Yes, had the same thought. Have tried updating based on 10-week moving averages of opponent attack and team defence, with and without a home advantage factor. That can make quite a big difference to individual combination scores but the overall impact is quite small (gain about 4 pts per GK pair).jabbamunter wrote: ↑23 Aug 2020, 11:26 Interesting. So the McCarthy/Ryan rotation wouldn't have worked out last year under this strategy!
I'd be interested to see how the table would look under a strategy of picking the team with the minimum of (GA + opponent GF). It could be further nuanced by home/away if desired. And also the proportion of correct decisions it results in (excluding tied scores).
Its naturally going to look worse for the better performing keeper. Part of what you get is protection against picking the wrong keeper up front. So the rotation should probably be assessed against the true upfront first choice or the average of the pair (even though it's an impossible score).
Can anyone remember the popular GK pair combination based on fixtures last pre-season?