Back to Drew's Views
September 7, 2020
Previous
Next

The Hot-Hand Fallacy Fallacy Fallacy?

Factors
Sports
Stock Picking

Many of you will be familiar with the so-called “hot-hand” fallacy, or (perhaps) the lack thereof.  In 1985, Tom Gilovich (along with Robert Vallone and Amos Tversky) published a paper suggesting that there is no such thing as a hot hand, but rather a “general misconception of chance…and random sequences”. [1] 

The analysis was originally focused on basketball, and as anyone who ever played hoops will attest, this was a hard one to swallow.  I’ve sure at least felt like I had a hot hand, and I’ve damn sure felt like I’ve had a cold hand.  And if a cold hand can exist, well, that means a hot one can too.

Yet for the ensuing two decades, there was much additional work corroborating Gilovich’s results.  It became intellectual heresy to claim that a hot hand existed in hoops (unless you were a New York Knick fan and talking about Reggie Miller, in which case you got a pass).  

I disagree.  I think.

In the last few years, however, there have been two papers that suggested that the Gilovich’s results may not have been as robust as everyone thought. The first one[2] pointed out that “hot” players start feeling so hot that they start taking shots from further away.  Moreover, opposing teams will also adjust their defense to focus more primarily on the hot shooter.  Using some pretty cool tracking data, they figured that the streaky shooter’s actual apples-to-apples FG percentage goes up 1-2% when he or she is “hot”.  So this all makes sense and is kind of interesting, but a 45% shooter becoming a 46% or 47% shooter isn’t enough to debunk many of the great conclusions from Gilovich.

A potentially more interesting paper was simply about a quirk of statistics.  This paper[3], by Josh Miller and Adam Sanjuro, suggests something very basic was at play, camouflaging the “hotness” of the shooter.  Basically, if you look at datasets and start observing outcomes after the shooter makes his first shot, you are creating a sample bias.  

Think about it like a coin flip.  If you are trying to test for skill at flipping heads, and you flipped a coin four times (or shoot a basket as a 50% FG shooter), there are 16 possible, equally likely, outcomes.

HHHT HTTH TTTH THHT HHTH HTHT TTHT THTH
HHTT HTHH TTHH THTT HHHH HTTT TTTT THHH

To get to the bottom of things, if we look at each sequence and start with the first head that we find, we need to then calculate how many of the next flips (shots) in the sequence were also heads. Of these 16 sequences, 14 of them have a heads (a make) that is followed by another shot.  Of these sequences, we can tally up the number of “heads” flipped directly after a heads was flipped, and then go to work.

As we see in the table to the right, the average FG percentage of all the sequences where the shooter (the coin-flipper) shoots again after his first make is just 40%!

The perhaps unintuitive observation is that if someone is a 50% shooter, and you go back in time to look at a string of finite attempts, and if you start with a make within any of those sequences, the average FG% of the remaining shots in those sequences is only 40%!

The motivation here is that there is this kind of statistical penalty given conditional probabilities in the analysis of finite data.  And thus if the analysis that Gilovich did only reveals that a 50% shooter is a 50% shooter after making a basket, that this means he actually is hot!  Basically, in this simple description of it all, anything better than 40% would imply hotness.

Finally!  The hot-hand exists!!  There was much ballyhoo over this in both the sports industry and in academia.  There a well-written description of the background here “How Researchers Discovered the Basketball ‘Hot Hand’”[4], and noted statistician Andrew Gelmer posted “Hey-guess what? There really is a hot hand!”.[5] Subsequently, Ben Cohen has written a tremendous book[6] describing the hot-hand in many aspects of life, and how Miller and Sanjuro ultimately turned the hot-hand fallacy into a fallacy itself.  

But hold on a second!  Let’s go back and add up all the attempts after flipping heads (after making a shot) and see how many of them were heads.  Within the sequences noted above, there are 24 attempts, and 12 makes.  That’s 50%, not 40%.  

In other words, the claim that  “a given flip of heads is more likely to be followed by tails than by another heads” doesn’t actually appear to be true.

So is the debunking of the hot-hand fallacy simply based on a feature of taking averages of averages?  Isn’t the likelihood of the 50% shooter making another shot after making the first one still just 50%?  

Our conclusion: if Gilovich’s data set took the average of the sequence percentages, we agree with Miller and Sanjuro, and there probably is a hot-hand fallacy fallacy.  If however Gilovich is looking at total shots taken within those sequences, then the hot-hand fallacy fallacy, is itself a fallacy.(i)

And in either event, Gilovich’s initial assertion is almost certainly still very true.  People do not have a good sense of randomness, and tend to see patterns (and causality) when they don’t exist.  None of this stuff is as intuitive as we’d like it to be.

‍

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

(i) An alternative interpretation of the heads and tails dataset is to look at further attempts remaining in the sequence after the first heads, and seeing how many of them were heads.  In this case, the “average of the average” is not as low as 40%, but still is not 50% (it is 46%), but perhaps more interestingly, of the 34 coin flips, or shots, after a head (or a make) is first revealed, only 16 of them are heads.  In this case the shooting percentage is 47% (16/34).  If this probability matters, then it too would help debunk the hot-hand fallacy, thus making this entire process a hot-hand fallacy fallacy fallacy fallacy. 😊

‍

FOOTNOTES

[1] https://www.cs.colorado.edu/~mozer/Teaching/syllabi/7782/readings/gilovich%20vallone%20tversky.pdf

[2] http://www.sloansportsconference.com/wp-content/uploads/2014/02/2014_SSAC_The-Hot-Hand-A-New-Approach.pdf

[3] https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2627354

[4] https://www.thecut.com/2016/08/how-researchers-discovered-the-basketball-hot-hand.html

[5] https://statmodeling.stat.columbia.edu/2015/07/09/hey-guess-what-there-really-is-a-hot-hand/

[6] Cohen, Ben, The hothand: the mystery and science of streaks, 2020, Harper Collins

‍

Download PDF

Topics

Factors
Sports
Stock Picking

DISCLAIMER

The views and opinions expressed here are those of the post’s author and do not necessarily reflect the views of Albert Bridge Capital, or its affiliates. This post has been provided solely for information purposes and does not constitute an offer or solicitation of an offer or any advice or recommendation to purchase any securities or other financial instruments and may not be construed as such. The author makes no representations as to the accuracy or completeness of any information in this post or found by following any link in this post.

‍

YOU MIGHT ALSO LIKE

Text Link

Heading

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Suspendisse varius enim in eros elementum tristique. Duis cursus, mi quis viverra ornare, eros dolor interdum nulla, ut commodo diam libero vitae erat. Aenean faucibus nibh et justo cursus id rutrum lorem imperdiet. Nunc ut sem vitae risus tristique posuere.

Text Link
Text Link

Investors? Possibly you!

FATMAN-G. No, that isn’t my rap name, although vaguely appropriate (I mean, the G part of course). That’s our new moniker for the Avengers[1] of the equity capital markets, to which we have now added Tesla. FATMAN-G is Facebook, Amazon, Tesla, Microsoft, Apple, Netflix, and Google.

Stock Picking
Markets
Read More
Text Link

Avengers Assemble!

Amazon is, very simply, a tremendous company. Google is also a tremendous company. The same goes for Microsoft, Apple, Netflix, and Facebook. These are the superheroes of capital markets. In these FAMANG stocks, the Avengers have assembled.

Markets
Stock Picking
Read More
Text Link

How Did This Even Happen?

No, I’m not talking about the winner of the Presidential election or COVID-19 or 2020. I’m talking about something I saw Rex Chapman tweet out over the weekend.

Sports
Read More
Text Link

Was “Value” Just a Hot Hand Thing?

In 1861, at the outset of the Civil War in the United States, Union Pacific issued 20 shares in its IPO.

Behavioural Finance
Factors
Valuation
Asset Pricing
Read More
Text Link

Cue the Camouflage

The S&P 500 is up 6.5% this year. That’s a total return number, and given what is going on out there, it’s pretty impressive.

Behavioural Finance
Markets
Stock Picking
Read More
Text Link

The Times that Try Stock-Pickers’ Souls

On first principles, there is one way to generate excess stock market returns over the long term, and it isn’t to “own winners at any price.”

Markets
Factors
Stock Picking
Valuation
Read More
Navigations
HomeTeamDrew's viewsPressContact
About Albert Bridge Capital

Albert Bridge Capital manages concentrated long-only equity portfolios for institutional investors.The hallmark of its Alpha Europe strategy is the application of tenets of behavioural finance to a rigorous, fundamental, process-oriented research process.

Regulatory Disclaimers
Stewardship codePillar three disclosureRTS 28Legal & regulatoryPrivacy policyCookies policy
How to get in Touch
Michelin House
Suite 109, 81 Fulham Road
London, SW3 6RD
info@albertbridgecapital.com
+44 (0) 203 826 0206

Subscribe to Drew's Views

No spam. Unsubscribe anytime.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
Albert Bridge Capital LLP is authorised and regulated by the Financial Conduct Authority of the United Kingdom © Albert Bridge Capital 2021.
Website by SW10media.com
Investors? Possibly you!
Avengers Assemble!
How Did This Even Happen?
Was “Value” Just a Hot Hand Thing?
The Hot-Hand Fallacy Fallacy Fallacy?
Cue the Camouflage
The Times that Try Stock-Pickers’ Souls
A Different Game?
Heads I Win
A New Ice Age?
Grandpa Stocks
Bubblicious?
On Negative Oil and Futures Prices
In Flew Enza
COVID-19 and Equity Markets
Regulators to the Rescue?
Perspective
We Don’t Make Pizzas
Ben Graham the Growth Investor?
Not a Bad Decade
On the Impact of the FAMANGs
Europe vs the US: Is it all about sector exposures?
Behavioural Finance is Finance
America’s Decade
Known Unknowns and Share Prices
Prediction, Publicity, and Paul the Octopus
Are Company Visits Good or Bad?
Everybody Was Kung Fu Fighting?
Voting Machines and Weighing Machines
The DCF is the Randy Watson of Valuation
The Sacrilegious Diaries: Measuring the Impact of Portfolio Turnover
Passively Irrational?
Imagine No Inflation
The Sacrilegious Diaries: The Benefits of Turnover
Stay in the Game
I’m Volatility?
Woody was Right
When You Can’t Wait For Tomorrow
James Harden and Alpha
Groundhog Day and Overnight Returns
Debiasing and Alpha
Peak “Peak Car” ?
The Right Way
The First Step to Regaining Credibility
The Futility of Market Timing
Visualizing the Arithmetic of Active Management
123 Years of the Dow
Sweet Emotion?
Share Buybacks, Bad Companies, and Bear Markets
Risk and Portfolio Theory
Keeping Calm and Carrying On
Reminiscences of a Stock Operator: The Volkswagen Chronicles, Ten Years Later
We’d Rather Not Sleep
Factor Timing, Should You Try?
The Mathematics of Maintaining Bet Size
The Grandfather of Behavioural Investing
On Sexual Chocolate and Semi-Annual Reporting
Island Economies and Risk
Build a Bear?
Data Science and Alpha
We’re all Value Investors
Hunting for Alpha
Career Risk, Alpha, and Contrarian Investing
Passive Flows and Wheelbarrows
God Bless the Shorts
Sell in May, and Go Away?
Equilibrium Happens
Horses and Stocks
Peak Quality?
Bill Sharpe and Hank Aaron
Unwarrented
The Search for Excellence and the Loser’s Game
Fooled by Non-Randomness
Half Hearted Is Half Minded
Still Superman, but without the cape
122 year Dow Jones Histogram: Putting 2017 into context
Pegs, P/E'S and the Value Premium
Rick Barry and Lewis Carroll
Secular Winners and Value Investing
Robots and Alpha
The Voting Machine
Prevous
Next
hometeamdrew's viewspressCOntactRegulatory