New$10 Sign Up Bonus & $30 Free Trial Reward
Start challenge

Gambler's fallacy

Also called Monte Carlo fallacy

The mistaken belief that, in a random process with independent outcomes, a run of one result makes the opposite result 'due', for example expecting red after several blacks at roulette.

Reading
2 min
Sources
4Research
Updated
On this page· 5 sections
  1. How it works
  2. Example
  3. Watch out
  4. Sources
  5. Related terms

A roulette wheel has no memory. After a long run of red, black is no more likely than it was before the run began, yet most people feel that black is now due[3].

Traders meet the same pull after a run of losing trades, when the next winner starts to feel owed.

How it works

Amos Tversky and Daniel Kahneman traced the fallacy in 1974 to “local representativeness”[3]. People expect every part of a random sequence to look random, with plenty of switches and few runs[3]. So chance is seen as self-correcting: a deviation one way should bring one the other way[3]. In fact deviations are not corrected as a random process goes on. They are merely diluted[3].

Matthew Rabin turned the idea into a model in 2002. In the model, a believer in the “law of small numbers” exaggerates how closely a short sequence of independent signals should resemble the long-run rate[1]. Such a person expects that early draws of one signal make the other signal more likely next[1].

The fallacy also shows up where money is at stake:

  • Rachel Croson and James Sundali studied security videotapes from a large casino in Reno, Nevada: 18 hours at one roulette table over three days in July 1998, with 139 players placing 24,131 bets[2]. After runs of one to four, the split between bets against the run and bets with it stayed within chance of 50:50[2]. After runs of six or more, 85% of even-money bets went against the run[2].
  • Charles Clotfelter and Philip Cook tracked 52 consecutive winning numbers in Maryland’s three-digit daily lottery, drawn in March and April 1988[4]. Three days after a number won, the amount bet on it had fallen in every case[4]. The median betting ratio fell from 1.00 on the drawing day to 0.64 after three days and recovered to 0.93 after 84 days[4].

Example

Say each trade of a method has an independent 50% chance of a loss.

  1. Before any trade, the chance of five losses in a row is 0.5⁵ = 1 in 32, about 3.1%.
  2. The chance of six in a row is 0.5⁶ = 1 in 64, about 1.6%.
  3. After five losses have already happened, the chance that the sixth is also a loss is still 50%.

The fallacy mixes up steps 2 and 3. Six in a row was unlikely before the streak began. Once five are on the record, only the sixth is still uncertain, and it is a coin toss.

Watch out

The fallacy applies to independent outcomes. Rabin’s model is about signals that are independent and identically distributed[1]; whether a series of trades is independent is a separate question about the method and the market.

Experience is no cure. Tversky and Kahneman report misconceptions of chance among experienced research psychologists too[3].

The opposite error sits close by. When unsure of the true rate, Rabin’s believer over-infers from short sequences and concludes the rate is more extreme than it is[1]. In the Reno data, 111 of the 139 players quit after losing a spin and 28 after winning, which the authors read as consistent with a belief in hot streaks[2].

All terms in Tronexopedia

Your challenge starts here

Start your challengeFree TrialNew
Registration bonus

Get $10 free in your Wallet when you sign up

Register and use your $10 Wallet credit to reduce the price of your first challenge.

The $10 bonus is credited to your personal Wallet and can only be used to reduce the price of a challenge purchase. Credits cannot be withdrawn or used for payouts.

Start now & get $10