Gameweek 22 Preview Data

Clean Sheet Rankings





Attacking Rankings



Individual Rankings


On a recent FanGraphs podcast, guest Dan Szymborski spoke briefly about the difference between projections and predictions. Szymborski is the creator of the ZiPS baseball projection system which uses a combination of historic data and minor league stats to project future production. The key point here for us is that Szymborski notes that he is keen to avoid having a computer model do things that a computer isn't very good at, such as judging whether a niggly knee injury will keep a player out or whether a manager gave a hint in his latest press conference that someone might be rotated. He therefore distinguishes between projections - which are computer generated and focus on factors we can quantify - and predictions which use those projections along with other factors that are best left to the human mind to factor.



Given a few comments I have seen where the below data has been referenced elsewhere, I thought it important to underline the fact that these are very much projections and do not include reference to rotation risk or minor injuries. Where a player is suspended or ruled out I try and exclude them from the listing, but I am generally leaving the impact of more subtle decisions to the reader.



You may also note the absence of players like Javier Hernandez or Daniel Sturridge who look at least likely to feature this week for their respective teams. The reason is that their sample size is extremely small and thus, for example, Sturridge would be the number one non-double gameweek player this week, which just doesn't feel right based on what we know. If people feel it would be useful to include this data, I could potentially add a filter that could be turned on and off, so if you'd like to see that, please post a comment below.