Last March I wrote about how Mark Lawrenson’s premier league predictions were a good example of System 1 biases (see here). I promised an update at the end of the premier league season. As one can see from the graph and table below, these biases persist over the whole season. He greatly under predicts the number of 0-0 draws, a particular error that I describe as a tendency to predict that something will happen rather than nothing will happen. He also over predicts the number of home wins and under predicts the number of away wins. His predictions of the final scores are biased toward lower scores (under predicts the number of games with more than 4 goals) and are also biased toward closer games (under predicts the number of games where the winning margin is greater than 2 goals).
So why doesn’t Lawrenson learn to correct his ‘mistakes’. Again it comes back to the whole idea of the automatic or impulsive part of the brain (System 1) overriding the reflective or controlled part of your brain (System 2). This generates predictions which are not rational especially when there is past information which should lead you to change your behaviour. It would be interesting to see if the trend continues next season. Assuming System 1 biases hold one would expect that it will.