Of course game flow plays a big part in xG on occasion, and nobody said it was perfect. It's meant to be used as a longer period of time marker anyway. So if over a long period, on average you dont concede many chances but create plenty, then whatever you are doing is working, even if finishing might not be favorable during a run. It's a stat like any other, but it's shown to correlate more to future points and predictors than others like shots or shots on target or actual goals.
United Hour - xG Gon' Give It To Ya
- Thread starter Rood
- Start date
No, I'm treating them independently as each shot is 0.3 on its own. Warning, maths incoming.
The probably of scoring at least 1 goal = 1 - the probability of scoring no goals
The probability of scoring no goals = p(missing the first AND missing the second AND them missng the third)
In general you multiply indepdent proabilities to do AND and add mututally exclusive probailities to do OR (e.g. probaility of rolling 1, 2 or 3 on a die = 1/6 + 1/6 + 1/6 = 3/6).
Therefore the probability of no goals = 1 - (2/3 * 2/3 * 2/3) = 0.703
Now, you can go the hard way around if you don't want to do the above:
Proability of at least 1 goal from 3 shots = p(scoring exactly 1) + p(scoring exactly 2) + p(scoring exactly 3)
This gets tedious:
p(scoring exactly 1) = P(score 1st AND miss 2nd AND miss 3rd) + p(score 2nd AND miss 1st AND miss 3rd) + p(score 3rd AND miss 1st AND miss 2nd) = 3 * (1/3 * 2/3 * 2/3) = 0.4444
p(scoring exactly 2) = P(score 1st AND score 2nd AND miss 3rd) + p(score 1st AND score 3rd AND miss 2nd) + p(score 3rd AND miss 1st AND miss 2nd) = 3 * 1/3 * 1/3 * 2/3 = 0.22222
p(scoring exactly 3) = 1/3^3 = 0.037
0.444 + 0.222 + 0.027 = 0.703 (same)
The reason the XG still somewhat balances out is because in that 30% chance of scoring at least 1, they might score, 2, 3, 4 etc.
I guess the only point I'm making is that XG can be misleading. In the example above an XG of 1 actually has a much larger probability of not scoring than scoring.
Last edited:
now that we can agree on !
I get your point but, regardless of the maths behind it, xG for each individual striker does have a close correlation to reality on the long term
The thing is that you're not looking for the probability of scoring exactly 1, 2 or 3. It's just finding the total xG based on the chances you created, you're overthinking things (also your numbers arent exactly right either). What you're trying to show is different from what xG tries to show, which is the expected total of goals you would score given all the chances as each chance is independent of one another. So if you have 3 penalties at 0.75 expected goals each, you would expect to score 2.25 goals from this. That number is different from the probabilities of scoring at least 1 goal (0.95), or 2 goals (0.86) or all 3 goals (0.58), as you'll always have 0<p<1. It's a different stat.
And no need to explain I took a good amount of stats in uni and use enough of it at work that I don't want to do it now! And that subject, these guys are statisticians for a living so they likely took far more than either of us as well so I'm pretty confident in trusting how they're collecting it!
I suppose that could apply to both Palace and Southampton in a way. xG has some relevance even game by game but I get the point that its too early to draw any meaningful conclusions.Cheers for this! Great to have that regularity now
On the xG front, I thought Castle's take was interesting yesterday, when he says that xG fails to address a few things. Such as winning a game in the first 30 minutes and not needing to create chances or falling behind in a match and desperately trying to create any sort of an attempt (which is what we've often had to do). And of course, xG for the small sample size of 4 games is almost irrelevant.
For me there are two fundamental problems we've seen this season that are quite coaching specific: We are nowhere near clinical enough and we don't come up with a way to play that mitigates that. So we're relying on individual brilliance there rather than a team philosophy. This is also painfully apparent in set pieces that seem to be not coached at all. Like, really, not at all. We counter. We press, at times. But we have no idea how to build meaningful attacks when we have the lion's share of possession. Secondly, we are still conceding sloppy goals. It's maybe not an easy problem to fix. Do we play a high line with two non-pacey centre backs? Do we solidify midfield to provide more than the appearance of defensive cover? Do we play Lindelof against a team that poses aerial threats or against teams with rapid counter attacks (like Leicester on Saturday)?
Isn't Kieran McKenna supposed to be some brilliant attacking coach? I worry that Ole is more a positive motivator - with an eye for talent, who has a coaching staff that's got virtually no experience bar Phelan when he really needs the expertise there to bring this team on. We can't develop these youngsters if all they get (I assume) is go out there and express yourself.
Clearly we are still a work in progress but there are some signs of what Ole wants us to be doing with the press and pacey forward line. Just lacking the clinical finisher, you would think that would be the main thing Ole could coach from experience.
A lot our play still relies on Pogba's creativity - will be a big test of the rest of the squad tomorrow when he is not there. We are likely to be forced to play a more defensive midfield with the creative responsibility falling to Mata/James/Peirera
The most positive signs are the players he's shipped out or dropped. Which takes balls. So it's clear he wants dynamism and pace. However, our attacking build up also needs technically sound players and that's where we come up short both in terms of players available and, I think, coaching. Early days of course. The midfield two scenario is also concerning. I don't see how we are getting the best out of Pogba, as you highlighted on the pod, and he also doesn't screen the defense. A 433 might be more suitable with the players available.
I am pretty desperate to see our set pieces improve from Sunday league level as well.
The doubts about whether he knows how to coach (or any of his staff) attacking play - intelligent quick movement, quick passing - as oppose to trusting in players figuring things out, still seem justified to me. But yes, I see positives also.
Respectfully disagree, and I don't think my numbers are wrong and I acknowledge that the probably of any event 0 <= p <= 1 and so won't go above 1, but I don't think you understood my point. I was merely showing that from 3 chances that the probably of scoring at least 1 = probability of scoring exactly 1 of them + probability of scoring exactly 2 of them + probability of scoring exactly 3 of them. In addition my point was that with an XG of 1 (under the conditions I showed) you are still more likely than not to score 0 goals.
Also, I think your maths is wrong, and I'm not sure you can expect to say "your numbers are wrong but hey don't tell me if I'm wrong" on an internet forum and not be called out on it.
With 3 penalties of 0.75 XG each the probabilities are:
- scoring 0 is 0.16
- scoring exactly 1 is 0.14
- scoring at least 1 is 0.98
- scoring exactly 2 is 0.42
- scoring at least 2 is 0.7
- scoring all 3 is (also) 0.42
Let's forget the maths lesson and focus on the fact that I actually got the score prediction right for once 
xG for this match checks out
1.1 Vs 0.5
xG for this match checks out
1.1 Vs 0.5