Beavis FPL @beavisfpl.bsky.social · 19/11/2024What are some of the most common pitfalls you still notice with the use of data in FPL? Here are 3 things I see a lot of: - Small sample sizes - Bad data points - Assuming that all points are equal More in each of these in the posts below. 10466
Rohan @brohan03.bsky.social · 19/11/2024agreed, each data point needs to be analysed with a pinch of salt.. be it negative or positive. 000
Nima ⚽️ @fplnima.bsky.social · 19/11/2024Great list and clearly explained Beavis 💪 My pet peeve is comparing two players solely on touches in the box and ignoring the context of these touches from the two players respectively. 010
FPL Redsky @fplredsky.bsky.social · 19/11/2024Agree with all these! Kind of fits into the first one but I would say not taking into account fixture difficulty. For example, using Arsenal's current xGC for the season to conclude they are no longer an elite defense 010
Filip Novák @filipnovak.bsky.social · 19/11/2024I agree with all that. Regarding FPL, the issue is generally not giving proper context to data when working with a low(ish) sample size. And I'm glad you mentioned grouping xG/xA into xGI, which is super bizarre yet widespread. You're one of the few who mentioned it 020
Alasdair @spicerackjack.bsky.social · 20/11/2024Good points. Small sample sizes are a big issue for me. There will be posts early in the season that a player had bad or good stats which ignore everything they did in previous seasons e.g. Salah has blanked for four GWs so he's rubbish! 000
FPL Crisis (Ash) @fplcrisis.bsky.social · 19/11/2024Using single metrics to justify a narrative. “DCL still getting good xG” yes but he’s a poor finisher. “Rogers is bad FPL player because he misses too many big chances” etc etc 010
Beavis FPL @beavisfpl.bsky.social · 19/11/20241) Sample sizes (1 of 2) - People get too easily swayed by small samples and convince themselves that very recent data is signal of future points, ignoring years of xG/xAG data as an indicator of player output. 150
FPL MUFASA @fplmufasa.bsky.social · 19/11/2024Most people just don’t really know how to use this data. They mostly just check out the fixtures, recent points, and what other players are saying. And when it comes to making mistakes, if they hear something from someone who only looked at a small sample size, they got influenced and make mistake. 000
NighthawkFPL @nighthawkfpl.bsky.social · 19/11/2024Succinct and to the point. Agree to all the observations. 020
FPL Hippie @fplhippie.bsky.social · 19/11/2024Yes bsky.app/profile/fplh...FPL Hippie@fplhippie.bsky.social"Whilst calculating the XG of my FPL team, our computer also predicted that this season, a large gorilla will climb a building and throw barrels at a plumber." 000
