In light of the recent paper about the existence of the Hot Hand submitted to the MIT Sloan Sports Analytics Conference, I wrote a piece that I thought this community would enjoy about the overvaluation of the hot hand and its negative effect on player performance. Whether you agree or disagree with my points, I would love to hear your thoughts and I look forward to contributing to the conversation on this site.
http://www.dleaguemvp.com/2014/03/the-o ... -hot-hand/
The Overvaluation of the "Hot Hand"
Re: The Overvaluation of the "Hot Hand"
Can you accept that a player may 'go cold' -- that is, he may lose all confidence in his shot, and subsequently miss most or all of them?
The effect can be quite dramatic. Hitting essentially none of your shots will likely get you a seat on the bench.
"Hot" exists, at minimum, as the opposite of "cold".
You can get hot or cold in any activity. A guy can suddenly find himself able to dribble the ball through a crowd; or unable to control his hands and feet. A passer can get in synch with teammates (and ahead of the defense) and thread passes he normally wouldn't.
Golfers get hot and cold. Tennis players. Skiers. Nothing to do with teammates or defenses.
Anyone can then get over-confident. Race car drivers, boxers, anyone. Results may be devastating; but that's a separate phenomenon.
The effect can be quite dramatic. Hitting essentially none of your shots will likely get you a seat on the bench.
"Hot" exists, at minimum, as the opposite of "cold".
You can get hot or cold in any activity. A guy can suddenly find himself able to dribble the ball through a crowd; or unable to control his hands and feet. A passer can get in synch with teammates (and ahead of the defense) and thread passes he normally wouldn't.
Golfers get hot and cold. Tennis players. Skiers. Nothing to do with teammates or defenses.
Anyone can then get over-confident. Race car drivers, boxers, anyone. Results may be devastating; but that's a separate phenomenon.
Re: The Overvaluation of the "Hot Hand"
I think the kindest thing that can be said about the paper in question is that it isn't a basketball-centric paper.
Taking the results at face value (and I do have some questions about the various assumptions involved in generating the results, not doubts per se, but questions), and someone please check me on this, what they show is that the "hot-hand" effect is real but also that it is shown to generate net negative, not positive, returns. That is the quality of subsequent shots of the emboldened "hot" shooters are of lower quality, such that, on net, these more than offset the estimated positive returns to "hotness".
Now, never mind that this fact (and, again, someone please check me on this) wasn't emphasized, neither by the authors (actually, to the contrary, where the local wisdom of the players etc. was weirdly celebrated) nor by the ESPN plumpers (journalistic tanking, anyone?) And never mind the legerdemain (from a basketball perspective) of emphasizing results in terms of the percent change in efficiency so as to get a bigger number, rather than the conventional percentage point improvement in shooting efficiency. The question I have is (again, accepting the results at face value): why was an estimate not provided in the paper as to the total, per game, hot-hand effect?
I suspect that it is because this effect would be utterly trivial (again, if not negative). After all, this is an effect which only arises after multi-possession, above-average result runs. And how many of these are expected per game? (And might there be offsetting "cold hand" effects?)
Multiply "not many" times "a very small effect" (perhaps negative) and one is returned to the position where all right-thinking people were before this paper was reported: any hot-hand effect should be presumed to be small and irrelevant to the game of basketball. And to celebrate such a result as a redemption of traditional wisdom is simply...well, everyone can finish this sentence for themselves.
Taking the results at face value (and I do have some questions about the various assumptions involved in generating the results, not doubts per se, but questions), and someone please check me on this, what they show is that the "hot-hand" effect is real but also that it is shown to generate net negative, not positive, returns. That is the quality of subsequent shots of the emboldened "hot" shooters are of lower quality, such that, on net, these more than offset the estimated positive returns to "hotness".
Now, never mind that this fact (and, again, someone please check me on this) wasn't emphasized, neither by the authors (actually, to the contrary, where the local wisdom of the players etc. was weirdly celebrated) nor by the ESPN plumpers (journalistic tanking, anyone?) And never mind the legerdemain (from a basketball perspective) of emphasizing results in terms of the percent change in efficiency so as to get a bigger number, rather than the conventional percentage point improvement in shooting efficiency. The question I have is (again, accepting the results at face value): why was an estimate not provided in the paper as to the total, per game, hot-hand effect?
I suspect that it is because this effect would be utterly trivial (again, if not negative). After all, this is an effect which only arises after multi-possession, above-average result runs. And how many of these are expected per game? (And might there be offsetting "cold hand" effects?)
Multiply "not many" times "a very small effect" (perhaps negative) and one is returned to the position where all right-thinking people were before this paper was reported: any hot-hand effect should be presumed to be small and irrelevant to the game of basketball. And to celebrate such a result as a redemption of traditional wisdom is simply...well, everyone can finish this sentence for themselves.
Re: The Overvaluation of the "Hot Hand"
" Heat Check: New Evidence on the Hot Hand in Basketball" was published recently at SSRN.
Abstract: The vast literature on the hot hand fallacy in basketball rests on the assumption that shot selection is independent of player-perceived hot or coldness. In this paper, we challenge the assumption of independence using a novel dataset of over 83,000 shots from the 2012-2013 NBA season, combined with optical tracking data of both the players and the ball. We use this data to show that players who have exceeded their expected shooting percentage over recent shots shoot from significantly further away, face tighter defense, are more likely to take their team’s next shot, and overall attempt more difficult shots. In other words, we show that the independence assumption fails. We then turn to the hot hand itself and show that players who are outperforming (i.e. are “hot”) are more likely to make their next shot if we control for the difficulty of that shot. We estimate a 1.2% increase in the likelihood of the typical player making his next shot for each additional prior shot he made.
It can be downloaded here: http://papers.ssrn.com/sol3/papers.cfm? ... id=2481494
Not sure if I should apologize for not having a comment on it but I welcome comments from others.
Abstract: The vast literature on the hot hand fallacy in basketball rests on the assumption that shot selection is independent of player-perceived hot or coldness. In this paper, we challenge the assumption of independence using a novel dataset of over 83,000 shots from the 2012-2013 NBA season, combined with optical tracking data of both the players and the ball. We use this data to show that players who have exceeded their expected shooting percentage over recent shots shoot from significantly further away, face tighter defense, are more likely to take their team’s next shot, and overall attempt more difficult shots. In other words, we show that the independence assumption fails. We then turn to the hot hand itself and show that players who are outperforming (i.e. are “hot”) are more likely to make their next shot if we control for the difficulty of that shot. We estimate a 1.2% increase in the likelihood of the typical player making his next shot for each additional prior shot he made.
It can be downloaded here: http://papers.ssrn.com/sol3/papers.cfm? ... id=2481494
Not sure if I should apologize for not having a comment on it but I welcome comments from others.
Re: The Overvaluation of the "Hot Hand"
That sounds really familiar. Was this data already presented/published somewhere?
If this is the research I'm thinking of, and I remember someone else's comment about it, it's that the practical value of this finding is that getting hot hurts a team. The increased difficulty of later shots wipes out any benefit of 'hotness', and overall the player isn't shooting any better and could be shooting worse.
If this is the research I'm thinking of, and I remember someone else's comment about it, it's that the practical value of this finding is that getting hot hurts a team. The increased difficulty of later shots wipes out any benefit of 'hotness', and overall the player isn't shooting any better and could be shooting worse.