When doing this, to be statistically correct, you should also remove the top 2 scores, as suggested above.I was talking about last year mate, there's a bigger sample of games to read...
Lid's first two games this year were lower than his usual scores because he had no match fitness. Ofcourse taking out his best 2 games will drop his average heaps, that's just under a third of his games. Last year there's a sample of 17 games, his two lowest were clear outliers, that's why I took them out when I calculated that average