@Cottage Analytica

New blogpost: Predicting Fulham’s Budget for the 2023 Summer Transfer window. In which I estimate Fulham's spending capacity for the summer and try to construct a plausible shopping list of potential transfers.... cottageanalytica.com/2023/06/29/pre…

It’s with great sadness that we’ve been informed that Chris from @CottageAnalytic died on January 18th from Lymphoma. Chris’ wife Jo asked us to let the Fulham community know on her behalf. She said that Chris loved creating the stats and reports about the club, and particularly enjoyed knowing that they were appreciated by so many Fulham fans. Chris will be part of the Celebration/Memorial Match held on Saturday 15th February vs Nottingham Forest. We’d like to extend our sincere condolences to Jo and her family at what must be an incredibly tough time.

@Serbiantennis77 I was not asked this for my 11 & 15 year olds 🤷

Fulham ticket office now demanding a copy of my sons birth certificate to renew his season ticket (that he’a had for 4 years and never needed to supply before) He’s 8 ffs

@Serbiantennis77 But the capital spend doesn't go through p&l though does it, or at least only in the form of depreciation which has historically been about £2m a year? Write downs could be a factor but Knockers and others carrying value must have been small by that season.

@CottageAnalytic Its definitely a bit higher than expected but could have been contract write downs for people like knockaert in there, for example. Capital spend could potentially be chunky also. I would have hoped to have been closer to break even with £183m revenues tho, no question.

Accounts not live on companies house yet. Headline loss looks bigger than I expected, but dont know if the club wrote any contracts down, what capital expenditure is in there etc. Revenues more or less exactly as expected.

Thanks for the shout out. Unfortunately I am dealing with some health issues at the moment, and don't currently have the energy to do the research behind my posts. I hope to be back in the future once I am better again. COYW!!!🖤🤍🖤 Chris

@CalbarFFC So it must then be the appetite for spending from the board, perhaps they want to run the business profitably which is not unreasonable

@CalbarFFC I don't have much to say, all my analysis says that since a big pinchpoint at the end of the 21/22 season we have had tons of ffp headroom, even ignoring the Mitro sale. Unless this is an undisclosed punishment for 21/22 then I don't think ffp is the issue. 🤷‍♂️

@CottageAnalytic fancy coming out of retirement to have a stab at where the Mitrovic money has gone? (Hope you’re doing well otherwise)

@timotheblueone I estimated the ratings of the promoted teams based on their performance in the championship. Those ratings have been fairly accurate, except Sheffield have been much worse defensively!

@timotheblueone It's based on something built by Excel4soccer, but they didn't release their money this year so built my own version. It uses my own system to predict xG and then adjusts based on actual performance. I have been surprised how often it predicts teams xG correctly. (1/2)

With 8 games played, I thought we are probably approaching a point where the premier league model I have been running has gathered enough data to make some reasonable predictions & observations, so here they are in a thread. Firstly, the predicted league table: (1/14)

@Serbiantennis77 Yes they would but I kept adjustments quite small so freak results wouldn't move things too much. In reality I think the adjustments are too small and the model is not responding as quickly as I would like, but if i made it more agile, then I would need a fix for outlier results!

@CottageAnalytic Ok got it - so potentially wierd/quirky results can throw the model disproportionately? (Our game v newcastle with 10 men for 85 minutes is a good example) Be interesting to see how things like that ultimately get absorbed over a full season 👍

@Serbiantennis77 The more a team exceeds the predicted xG, the bigger the adjustment. It's surprising how often the xG prediction is about right though

@CottageAnalytic Interesting - thanks as ever. One question, are the models’ adjustments equally cumulative (ie adds all performance data for all matches them averages them), or does it have any rules for one offs (sheff utd v newcastle surely not typical), or recency trends?

West Ham: Defence weaker than predicted Wolves: Attack stronger than predicted in the initial pre-season model. Other than these, the model has not significantly adjusted the team ratings since the start of the season. (14/14)

Everton: attack and defence both stronger than predicted initially Newcastle: attack much stronger than initially predicted. Defence slightly stronger as well Forest: Defence much stronger than predicted Sheffield: Defence much weaker than predicted (13/14)