• A reminder that if you give a thumbs up or similarly positive reaction to a racist comment you may also receive a ban along with the user that wrote the post.

Analytics in Football - Official Thread

Regardless of any of that, a massive portion of United's xG came at three down. City were clinical and didn't really face much threat at the other end until late on when the game was over.
It maybe reflected Utd should have gotten a goal or two at the end though no?

xG doesn’t always pass the eye test.
 
Interesting. The other thing not captured by that stat is when Utd break four on four (or better) and conspire to overhit a pass, run offside, turn back to pass it to Bruno rather than the sort of two pass movement that used end with Rooney crashing the ball past the keeper. We miss situations that most other Premier League teams translate into goals. They are creating these more routinely now though. They also won ball back high up the pitch but then mess up simple passes to turn same into high quality chances.

There’s no point talking down the chances due to context if going to ignore city being 2-0 up, Utd pressing up the field exposing the defenders they have that can’t play a high line that resulted in simple breakaway situations. Could take them off as well as didn’t require city to do much to create them.

What stat covers the ongoing capacity of Bruno, Dalot, Shaw and others to have utter brain farts that might lead to a goal directly or more insidiously cause the corner or loss of possession that led to the goal ultimately
This is the sort of situation I’m talking about (not even a very high quality one) but worth a look in first few minutes of this.


You HAVE to put the ball in early from Mazraoui. The likelihood of it resulting in a header or even falling to someone arriving tail off so much when you delay. Gary Neville puts that ball in first time. Valencia puts the ball in first time from there. Your average Premier League player puts it in first time there and you’ll get the benefit of it over time.

Mazraoui there is worrying about losing the ball, should he pass it, give and go, cut in. None of which increase the chance of the team scoring but do avoid him getting the blame.

The Antony’s/Garnachos there have a different problem where it’s “how do I make this about me?” while doing nothing that will genuinely trouble a top full back.

I like Diallo there as while he tends towards the latter group, he can at least surprise.

I think Amorim is not all at fault here or even mainly because this shit came in in two phases under Ten Hag and earlier under Van Gaal as Utd lost their dynamism entirely. Fielding wingers but telling them to play like a CM out there just recycling possession.

Utd would help themselves a lot by playing some sort of ball early and into space.

I don’t think there’s anything wrong with this system and it’s increasing the number of good positions we’re getting into. The execution is still poor.

Notably from what I can see the Utd youngsters have stopped playing Ten Hag style football (high risk, high precision, highly spread out) but aren’t playing this 3-4-3 formation. It is however more similar in style to what Amorim wants which I think is the man’s point re the formation is really something he’s using to force players into the style.

If I were him, I’d say I’m changing to 4-3-3 but say I’m playing Mazraoui in DM with a back four of Dorgu, Yoro, De Ligt, Diallo and Ugarte and Mainoo completing the midfield 3.

Then in game you realise the full backs are pushing on, with the two wide players tucking in and Mazraoui tends to fill in right behind Diallo and you end up with EXACTLY the same formation he’s actually playing but the pundits shut up about it for a couple of weeks.
 
Which is why it's only accurate over a long period of games. Two seasons, in my opinion. That's 68 to 76 games, and some people use it on a single game basis.
:lol!:

I re-found that article where it talked about rules of thumb around using xG for certain volumes of games. Nice idea, conveyed well by David Sumpter. My memory of the ~11 game threshold (for xGD to carry more signal than noise) must've been from elsewhere, but it's broadly consistent with David's article below


I had a play around last night re-creating it with PL data from recent seasons - the results are below.

1758014752348.png

I then took a slightly different view of it and, instead of averaging each of the 5 seasons per gameweek, I laid them out sequentially.

@KolaKubes this illustrates that in the long run, with a big enough sample, xG is highly reflective of the amount of goals you'll actually observe

1758014841576.png

@Cloudy O'Ranieri on your point about the 68-76 game mark, here's the data isolated from the last two seasons

Looks like it really starts to level off conisderably after ~200 games (roughly half a season)

1758014908752.png

And a drill-down into first 300 games of each of the 5 seasons (averaged). After ~150 games (~40% of the season) the averages meaningfully converge.

1758017766748.png
 
Appreciate that Phil. My point is that, largely, xG will be a good pointer reasonably early in a season, but there'll always be some outlier teams across a whole season (making it somewhat dangerous to pile betting funds into/against a team just because they are not obeying the xG, for example).

That will usually fix itself the following season. I would say after two or three seasons, you'd be hard pressed to find any team over or under performing by a significant amount. Again, I'm sure there will be exceptions to this, as always where stats are concerned, but it's a case of "the bigger the sample, the more outliers get pulled back towards the mean".
 
Last edited:
Yep, fully agree.

The task then becomes trying to drill in and understand why that handful of outlier teams are over/underperforming xG for an extended period.

Are they doing meaningful things which don't show up (enough) in the xG or have they just been lucky.
 
Yep, fully agree.

The task then becomes trying to drill in and understand why that handful of outlier teams are over/underperforming xG for an extended period.

Are they doing meaningful things which don't show up (enough) in the xG or have they just been lucky.
For me, that's where level game state xG comes into play. I think it might be slightly better than overall xG for immediate short term prediction, or it certainly gives a better picture of where a team are right now anyway.

Again, you need a good whack of games (maybe 15-20) to get a decent sample size and break it down to per 90 or ratio.
 
For me, that's where level game state xG comes into play. I think it might be slightly better than overall xG for immediate short term prediction, or it certainly gives a better picture of where a team are right now anyway.

Again, you need a good whack of games (maybe 15-20) to get a decent sample size and break it down to per 90 or ratio.
Like this stuff guys.

A more long range appreciation of stats might lead to assessing players on longer time lines for who is or isn’t playing well.

You can then make decisions about players and managers over an appropriate sample size of games.

One of my favourite ones is goals per appearance which makes no sense if they’re mainly coming on for 15 minutes but also depending on the game state a player coming on at Liverpool with them 3-0 up is more likely to score a goal and their stats will be juiced up because of that.

I’m starting to feel nostalgic for people just analysing how good a particular goal was and leave it at that.
 
Martin O'Neill: "Expected goals (xG) is total nonsense. You’ve got to remember what the game is about: winning football matches, and that means scoring goals, not recording the expectation of them. It’s a clueless development. Some people just use these words to try to sound clever."

Who is right? A successful player and manager or social lepers like Kloudy and Philomena?
The statistics should be used when applied across a long timeline and large enough sample size for comparing player and team performances.

They are not useful for analysing a game.
No improvement on simply looking at video and saying that was a good chance wasn’t it, Martin?
 
The Crazy Gang. Real football. Real men.

Unlike the xG dullards on here.
I think they could get in the automatic offside and then arguably the two lads on the line could be going in to watch the shit going on in the penalty area at set pieces. There’s an extra pair of eyes and so long as neither has raised a problem, goal stands.

I also suspect VAR is more open to corruption in an already corrupt sport.
 
What's On Today

Live Music

Ballads & Banjos

The Welcome Inn, What's On Today @ 9:30 pm

More events ▼
Top