Sunday, May 24, 2015

End of Season Predictions (Spoiler: Summer is Coming)

Last round Mark Lawrenson got five games out of ten right for seven points.  His celebrity guest got three games right for three points.  Merson got six games right for eight points.  I got four games right out of eleven for four points.

Since I started making these predictions I have predicted 133 out of 270 games correctly for 196 points.  Lawrenson predicted 131 games out of 266 correctly for 181 points. Merson predicted 134 games out of 262 correctly for 189 points.

I predicted 49.256% of the results correctly.  I earned .73 points per game.  Lawrenson predicted 49.248% of the results correctly and earned .68 points per game. Merson predicted 51% of the results correctly and earned .72 points per game.

Overall there was not much of a difference between the accuracy of our predictions. My system was slightly better at getting the score right—almost certainly because I never predicted a two goal victory.  I was significantly ahead before the relegation run in when the experts were allowed to predict victories for clubs fighting relegation whereas my system had no knowledge of relegation.  Likewise, my system did not notice how bad Newcastle had become.  I was also hurt by surprise results in the four or eight games I predicted which the experts skipped.

If my goal was to suggest that the expertise of paid pundits did not really enhance their ability to pick the results of Premier League games, I have succeeded.  Of course, I was relatively confident of this because any pundit who could reliably pick winners would be making money by betting on the games, not by writing about them.

Over the course of the summer I will decide whether I want to do this again and, if so, I will evaluate and revise my prediction system.  The problem is that any improved system will inevitably become more complex which defeats the purpose of the whole thing.  For example, I might base predictions on the results of the last ten games, but that is not so easy to figure out (except exactly ten games into the season, of course).  In any case, I would have to come up with a system to predict the results of the first several games of the season when I have no current season results to use.

As for my summer predictions:
We will sell some players to other clubs.  Some fans will panic.  We will buy players from other clubs.  Some fans will be quite certain they are not good enough.  Some of them won’t be.

MAGIC NUMBERS
Here are our final magic numbers for the entire League as explained here:

Magic numbers had no real meaning now that the season is over, but I am going to finalize the list anyway.

Chelsea                 30

Man City              20

Man U                  11

Arsenal                 21

Tottenham           5

Liverpool             3

West Ham           -12

Swansea              -3

Stoke                    -5

Newcastle           -20

Everton                -12

Palace                   -11

Sunderland         -19

West Brom         -16

Aston Villa           -19

Burnley                -26

Hull                       -24

QPR                      -29

Leicester              -18

The points total mathematically needed to avoid relegation is now 36. We finished in 7th place.

No comments:

Post a Comment