individual doesnt have to move around much, such as my old sport, compe ive shooting. Check the brain activity vs actual performance across a given period of time against a large number of individual compe ors. It would be interesting to see if there was any correlation to certain brain activity when being in "the zone" is perceived by the individual compe or. Another easy way to do this, while not a physical sport, would be to hook up brain monitoring systems to "professional" PC/XBOX/PS3 players vs your average gamer. It may not be physical but the mental part still plays a factor and might even generate an even better result as you take the physicality our of the equation and only focus on the mental side of it.