The irony here, tho, is they had a very decent defensive run the last 5 league games - or certainly a very decent one for a lower-table team. Giving up early goals hurt, but they just couldn't finish. My guess is they would have stayed up with Jesse if he didn't lose the lockerroom, but it sounds like he did, and that's a situation that's awfully hard to overcome.
Mostly, the ball didn't bounce right for them. Measured by expected goals, Leeds should be about 12th https://understat.com/league/EPL
We’ve discussed some of these xStats and I’m unconvinced about how much we can take from them. Perhaps it just demonstrates the limitations of such stats to tell us much. West Ham vastly out performed expectations last season and are now going the opposite direction - but why? Is it good luck and bad luck? Or is sustained success just incredibly difficult in the PL? For me the biggest red flag was JM said he’d never seen Ted Lasso. Those that fail to learn from history are doomed to repeat it…or something.
They are fairly accurate for pretty much everyone in the league except for Leeds and West Ham. And over time they always balance out. West Ham underperforming this year is those stats balancing out,
My favorite word from grad school - heteroskedasticity - uneven variation from the mean. Over a robust sample size performance should mirror Expected Goals (if your model is reliable) but even a 38-game season is not a huge sample size so there will be clubs that vary, some widely. I await the correction of those who know better (and I guess a sample would be an individual attacking move or shot opportunity within a game). mostly though, I saw a manager Sunday who had run out of ideas and lost the ability to inspire a higher level of performance. It reminded me of Ben Olsen coaching DCU in big games, where other teams were able to increase the effort/intensity but Olsen always kept the team in such a high state that they had no next gear when it was needed.
Well it is true that the longer the sample size the more the underlying numbers and actual numbers will align. And 38 is definitely not a large sample size.
I don't disagree with this statement AT ALL. However, there was nothing JM could have done to push Gnonto to finish that outside-the-boot shot he had in the first 10 minutes, or Bamford to not completely whiff on a slow rolling ball inside the box (or take a donkey touch on that deflected pass that went into his stride in space or otherwise just playing like crap), or the other dude who the ball did reach from hitting it right at Navas. Their finishing was just brutal whereas NF basically scored their only real chance of the game. I talk myself in circles about whether Jesse was unlucky this season or if he bears most of the blame for not being able to create his own luck. I think getting fired without having a chance to use the new acquisitions was very harsh, but as others have noted results are what matters. Of course, on the flip side, Potter hasn't been fired from Chelsea despite performing worse compared to expectations than Marsch has....
In 9 seasons, Lewandowski scored 252 goals with an xG of 252.4. He had one season where he underperformed xG by 11 goals. Just an example.
Was one of my favorite words from grad school too. I think it is important to note that heteroskedasticity is not like regression to the mean. It isn't something that will necessarily "go away" in the limit (i.e. become homoskedastic). It just means that when modeling/predicting the performance of clubs (E.g. xG), that the variance is non-constant across clubs. If you were treating this like panel data (each club over time) then you could also have different variances over time (across and/or within N/t). If I had gotten re-obsessed with soccer earlier in my life I may have gone into sports analytics instead of spatial/temporal. Get paid to watch and then analyze soccer? What could be better?
Jesse was unlucky, he's also not astute nor creative enough to make any in-game changes that have any real impact. Would've loved to have seen him get two more games, especially with the impact I believe Wes could've had with this team. But in the end, the club moved to try to save itself and no one can really complain. I don't think any of our guys there take a big hit in playing time. Wes has already proven himself to how many coaches across Schalke and Juve? A Leeds coach in the Premier League will be no different I'd think. As for BA, the kid deserves a smaller role and hopefully he can regain his early season form as a super sub
The only person who has outperformed his xG over time is Messi I believe. And that’s just Messi being Messi I think.
Based on Bamford's post match interview and the way the UK media interpreted them, you are spot on in assuming he was a player happy to see Marsch go. He clearly did not think the tactics were up to par. Of course, Bamford should worry about other things - like having a decent first touch and staying healthy for an entire calendar month.
Every team Leeds plays seems to score on their real chances. I'm not sure about the heteromegaectoplasticity or even the sample sizing of that but it seemed to occur in a frequency that would suggest the manager was missing something.
Bamford spent the majority of Marsch's time at the club in the treatment room. Maybe there wasn't enough reading material or video games to keep him occupied while doing rehab and he held Jesse accountable for that. Cant possibly see how it would be game related because that would be a small sample size.
We all also know the reality of high level professional coaching. Established coaches who have proved themselves in situations considered “high level” tend to get a longer leash - as do famous/respected former players. Newer/lesser known coaches tend to get a shorter leash. A Lampard or Pep or van Gaal might get a few more games or even the chance to finish out a crap season (but even they can get canned of course). But a lesser know, trying to prove themselves coach usually gets less runway. There is also a recruiting factor, I’d imagine. While the ability to attract talent is largely financial, it is also somewhat coach driven. If you are losing, some players lose faith in you and you aren’t finding and convincing undervalued players to come play… your job is going to be in balance - fairly or not.
I would definitely not say clearly. In his last full season’s worth of games, their GD was like -30, significantly worse than Marsch’s full season’s worth. He definitely had them riding a high following promotion, but once he was figured out by the rest of the league, they were toast. Funnily enough, that’s really the same criticism against Jesse. There’s definitely a recency bias against Marsch. And just like with a breakup, you remember the good times with your ex (Bielsa) and not the bad… they were definitely going down with Bielsa last season. They were in 16th or 17th but all the other relegation candidates had one or two games in hand. Whether they were going down this year before the Marsch firing is up for debate. I lean towards they would stay up, the defense has been much improved with Wober, and they were just beginning to integrate Rutter and McKennie. Plus Jesse usually hs leeds perform wel against the bigger clubs, and they have Man U twice in a week. Weird time to fire him.
Fulham, Crystal Palace, Southampton, West Ham, Leeds. They are accurate until they aren't. If they are inaccurate over a full season - that's a pretty big balancing out. And Liverpool may not be underperforming in the expected stats - but they are sure are underperforming expectations big time. Which is to say - soccer is unpredictable and sustaining success is really difficult especially when there are multiple teams spending as much or more than you. Before the season started, people were ready to just give City the trophy - Haaland, KDB, Cancelo, etc.. If the quality of the squad is not relegation fodder but is more mid-table, that's arguably even more of an indictment of the coach - unless you believe bad luck over a full or half season is the reason. But Klopp isn't a bad coach - which leads back to the unpredictable and difficult to sustain things.
I don't buy the Jesse usually has Leeds perform well against big clubs. While it's true they beat Chelsea and Liverpool, that has at least as much to do with those teams being messy this year as it does any great performances on Leeds part. Anyway, those two wins against big clubs struggling are half their wins. Ultimately, Leeds are going to stay up or not based on their results against the rest of the table, and they were not getting it done there.
Bamford has been utter dogshit this season, he should never play for Leeds again as far as I’m concerned. You can tell he was dogging it. No professional player can have that deficient of a first touch after showing the complete opposite in previous seasons. Jesse, like a fool, put his faith in that Judas. Should’ve put Rutter in from the beginning. screaming at the TV to get his ass off the field. Every time he touched the ball it was a disgrace.
Looking at Understat, Messi at (+33 goals above his xG) is one of the best out performers of xG since 2014-15 Son (+28), Mbappe (+18), Haaland(+18) and Kane (+24) are also very high. Lewandowski was basically 0 which was surprising. Bamford is -7 and Rodrigo -10 (although his Leeds values are quite good)
I'm not a Jesse fan but I would have given him more time. Really curious to know if he lost the locker room and if WM would have come regardless if he knew this was coming. Or if this was just the fans losing their minds. I would agree that with Bielsa, it's complicated. I'm not sure it's just that the league figured them out - I'd say it's more that his success required unsustainable demands on the players. These players were new to the league, riding a high of being at the top level after so long - and after that euphoric season, they crashed down to earth. It's also a little unfair to compare them - to Jesse's team. But Stuart Dallas (!) had 8 goals, 3 assists and 12 clean sheets that first season. Bamford 17 goals and 11 assists - which is mindboggling in retrospect. For whatever reason - Jesse hasn't really unlocked many mediocre players to play above their head and that's arguably a credit to Bielsa and his system.
I've not been all that interested in the xG etc movement, but it seems to me you are identifying two separate questions: Does xG ~= expected results over time Does a team's xG (and/or results) accurately reflect the expected quality - or is expected quality influenced by things that never show up on the field (club rep, players who used to contributed but don't anymore, etc.) In other words, a team you expect to do well can be poor in both xG and league points - and they are less likely to revert to the mean - the expectations are just not realistic. They just aren't living up to their rep/sum of their parts. And a team whose xG suggest they should have more points may/will eventually "revert" to their expected point level - the question is just if they can do it in the allotted time frame. And that's where, to some large extent, coaching quality comes in. Because reverting to the mean over 45 games doesn't help if you need to match your xG by May.
Jesse did get better performances out of Rodrigo, Struijk, Meslier. But yeah, some of Bielsa’s main guys never got going under Jesse…. cooper, Bamford, Harrison, Ayling. Dallas was never healthy under Marsch. Who knows if he’d have made a difference. Many of Leeds best players this year have been new signings under Marsch, which I think would’ve been another reason to give him some more time with the January signings. I agree, it seems a little shady to recruit McKennie, presumably with Marsch as a main selling point, then fire him after 30 minutes with his full squad.
fans and the players may have been willing to smash through walls for Bielsa ... perhaps for some time they would have for Jesse ... but just not this year?