The well established probabilities say that the expected value of all players is around 7, for the winners around 8.4, and for the losers around 5.7. This is in line with what they observed in the experiment. What the “researchers” forgot is that the score of the winners is not the expected value of a two-dice throw,…

The expected value of the winner of such a game is around 8.37 (not 7), while for the loser is around 5.6. Of course if you have limited amount of plays then the average can vary around these values a lot. This mistake is so elementary I cannot even believe this.