60-Second Mind

Artificial Intelligence Predicts Gambling Behavior

A simulated neural network is able to predict the bets and wins/losses of gamblers. Christie Nicholson reports














Share on Tumblr

Listen to this Podcast

[Below is the original script. But a few changes may have been made during the recording of this audio podcast.]

Artificial intelligence works to replicate our human mind (and a lot more). So scientists need to know how humans make their next moves. And while predicting behavior is inherent in AI, prediction itself is not often an end goal for AI.

But recently scientists have created a system that can predict the bets of gamblers, with an accuracy bordering on spooky.  Their research was published in the Journal of Gambling Studies.

Researchers used data from a total of 675 online games from six Texas Holdem gamblers. They built a mathematically simulated neural network based on the players’ initial several plays. Meaning the network learns and rewires itself based on guesses that either turn out to be right or wrong.  This method of AI is called propagation of error.

They found that their neural model could predict each of the six gamblers bet amounts with an accuracy to three decimal places of the dollar. Additionally they could also predict with similar accuracy their cumulative wins/losses.

So based on their first few games, the gamblers subsequent behavior, strategy and ultimately their wins and losses, was consistently predictable.

Proves, yet again, sadly there’s no magic in gambling.

—Christie Nicholson

(via Mind Hacks)


3 Comments

Add Comment
View
  1. 1. eco-steve 06:32 AM 7/23/09

    Unfortunately, this model did not predict the effects of bankers gambling on the bubble of house prices before the crash...

    Reply | Report Abuse | Link to this
  2. 2. JustMe 04:21 PM 7/23/09

    I'd like to know more about the players they took data from, and the nature of the games they were playing at the time.

    If they had been applying game theory to their actions then their bet sizes would have been sufficiently random as to defy profitable prediction.

    On the other hand, you only do this in poker when you don't know you have a big enough skill edge over your opponents - if these players knew they were sufficiently better than their opponents then they would allow themselves to be predictable (and hence exploitable) in order to maximize their returns from the lessor players who don't know how to exploit the informtion they are providing.

    Net result - this article doesn't really tell us anything about the quality of the experiment or its results.

    Reply | Report Abuse | Link to this
  3. 3. royniles 02:11 PM 7/29/09

    This is not so much about gambling as it is about the effectiveness of different strategies. Because poker is not simply a game of chance, it's a game of skill. The chance evens out, and you can "bet" that the most skilled strategist will win most in the end. Which if I'm not mistaken is essentially what this study predicted.
    The conclusion that there's no magic in gambling was a non sequitur. Try to use this system with blackjack or roulette and see what's predictable.

    Reply | Report Abuse | Link to this
Leave this field empty

Add a Comment

You must sign in or register as a ScientificAmerican.com member to submit a comment.
Click one of the buttons below to register using an existing Social Account.

More from Scientific American

See what we're tweeting about

Scientific American Editors

More »

Free Newsletters


Get the best from Scientific American in your inbox

  SA Digital

Latest from SA Blog Network

  SA Digital

Science Jobs of the Week

Email this Article

Artificial Intelligence Predicts Gambling Behavior

X
Scientific American Mind

Subscribe Today

Save 66% off the cover price and get a free gift!

Learn More >>

X

Please Log In

Forgot: Password

X

Account Linking

Welcome, . Do you have an existing ScientificAmerican.com account?

Yes, please link my existing account with for quick, secure access.



Forgot Password?

No, I would like to create a new account with my profile information.

Create Account
X

Report Abuse

Are you sure?

X

Institutional Access

It has been identified that the institution you are trying to access this article from has institutional site license access to Scientific American on nature.com. To access this article in its entirety through site license access, click below.

Site license access
X

Error

X

Share this Article

X