MBB: Michigan Intercollegiate Athletic Association

[KnightSlappy]

Just to play devil's advocate to this argument today, if a shooter is a 40% shooter and makes 5 straight threes, the shooting percentage would also go up.  So, now the shooter is a 43% shooter (let's say, obviously dependent on the total number of shots) so the probability of making the next shot would be slightly higher (43% vs. 40%).  The argument then becomes whether the shooter is truly a 40% shooter or a 43% shooter?

I think the phrase "60% of all statistics are made up" might apply to my argument above but I tend to believe that hot streaks are random sampling (contrary to my argument).  Just because you flip a coin and it comes up heads three times in a row doesn't increase your chances of the next flip being heads.

If only shooting a basketball was as simple as flipping a coin....

Don't think you can really compare a coin flip streak of multiple heads in a row to a basketball shooting streak of multiple makes in a row.

A coin flip is non-emotional event, while making or missing a basketball shot creates an ebb and flow of emotions that also factor in to the next shot taken.

It fits in that the statistical analysis determines that a shooter is no more or less likely to make a shot based on the outcome of the previous shot just as a coin is no more or less likely to come up on one side based on what side came up the time before. If you want to throw out what the studies have concluded then that is fine but that sort of ends the discussion.

I would agree with that if it was robots playing all the games.

Are you saying then that game situation and nerves play no part whatsoever in the likelihood of the outcome in an emotional event (as basketball games often are).  Because I've seen a lot more good free throw shooters miss one at the end of the game when they hadn't taken any FT's through the course of the game than when they've already taken a few and are in a good rhythm.

In something such as sports you can't apply ONLY statistics, if that was the case then Derek Jeter wouldn't be thought of as highly as he is.  ARod has better overall statistics for the Yankees but with the game on the line and just needing a base hit in the bottom of the 9th most people would rather Derek Jeter at the plate than Arod regardless of what the overall statistics say for a 9 inning game.

The reason prior results don't play a part in coin flips is because a coin is an inanimate, non-breathing, non-emotional subject that has no memory.  The same can't be said for a human.

We should be careful not to confuse end game scenarios with hot streaks. It's probably an entirely different scenario.

Tom Tango (insidethebook.com) has done projects the couple of years on the 'clutchness' of hitters. Fans from each team voted to determine who they wanted at the plate in end game scenarios, and Tom simply took the best hitter on the team. Tom's players did better in high leverage situations than the players who were perceived to be better 'clutch' hitters.

http://www.fangraphs.com/clutch.aspx?type=0

[realist]

FDF:  Unlike the other politicans I will answer the question you asked.   I go with player A.  My reasoning is most coaches also know what the %'s are, and they are going to guard the guy "with the hot hand that night (aka player B)"  Player A is likely to have just the extra split second to prepare for a more open shot.

[Happy Calvin Guy]

This debate comes down to one issue:  what holds predictive power?  No can dispute that sometimes players make 4 shots in a row or miss 4 shots in a row.  Obviously that happens sometimes.  The question is then, given one scenario or the other, does that lend any predictive power to what they are going to do on their next shot, which is really all that matters in decision making.  

As humans, we are subject to confirmation bias.  i.e. if we think that a player has a hot hand and is going to make the next shot, and then the player actually makes it, we say "see, he has the hot hand!". We have confirmed our belief. If he misses, then we usually quickly forget that it happened and don't use that as an opportunity to challenge or dispute our own beliefs about the existence of hot hands.  This is just human nature and there are plenty of studies on this.  I'm certainly not a psychologist but have done my share of reading on the issue of randomness, bias, and human nature.

By the way, it can be more fun to believe in a player who has the hot hand, so if you choose to cling to that belief, then don't let this debate ruin any of your fun in being an MIAA basketball fan.  It's not worth it!

[ziggy]

This debate comes down to one issue:  what holds predictive power?  No can dispute that sometimes players make 4 shots in a row or miss 4 shots in a row.  Obviously that happens sometimes.  The question is then, given one scenario or the other, does that lend any predictive power to what they are going to do on their next shot, which is really all that matters in decision making.  

As humans, we are subject to confirmation bias.  i.e. if we think that a player has a hot hand and is going to make the next shot, and then the player actually makes it, we say "see, he has the hot hand!". We have confirmed our belief. If he misses, then we usually quickly forget that it happened and don't use that as an opportunity to challenge or dispute our own beliefs about the existence of hot hands.  This is just human nature and there are plenty of studies on this.  I'm certainly not a psychologist but have done my share of reading on the issue of randomness, bias, and human nature.

By the way, it can be more fun to believe in a player who has the hot hand, so if you choose to cling to that belief, then don't let this debate ruin any of your fun in being an MIAA basketball fan.  It's not worth it!

^This.

[realist]

"What causes the "hot hand" feeling?
Many athletes and other people believe in a "hot hand" phenomenon: that a basketball player who sinks several baskets in a row is "hot" and is more likely than average to sink his next basket as well. But based on statistical analysis this doesn't seem to be true.

Here's a possible reason that people might feel as if it were true. When you're learning to perform a task for the first time, a "hot hand" type of belief is probably both factually correct (you actually are more likely to get it right if you just got it right a minute ago) and also adaptive (do it right --> more confidence --> do it the same way again, do it wrong --> less confidence --> do something different.)

The basketball players under discussion have skills that are more or less mature; they're not going to get measurably better at shooting over the course of a game. But maybe part of the brain doesn't "know" that. From the point of view of this learning mechanism in the brain, maybe the fact that you just sunk a few baskets indicates that you've learned something new about shooting, so it's time to positively reinforce that learning with a flush of confidence. As long as this "hot hand" feeling isn't actually harmful, there's not much reason for it to distinguish between developing skills and mature skills, if such a distinction is even possible. "



Richard Mason



--------------------------------------------------------------------------------

[wiz]

Here's a testimony that you simply put the ball in the best player's hands.
http://www.youtube.com/watch?v=q6CjOOXmkJ8

[KnightSlappy]

Here's a testimony that you simply put the ball in the best player's hands.
http://www.youtube.com/watch?v=q6CjOOXmkJ8

But sometimes that same strategy doesn't work.

http://www.youtube.com/watch?v=k3mJvgIMycc&feature=related

[Gregory Sager]

The problem with this analysis is that you guys are all looking at the shooter in isolation. Basketball is a team game. Unless you're talking about a point guard who is simply going to bring the ball up the floor and then take a shot, every three-pointer attempted comes within the context of an offensive set. Another aspect of the game to consider that I haven't seen addressed much in this conversation is that not all trey attempts are alike. A 21-footer is easier to make than a 25-footer; an open shot is easier to make than a contested shot; and a spot-up stationary shot is (usually) easier to make than a shot on the move. These three factors alone will affect the percentages of Shooter A and Shooter B in that end-game scenario.

Things to consider:

* Is the shooter in question moving well without the ball, and thus doing a good job of helping to create an eventual open shot?

* Does the team have a playmaker or playmakers who can improve the shooter's odds of getting an open look by orchestrating rapid perimeter ball movement or by breaking down the defense with dribble penetration prior to a kickout?

* Where is the shooter's favorite shooting spot on the floor? Some shooters are better from the deep corners; some are better from the elbows; yet others prefer the top of the key. This leads to the secondary question: Is it possible to run a play that puts the shooter in his ideal spot on the floor that won't be identified and checked by the opposing coach?

* How good are the shooter's other teammates at aiding his shooting? Do they set good screens for him? Do they set any screens for him? Is/are the post player(s) adept at finding and hitting the open man on the perimeter when a double-team comes at the post?

* What are the ancillary skills of the shooter? Does he have the handles to penetrate? Does he have the quickness and the hops to penetrate? Is he good at scraping off of screens? Does he at least have a good jab step? Obviously, if he has one or more of these abilities, then his man will have to guard him more honestly by playing off of him a bit, which opens up more shooting room for the shooter.

* What is the defensive capability of the player assigned to guard the shooter? Does he drift away off the ball, or stick like glue to his man? Does he fight through screens, or is he quick enough to go around them without losing his man? Does he have the height and reach to interfere with a shot, either by blocking it or by impeding the shooter's vision, if he's close enough? Is he quick enough to recover and close? Does he have a big man who will come out and hedge at the top of the key to disrupt ball movement, or back-liners who will converge and play help defense for him if he commits to guarding the shot rather than the drive by playing right in the shooter's face? Does he have teammates who are verbal and alert enough to call out screens that are being set in his path?

* What about the coaches? Does the shooter's coach recognize his ability as a shooter and design plays around him, or does he not consider him to be a featured shooter? Does the opposing coach recognize his ability and assign a good defender to guard him, or does he put one of his poorer defenders on him? Anybody who's ever sat in a gym and watched basketball for any length of time is aware that opposing coaches know full well which outside shooters can hurt them, because an opposing coach yelling "Shooter!" when a particular player gets the ball in his hands on the perimeter is one of the classic sounds of the game.

Sorry, guys, but it's not all just a tidy set of mathematical formulations. Nor is it even strictly a matter of emotion being the random element in the game. The game itself, in all of its permutations, affects how well a shooter is going to shoot in any given situation.

[KnightSlappy]

Venerable Mr. Sager, the primary discussion was centered around streaks and not end-game scenarios (although the topic did come up).

In the vein of "streaks" how many of your points to consider would change from shot to shot, and how would they apply differently if the previous shot was made or missed?

End game strategy does indeed contains many of the variables you listed plus quite a bit of game theory.

[Happy Calvin Guy]

Here's a testimony that you simply put the ball in the best player's hands.
http://www.youtube.com/watch?v=q6CjOOXmkJ8

But sometimes that same strategy doesn't work.

http://www.youtube.com/watch?v=k3mJvgIMycc&feature=related

Sometimes a good analysis of the empirical evidence is also the most enjoyable. 

[Knightmare]

The problem with this analysis is that you guys are all looking at the shooter in isolation. Basketball is a team game. Unless you're talking about a point guard who is simply going to bring the ball up the floor and then take a shot, every three-pointer attempted comes within the context of an offensive set. Another aspect of the game to consider that I haven't seen addressed much in this conversation is that not all trey attempts are alike. A 21-footer is easier to make than a 25-footer; an open shot is easier to make than a contested shot; and a spot-up stationary shot is (usually) easier to make than a shot on the move. These three factors alone will affect the percentages of Shooter A and Shooter B in that end-game scenario.

Things to consider:

* Is the shooter in question moving well without the ball, and thus doing a good job of helping to create an eventual open shot?

* Does the team have a playmaker or playmakers who can improve the shooter's odds of getting an open look by orchestrating rapid perimeter ball movement or by breaking down the defense with dribble penetration prior to a kickout?

* Where is the shooter's favorite shooting spot on the floor? Some shooters are better from the deep corners; some are better from the elbows; yet others prefer the top of the key. This leads to the secondary question: Is it possible to run a play that puts the shooter in his ideal spot on the floor that won't be identified and checked by the opposing coach?

* How good are the shooter's other teammates at aiding his shooting? Do they set good screens for him? Do they set any screens for him? Is/are the post player(s) adept at finding and hitting the open man on the perimeter when a double-team comes at the post?

* What are the ancillary skills of the shooter? Does he have the handles to penetrate? Does he have the quickness and the hops to penetrate? Is he good at scraping off of screens? Does he at least have a good jab step? Obviously, if he has one or more of these abilities, then his man will have to guard him more honestly by playing off of him a bit, which opens up more shooting room for the shooter.

* What is the defensive capability of the player assigned to guard the shooter? Does he drift away off the ball, or stick like glue to his man? Does he fight through screens, or is he quick enough to go around them without losing his man? Does he have the height and reach to interfere with a shot, either by blocking it or by impeding the shooter's vision, if he's close enough? Is he quick enough to recover and close? Does he have a big man who will come out and hedge at the top of the key to disrupt ball movement, or back-liners who will converge and play help defense for him if he commits to guarding the shot rather than the drive by playing right in the shooter's face? Does he have teammates who are verbal and alert enough to call out screens that are being set in his path?

* What about the coaches? Does the shooter's coach recognize his ability as a shooter and design plays around him, or does he not consider him to be a featured shooter? Does the opposing coach recognize his ability and assign a good defender to guard him, or does he put one of his poorer defenders on him? Anybody who's ever sat in a gym and watched basketball for any length of time is aware that opposing coaches know full well which outside shooters can hurt them, because an opposing coach yelling "Shooter!" when a particular player gets the ball in his hands on the perimeter is one of the classic sounds of the game.

Sorry, guys, but it's not all just a tidy set of mathematical formulations. Nor is it even strictly a matter of emotion being the random element in the game. The game itself, in all of its permutations, affects how well a shooter is going to shoot in any given situation.

What I was trying to get at but much better stated.  Also KnightSlappy, you are correct to make the distinction between end-game scenarios and "streaks" BUT even with "streaks" all other variables come into play.  If basketball, or any other sport for that matter, could be broken down into strictly statistical analysis then there would be no point in playing the games because we could just plug everything into a computer and come up with the result instead.  Unforeseen or unpredicted play and results is why they play the game otherwise we could just use statistical analysis to determine the winner.

[ziggy]

What I was trying to get at but much better stated.  Also KnightSlappy, you are correct to make the distinction between end-game scenarios and "streaks" BUT even with "streaks" all other variables come into play.  If basketball, or any other sport for that matter, could be broken down into strictly statistical analysis then there would be no point in playing the games because we could just plug everything into a computer and come up with the result instead.  Unforeseen or unpredicted play and results is why they play the game otherwise we could just use statistical analysis to determine the winner.

That is a gross misapplication of the implications of the statistical analysis presented over the past couple pages.

Strategy was never discarded in favor of a strict statistical method as part of the discussion. The only "strategy" that was discredited was the riding of a hot hand, something statistical analysis shows does not exist. Game situation obviously plays a role but that is all rolled up into the numbers and it is understood that there are a lot of things that can be done to positively or negatively impact the likelihood of a particular result. That is far different than the wild swings in probability based on simply being "hot" or "cold".

[Gregory Sager]

Venerable Mr. Sager, the primary discussion was centered around streaks and not end-game scenarios (although the topic did come up).

... and the fact that the topic came up is the reason why I addressed it.

In the vein of "streaks" how many of your points to consider would change from shot to shot, and how would they apply differently if the previous shot was made or missed?

Depends. If a player hits three treys in a row, the opposing coach might switch defenders or sub in a new one in order to counteract what he perceives to be the "hot hand," even if the shooter is typically a mediocre or poor shooter. If a normally good shooter isn't hitting his shots, his coach may decide to augment his chances by having his players specifically set screens for the shooter, or by having the shooter run baseline in an attempt to shuck his man, or any number of other options designed to free up a shooter.

As to whether or not a previous make or miss could affect a shot ... well, if a player misses a wide-open trey, his man might slough off of him just a tad the next time down the floor, thinking to himself, "This guy's no threat from behind the arc." Or, conversely, if he makes a 25-footer, his man may think to himself, "Geez, I gotta sit right on top of this guy from halfcourt on, or he's gonna burn me again," and get right up in the shooter's grill. Basketball is a game of constant adjustments, and one play could very easily lead to an adjustment by either shooter or defender, or by their teammates and coaches, that could tilt the odds from one shot to the next.

[ziggy]

Depends. If a player hits three treys in a row, the opposing coach might switch defenders or sub in a new one in order to counteract what he perceives to be the "hot hand," even if the shooter is typically a mediocre or poor shooter. If a normally good shooter isn't hitting his shots, his coach may decide to augment his chances by having his players specifically set screens for the shooter, or by having the shooter run baseline in an attempt to shuck his man, or any number of other options designed to free up a shooter.

As to whether or not a previous make or miss could affect a shot ... well, if a player misses a wide-open trey, his man might slough off of him just a tad the next time down the floor, thinking to himself, "This guy's no threat from behind the arc." Or, conversely, if he makes a 25-footer, his man may think to himself, "Geez, I gotta sit right on top of this guy from halfcourt on, or he's gonna burn me again," and get right up in the shooter's grill. Basketball is a game of constant adjustments, and one play could very easily lead to an adjustment by either shooter or defender, or by their teammates and coaches, that could tilt the odds from one shot to the next.

You prove the point. The success of the next attempt has nothing to do with simply being "hot" or "cold".

[Gregory Sager]

Depends. If a player hits three treys in a row, the opposing coach might switch defenders or sub in a new one in order to counteract what he perceives to be the "hot hand," even if the shooter is typically a mediocre or poor shooter. If a normally good shooter isn't hitting his shots, his coach may decide to augment his chances by having his players specifically set screens for the shooter, or by having the shooter run baseline in an attempt to shuck his man, or any number of other options designed to free up a shooter.

As to whether or not a previous make or miss could affect a shot ... well, if a player misses a wide-open trey, his man might slough off of him just a tad the next time down the floor, thinking to himself, "This guy's no threat from behind the arc." Or, conversely, if he makes a 25-footer, his man may think to himself, "Geez, I gotta sit right on top of this guy from halfcourt on, or he's gonna burn me again," and get right up in the shooter's grill. Basketball is a game of constant adjustments, and one play could very easily lead to an adjustment by either shooter or defender, or by their teammates and coaches, that could tilt the odds from one shot to the next.

You prove the point. The success of the next attempt has nothing to do with simply being "hot" or "cold".

I wasn't disputing it in the first place. My point is that you can't reduce this discussion to statisical analysis, either pro or con the idea of a shooter being either hot or cold (or lukewarm, for that matter). Yes, statistics do give a general basis of probability that will affect the outcome of any particular play -- layups are more likely to go in than trey attempts, Veltema is more likely to make a trey than would Mantel under identical circumstances, et cetera, et cetera -- but statistics do not exist in a vacuum. Game play, in all of its permutations, shapes and distorts how statistics work. And, in the case of hot and cold, they also can determine how coaches coach and defenders defend, a la confirmation bias.