Welcome to part 21 and the last post on the topic of possession.
Lets look now at expected goals conceded.
EPL average possession was 50%
EPL average goals conceded was 53.3
EPL average goals per 1% of possession was 1.06
Goals Against, Possession and Goals Per 1% Of Possession Conceded
Note- Goals against per 1% =(goals against/average opposition possession %---Stoke's 40.2% for becomes 59.8% against and that is the number we use)
Actual Goals Against v Expected Goals Against
The comparative numbers.
Actual Goals Against v Expected Goals Against (+/- differential)
So the differential graph shows us a few interesting things.
Arsenal are 5 goals below their expectancy number. Why? Well, Arsenal were outperforming (+) their individual shots against expected number, so it's fair to concede that if Arsenal conceded less shots than was expected yet conceded more goals the finger of blame must hover toward the GK and his save %. Arsenal's save % was the 2nd worst in the EPL.
Swansea are again worse than expected, but the - number is a small one. Wigan struggle,they have a weak defense but a good GK.
Wolves, a poor number meant relegation. QPR, Norwich and Bolton all are well below their individual expectancy numbers.
Blackburn, despite conceding less shots and shots on target than expected, conceded more goals (-16.7) than was expected when looking at the possession conceded numbers. Why? Same answer as Arsenal. Poor goalkeeper and, in all probability, a higher quality of shot conceded due to poor individual performances. Although, is it a coincidence that the 2 worst GK's in the EPL were Paul Robinson and Szczesny?
Goals Against v Expected Goals Against (+/- differential per game)
Here we see the same number as the previous graph but altered to +/- per game.
Man City conceded 0.41 less goals per game than was expected. That is a significant number over a 38 game campaign.
Man Utd (+0.32 per game), despite giving up more shots on target and shots than was expected ended up conceding less goals than their individual expectancy number. The leagues best team save % was at play here.
Liverpool conceded 0.20 goals per game less than their expected number. This just adds to the theory that their poor actual goals for v expected goals for differential number was the main, and, maybe only, reason for underachieving in 2011/12.
Fulham and Newcastle tread water in the big pool. Everton are the League leaders in this category. They had better numbers than was expected in regards to shots against and shots on target against and they followed those numbers up with a good team save % and thus conceded 0.42 less goals per game than they should have.
Sunderland were excellent by this measure and Stoke continued their better than expected results in the shots against, shots on target against and now goals against categories. They are a stubborn bunch and without these excellent and better than expected results in terms of conceding, I believe the awful shots for and goals for numbers would be left exposed and that relegation would be a real threat.
Edit- Good catch by Danny on twitter when he spotted that the 3 relegated sides had the worst actual GA v Expected GA differential. I was too tired to spot it first time around!
Thanks for reading.
I am going to sleep for a long while!