Friday, December 10, 2010

Correlation between points and cost

In a post to start the season I looked at the correlation between points and a player's cost for the 2009/10 season. The correlation noted between points and cost for the season was as below:
  • Goalkeepers 62%
  • Defenders 56%
  • Midfielders 70%
  • Forwards 73%
These numbers were fantastic but they did suggest that it was probably worth paying for a couple of premium midfielders and forwards as they were much more likely to deliver on their high price tag.

With almost half a season gone I thought it might be useful to perform a similar exercise and see where the trends are going (I have used points per 90 minutes to eliminate the high level of injuries for the premium players):

Conclusion: there is basically no correlation for any position! Midfield (59%) and defense (53%) are showing small signs of correlation while forwards (35%) are the worst by far. At the forward position we are not too far from saying a 6m forward has the same chance of a big season that a 12m forward.

Sometimes not learning something in one area is useful for learning something elsewhere and I think this is the case here. While I have been only including stats for top players for captain consideration, this analysis suggests that anyone should be open for consideration and captaining a Nasri or an Elmander is no longer an undue risk. I will update my future captain rankings accordingly and probably sell my expensive players too! Now to find something which does correlate to points . . .

The weekly rankings for GW17 can be found here 

1 comment:

Andrew said...

Thanks for running this regression again!

Just an interesting fact I was looking at today: Andy Carroll has more games of 5+ points than any player in the game, and now has 9 in a row. Yet, he has only scored in double digits twice.