Thanking Ken B. for his Willful Ignorance...
recommendation.
Ken recommended the book Willful Ignorance to me. It arrived today; I randomly* opened it up and found a section on "The Ignorance Fallacy." In the section, the author, Herbert Weisberg, discusses the "hot hand fallacy." After a quick review of the evidence, he writes:
Let us trust these athletes self-reports for a moment, but further posit that such periods of heightened awareness appear and disappear in an unpredictable fashion. Then athletes' reports of having a "hot hand" would be entirely accurate, but also be compatible with the analysis showing that the data matches a random process.
By the way, a great example of something that appears random but is actually deliberately caused is the random number generator in your favorite programming language. It took careful design on the part of many engineers to make seemingly random numbers pop out of your computer!
* Or was I guided? Perhaps by the "hot hand" of... Satan?
Ken recommended the book Willful Ignorance to me. It arrived today; I randomly* opened it up and found a section on "The Ignorance Fallacy." In the section, the author, Herbert Weisberg, discusses the "hot hand fallacy." After a quick review of the evidence, he writes:
So, it appears that streakiness is just a myth. Or is it?And this is precisely what I have pointed out a number of times in the past: the findings "debunking" hot hands are all entirely consistent with the actual existence of hot hands. For instance, athletes who are "hot" often report that during their streak, they felt a sense of heightened awareness, and say things like "the baseball appeared as big as a grapefruit to me."
Let us accept for the moment the hypothesis that pure randomness can plausibly explain almost any hot hand streak in sports or games. Does that necessarily imply that such streaks do not really exist? Consider that there are a great many factors, most not measurable, that might influence any individual outcome, such as one particular game or at bat... What the research certainly tells us is that if such factors exist, they must be haphazard enough to appear essentially random.
Let us trust these athletes self-reports for a moment, but further posit that such periods of heightened awareness appear and disappear in an unpredictable fashion. Then athletes' reports of having a "hot hand" would be entirely accurate, but also be compatible with the analysis showing that the data matches a random process.
By the way, a great example of something that appears random but is actually deliberately caused is the random number generator in your favorite programming language. It took careful design on the part of many engineers to make seemingly random numbers pop out of your computer!
* Or was I guided? Perhaps by the "hot hand" of... Satan?
Always trust Ken B.
ReplyDelete'ceptin with the womenfolk.
The original hot hands discussion centered on the claim that basketball players are more likely to make a successful shot if their previous shot was successful. Is that how "hot hands" is being defined here ?
ReplyDeleteIf so, and if one assumes that periods of heightened awareness help a player make a shot, but that such periods appear and disappear in an unpredictable fashion then I don't see how they are at all relevant to the discussion since the presence of heightened awareness will be a random factor equal for any shot and won't increase the probability of a second successive shot being made.
If instead one assumes that the periods of heightened awareness is likely to last more than one shot then one is assuming that hot hands (as defined in the first paragraph) exists. If so, then as long as one could find a way of accurately measuring it, wouldn't the statistics show that the chances of a second successive basket was greater than the chances of the first ?
"The original hot hands discussion centered on the claim that basketball players are more likely to make a successful shot if their previous shot was successful. Is that how "hot hands" is being defined here ?"
DeleteNope. In fact, what is being argued is that is a *bad* definition.
"then I don't see how they are at all relevant"
Of course you don't rob. There is apparently no post on any subject you are not capable of misunderstanding.
Gene,
DeleteI just noticed that in your initial quote from the book you replace a clause that indicates (to me anyway) that authors' views are somewahtr different from yours with "...". Would you agree that this omitted clause changes the author's meaning away from what you had in mind ?
"Would you agree that this omitted clause changes the author's meaning away from what you had in mind ?"
DeleteNO!!!!! And you are accusing me of lying here, you f'ing asswipe!!! Bye!!!!
Gene,
DeleteI think you're overreacting a bit here. What did I say to imply I thought you were lying ?
The "..." can be seen in below link.
Its google books and I can't cut and paste but is something to the effect of people who were able to understand the apparently random factors might be able to profit from them. This is not the case in your models, right? Hence my question. I don't honestly see why this is an unreasonable question to ask since it the premise for your post !
While I have to hold my hands up to occasionally posting comments that I think may get a rise from you (and perhaps this was such a case) I don't think I ever purposely step over any lines. Perhaps there is an etiquette for these kinds of sensitive matters that I inadvertently broke - if that is the case then I apologize.
The link is:
https://books.google.com/books?id=L-JvBAAAQBAJ&pg=PT59&lpg=PT59&dq=So,+it+appears+that+streakiness+is+just+a+myth.+Or+is+it?Let+us+accept+for+the+moment+the+hypothesis+that+pure+randomness+can+plausibly+explain+almost+any+hot+hand+streak+in+sports+or+games.+Does+that+necessarily+imply+that+such+streaks+do+not+really+exist?+Consider+that+there+are+a+great+many+factors,+most+not+measurable,+that+might+influence+any+individual+outcome,+such+as+one+particular+game+or+at+bat...+What+the+research+certainly+tells+us+is+that+if+such+factors+exist,+they+must+be+haphazard+enough+to+appear+essentially+random.&source=bl&ots=YE4S6HEOtK&sig=zxNartWpUUxZ7A0mWDNmJRZHa2I&hl=en&sa=X&ved=0ahUKEwjJppupiP_UAhVVImMKHaEaDkUQ6AEIKDAA#v=onepage&q=So%2C%20it%20appears%20that%20streakiness%20is%20just%20a%20myth.%20Or%20is%20it%3FLet%20us%20accept%20for%20the%20moment%20the%20hypothesis%20that%20pure%20randomness%20can%20plausibly%20explain%20almost%20any%20hot%20hand%20streak%20in%20sports%20or%20games.%20Does%20that%20necessarily%20imply%20that%20such%20streaks%20do%20not%20really%20exist%3F%20Consider%20that%20there%20are%20a%20great%20many%20factors%2C%20most%20not%20measurable%2C%20that%20might%20influence%20any%20individual%20outcome%2C%20such%20as%20one%20particular%20game%20or%20at%20bat...%20What%20the%20research%20certainly%20tells%20us%20is%20that%20if%20such%20factors%20exist%2C%20they%20must%20be%20haphazard%20enough%20to%20appear%20essentially%20random.&f=false
The "..." = "Like George a person with special insight or inside information might be able to profit handsomely by acting on it".
Delete"Its google books and I can't cut and paste but is something to the effect of people who were able to understand the apparently random factors might be able to profit from them. This is not the case in your models, right? "
DeleteWRONG!!!!!! Anyone who understood when the hothand setting was on could profit enormously betting against those who didn't.
I cut out that snippet BECAUSE THE QUOTE WAS LONG, and because I would have to also quote the previous two pages EXPLAINING WHO GEORGE IS.
In fact, if you read those two pages, you'll see the author presenting a model VERY MUCH LIKE MINE.
And yes, rob, if I HAD deliberately snipped out a part contradicting my post, that WOULD be lying! So you were accusing me of lying.
But it is to much fun knocking apart the rubbish you post to ban you!!!!
OK, I see now how my comment might justifiably have upset you - so I apologize again.
DeleteOK, apology accepted.
DeleteIt's been recently shown/argued that the basic 'hot hand is a fallacy' argument is not correct anyway - see e.g. http://andrewgelman.com/2015/07/09/hey-guess-what-there-really-is-a-hot-hand/
ReplyDeleteInteresting: thanks! This is an important point, but orthogonal to what I am saying.
DeleteGene, would you say that hot hands in gambling are a myth? Like in slot machines, for instance.
ReplyDeleteIn a game of pure chance, like roulette or a slot machine? Of course! The machine is *designed* to give random results, and no skill is involved, so of course one cannot get "hot" in such a situation.
DeleteSlot machines? Keshav could reverse engineer the randomizer, determine that there are dependencies, wait for his chance, and take advantage. This is not possible in practical terms but in theory a hot hand is possible for a slots player.
DeleteWell, to my mind there's no "of course" about it. Or do you think that notions of luck, fate, etc. are all self-evidently absurd?
DeleteKen B: Well, I meant under ordinary conditions.
Delete"Or do you think that notions of luck, fate, etc. are all self-evidently absurd?"
DeleteIf God is intervening, that is beyond science and math. "No humanly verifiable hot streaks."
I am pleased to note you are boyhood wrong. The hot hand is the claim that in a purportedly random sequence of 1 and 0 there are more contiguous subsequences of just 1 than is plausibly attributed to the hypothesis the events are independent.
ReplyDeleteThis comment has been removed by the author.
Delete"The hot hand is the claim that in a purportedly random sequence of 1 and 0 there are more contiguous subsequences of just 1 than is plausibly attributed to the hypothesis the events are independent. "
DeleteThat is the BAD DEFINITION used by TGV. I do not need to use their BAD DEFINITION along with them!!!!
"Or was I guided? perhaps by the 'hot hand' of... Satan?"
ReplyDeletei believe God exists. I do not believe Satan exists.