ClaphamFox

Very good discussion, as always. I think between you, Jake, Jordan and yourself came to a pretty fair and accurate summary of Cooper: tactically limited, unadventurous and prone to some maddening decisions, yet has made us much more determined and competitive than we were with a far superior squad during the relegation season. I agree that he isn't going anywhere unless we get dragged into the relegation zone, but I've got a feeling we won't. Jordan completely nailed it in his explanation of why xG stats are so limited and should be treated with caution. Re: Ayew, I suspect this will prove to be one of those maddening aspects of Cooper's approach mentioned above, and that we'll probably just have to learn to live with it. Like everybody else I would love to see us play Mavididi and Fatawu on the wings against Southampton, but I am as certain as I can be that Ayew will be on the right with Ndidi at number 10, with Winks and Skipp behind him.

Replacing one of our three DMs with a more naturally attacking midfielder also made perfect sense on Saturday and is exactly what many people were demanding. You can argue the toss over which one of Ndidi, Skipp and Winks should miss out, but based on recent performances I suspect many folk would agree that Winks was the right choice. So dropping Winks to the bench wasn't a remotely surprising decision in the circumstances.

Or possibly forced thinking due to lack of interested candidates (due to threat of points deduction) and/or lack of funds (due to PSR constraints)....

Mavididi has played in every league game this season, hasn't he?

I agree it isn't. But xG and xGA stats aren't very useful indicators either. The board is not going to sack Cooper based on the results so far and they're not going to sack him based on underlying stats. The only scenario in which he'd get the boot now is if there is significant player unrest and the board is sufficiently concerned about that to take action, but this doesn't seem a high probability outcome at the moment.

This is not true though, is it? We're in 15th place, with a significantly better goal difference than most of the teams below us. The results he has delivered so far have kept us out of the relegation zone, which is our sole objective for this season. If you offered the board a 15th place finish right now, they would snatch your hand off. Obviously if our results decline significantly and we get sucked into the bottom three, it will be a different story. But as long as we stay ahead of the teams below us, Cooper is going nowhere.

Good example - this is exactly what I’m talking about. In our last PL season we conceded eight more goals than the xGA stats say we should have. Why? Because we allowed a very good keeper to leave, didn’t replace him and allowed a poor quality reserve keeper to step in. If we hadn’t have made that mistake, we’d probably have survived. This season it’s the opposite - we have a keeper who is a net positive rather than a net negative. This means that if our xGA is exactly the same over the course of the season as a team with a worse keeper, we’ll very likely concede fewer goals. Personally I’d extend this to the defence as a whole. We may have had better players in 2022/23, but defensively we looked all over the place. Every time the opposition got a set piece or found space on the flanks, we looked so vulnerable. While we don’t have as many high quality attackers this season, our defence looks far more robust to me than in 2022/23 (although we’re still allowing the opposition lots of shots, we’ve conceded 10 fewer goals in our first seven games this season than we did in the first seven games of the relegation season). The point remains: some teams are simply more effective at one or both ends of the pitch than their rivals even if their overall play really isn’t that impressive. Those marginal differences in efficiency can persist over the course of a season and if they do, they can make a huge difference.

"When?!" Your absolute trust in the reliability of xG/xGA stats is akin to a form of religious faith. Yet I've yet to see any evidence that they are so reliable that mean reversion is inevitable over the course of a season. Indeed, the understat data discussed above suggest that xG/xGA stats are actually pretty unreliable at predicting how many points a team will end up with (here's the link again to save you from scrolling: https://understat.com/league/EPL/2023). At the risk of repeating myself, xG and xGA stats are useful in showing the quality of chances teams create and concede, but they are not even close to being a perfect predictor of outcomes over the course of a season. If they were, all teams would end up with points totals roughly in line with their xPts, but that does not typically happen. Last season the bottom five teams all significantly underperformed their xPts, which implies they were inefficient at taking their chances and defending against opposition chances. The same thing happened in our relegation season, when the x stats say that we should have ended up with 45 points (ie, comfortably mid-table) but somehow ended up with 11 points less and went down. Does that mean that poor old Brendan was just unlucky? Should we send a bunch of flowers and a crate of fake tanning lotion up to Celtic Park by way of an apology for all being so mean about him? I'm genuinely not a fan of Cooper. He's a tactically limited coach who appears scarred by his first few months in the PL with Forest and now defaults to conservative team selections and tactics, which make for a pain spectacle. He's also too slow to react and when he does react, he doesn't always get it right. But our performances so far suggest that he might just be quite good at setting the team up in a way that means we always remain in games and are a bit more efficient and attacking and defending than the x stats show. Anyway, fairly soon this discussion will be settled one way or another: either this mean reversion that some folk believe in so much will kick in and we'll be in the relegation zone (in which case Cooper will be gone), or we'll carry on grinding out results that the stats suggest we don't deserve and Cooper will go nowhere. I'm not ruling out the latter.

It also shows that a large number of teams obtained significantly more or less points than predicted by the x stats. If xG and xGA were really accurate predictors of points acquired over a season, surely we would not be seeing the kind of wild discrepancies shown in the red and green figures in the final column of that table? It's very notable that the bottom five teams all comfortably fell short of their expected points for the season. Is it not possible that while xG and xGA might be good art rating the quality of chances created and conceded, they are very poor at showing how efficient individual teams are at scoring and/or defending those chances? The assumption of mean reversion implies that all teams are equally efficient over the course of the season and therefore that the xG/xGA stats will be a good indicator of final points totals, but the table in that link suggests otherwise.

In 2022/23 our expected points total was far higher than the points we actually got, but I can't remember many people saying that Rodgers was unlucky and we should have stuck by him for longer. If it's that good, why does it predict teams' final positions incorrectly more often than not?

This article by Opta explains in some depth how their expected points model performed last year. It shows that seven teams finished in the positions predicted by the model and thirteen teams (65%) finished either higher or lower than predicted. So it's not exactly the hyper-efficient mean-reverting monster that some people seem to believe it is. https://theanalyst.com/2024/05/opta-supercomputer-premier-league-2023-24-predictions#:~:text=The Opta supercomputer has kept us

Yep. The same Dibling that has run the show so effectively for Southampton this season that they've amassed one point from seven games and a -11 goal difference.

While accepting that xG and xGA have their place, I’m not convinced they provide as full a picture as is sometimes claimed. Surely if they were a very reliable indicator of a team’s quality, mean reversion would dictate that all teams would fall back close to their expected results over a season? Yet a glance at last year’s final table suggests otherwise. Only three teams (Chelsea, Palace and Fulham) achieved final points totals that were close to their expected xPts. Every other team ended up a differential of more than five between their xPts and their actual points, and six teams ended up with a differential of more than 10. Interestingly, the bottom five teams at the end of the season all gained significantly less points than their xPts. It it possible that these xG/xGA stats aren’t really an accurate indicator of final points totals, but rather a measure of how efficient teams are at both ends of the pitch? If some teams consistently score more and/or concede less than they are expected to, could it be that they are simply quite good at the two most important aspects of the game? It just makes me wonder whether this habit we’ve developed under Cooper of remaining competitive in games while only performing well in patches and allowing the opposition lots of shots is not necessarily an early-season aberration, but rather just how things are going to be under this manager. Maybe shithousing results is our identity, for the time being at least? It might be a grim watch at times (well, a lot of the time), but if we continue to outperform our expected stats (and last year’s table would suggest this is possible), we’ll survive this season. The question will then be whether this approach is the right one to take us to the next level. https://understat.com/league/EPL/2023

They’ll absolutely batter us. Our best hope is that they have an off day and we keep it down to 6-0 or 7-0.

You’re not getting it. Every team we play is going to batter us until they don’t. Then it falls upon the next team we play to batter us. And on it goes.

But he’ll only be here all season if we win more than five or six games. If it looks like we’re heading for relegation, he’ll be sacked. His position is purely dependent on results, and so far ours have been better than five other teams. Until that changes he’s going nowhere.

Not a chance he’ll drop Ayew.

Last week there were reports that the players were unhappy because Cooper’s management style was too open and collaborative, and that they missed Maresca’s more dictatorial style because they like to have clarity. Now we’re told that Cooper is in fact a dictator who won’t listen to players’ concerns at all. And all the while this is happening the players are showing genuine determination and fight on the pitch. Wild times.

If the team keeps on grinding out results and stays out of the relegation zone, I doubt the board will give a **** about the players’ body language. It may be a grim watch, but if Cooper’s mandate is survival, he’s currently on track to deliver it. The board won’t do anything unless that changes.

I’ve said it before (and been greeted with scepticism), but I honestly believe Ayew only agreed to sign for us because he was assured he’d be in the first 11 when fit and Cooper is determined to honour that promise.

We’ve conceded first in five of our six games so far. In two of those games we’ve recovered to get a point, and in two others we got back on level terms before eventually losing. And even in the game where we never got back on terms (Villa), we arguably played well enough to do so. This shows that plan A is consistently not working, and that we only really start play playing when Cooper’s game plan is thrown out of the window and we introduce an element of chaos as we’re chasing the game. You’d think he might have learned something from this, but there are no signs as yet that he has. I think his rigid adherence to a conservative approach that doesn’t work will end his time with us sooner rather than later, but probably not today.

Excuse me? Who says I’ve bought into anything? I don’t think Cooper will last long with us, but when he goes it will more likely be because of an accumulation of narrow losses and draws than a thrashing at home. I don’t think we’ll get whacked based on what I’ve seen of us so far this season, but that doesn’t mean I like Cooper’s tactics or have ‘bought into’ them.

I think it the most likely outcome is a draw, followed by a narrow win for them and then a narrow win for us. I don’t think we’ll get thrashed because Cooper will set us up to avoid that. The problem is, he won’t set us up to win either, so it will probably be a 1-1 draw or 2-1 loss. I also don’t see much evidence that morale is low. I see a group of players who are being held back by a tactically very limited coach, but who are trying their best within those constraints.

Why will Bournemouth ****ing batter us when nobody else has done so and we’ve played several better teams than them?