Mic09 Posted July 16 Share Posted July 16 5 minutes ago, Chindie said: xG is one of the worst things to happen to the game. That's like saying probability is the worst thing to happen to maths. Something is likely or less likely to happen. That's it. Some people have a piano land on their heads, but the xPiano is well <0.01. 2 Link to comment Share on other sites More sharing options...
Genie Posted July 16 Share Posted July 16 Just now, Mic09 said: That's like saying probability is the worst thing to happen to maths. Something is likely or less likely to happen. That's it. Some people have a piano land on their heads, but the xPiano is well <0.01. How do they decide how likely it is that Saka scores from that position? It is completely flawed and I’m in the camp of waste of time too. Link to comment Share on other sites More sharing options...
Panto_Villan Posted July 16 Share Posted July 16 I'm glad Southgate has stood down, both for our benefit and for his - I think another tournament would have shredded his reputation. Ideally he'd have gone after the last tournament too, but I think overall he can leave with his head held high - he's built some good foundations and led England deep into several consecutive tournaments. A better manager would have gone all the way, and Southgate is clearly not elite, but I don't think he's as terrible as most people on here make out. I really hope we get a top-notch manager to follow him, as we were never going to actually win a tournament with Southgate's brand of football. Link to comment Share on other sites More sharing options...
mykeyb Posted July 16 Share Posted July 16 8 minutes ago, StefanAVFC said: Southgate to Gerrard would be so ridiculous I'd laugh like a drain That would kill the very tiny amount of interest I have in the national side. Link to comment Share on other sites More sharing options...
Davkaus Posted July 16 Share Posted July 16 1 minute ago, Stevo985 said: No, that's exactly why xG works. Because it is a benchmark for the average player. If you tailored it to individuals (those stats do exist) then it actually would be meaningless. xG is not a predictor But it's basically being used as one in that tweet, which is why I said in this regard Link to comment Share on other sites More sharing options...
Stevo985 Posted July 16 VT Supporter Share Posted July 16 8 minutes ago, StefanAVFC said: Southgate to Gerrard would be so ridiculous I'd laugh like a drain There is absolutely no logic that exists that would conclude in Gerrard being right for this job. The ideal candidate is someone who can carry on the positive culture that Southgate has helped to create, and actually has a bit about them tactically to evolve the team on the pitch. There could be an argument for getting a candidate that gives you one of those things, depending on who is available. Gerrard would give you neither Link to comment Share on other sites More sharing options...
tomav84 Posted July 16 VT Supporter Share Posted July 16 1 minute ago, Genie said: How do they decide how likely it is that Saka scores from that position? It is completely flawed and I’m in the camp of waste of time too. historical data. how many goals have been scored through similar situations. it's akin to risk management techniques i'd have thought Link to comment Share on other sites More sharing options...
MrBlack Posted July 16 Share Posted July 16 22 minutes ago, HanoiVillan said: That tweet just shows how weak xG is as a measure of anything. The claim that 'relying on screamers' is 'unsustainable' is rather belied by the fact that something he thinks has a less than 1 in 20 chance of happening happened in 4 out of 6 matches. The claim is basically 'well it worked in reality, but it shouldn't have worked according to my model'. Also, regarding you questioning the betting maths... if we take your figure of a 1 in 20 thing happening 4 times of of 6.... That has a 1 in 177,285 chance of happening. If we assume Southgate need it to happen a 5th time to get us to extra time and a win on pens, it takes it up to a 1 in 3,368,421 chance of happening. Funnily enough, that didn't happen. Now I know those figures were plucked out the air, and I know that xG has its issues. But it's a fact Southgate played low percentage football that kept things tight at the back but gave us little chance of scoring but for world class strikes from our world class forwards. 1 Link to comment Share on other sites More sharing options...
Stevo985 Posted July 16 VT Supporter Share Posted July 16 2 minutes ago, Genie said: How do they decide how likely it is that Saka scores from that position? It is completely flawed and I’m in the camp of waste of time too. They log thousands and thousands of shots that have been taken from that position before and work out how many of them went in. That give you your probability. Who is taking the shot is irrelevant Link to comment Share on other sites More sharing options...
penguin Posted July 16 Share Posted July 16 Just now, Genie said: How do they decide how likely it is that Saka scores from that position? It is completely flawed and I’m in the camp of waste of time too. From how often a goal is scored when attempted from that position, it’s not perfect as defender positioning etc. isn’t taken into account but overall it’s pretty accurate when looked at over a big enough sample. Link to comment Share on other sites More sharing options...
Mr_Dogg Posted July 16 Share Posted July 16 28 minutes ago, HanoiVillan said: That tweet just shows how weak xG is as a measure of anything. The claim that 'relying on screamers' is 'unsustainable' is rather belied by the fact that something he thinks has a less than 1 in 20 chance of happening happened in 4 out of 6 matches. The claim is basically 'well it worked in reality, but it shouldn't have worked according to my model'. Steven Gerrard likes this. 1 Link to comment Share on other sites More sharing options...
Stevo985 Posted July 16 VT Supporter Share Posted July 16 2 minutes ago, Davkaus said: But it's basically being used as one in that tweet, which is why I said in this regard It really isn't. It's being used to show that those goals we from low percentage chances and that in general results are unsustainable if that's all you're creating. 1 Link to comment Share on other sites More sharing options...
Davkaus Posted July 16 Share Posted July 16 Just now, Stevo985 said: It really isn't. It's being used to show that those goals we from low percentage chances and that in general results are unsustainable if that's all you're creating. But those low percentage chances are, for a team of the quality of England, probably far less likely than the stat would suggest. And yes, that's how xG works, I know that, so we're just going to go in circles here. I regret posting it because this had lead to an entire page of incredibly tedious posts 1 Link to comment Share on other sites More sharing options...
Mic09 Posted July 16 Share Posted July 16 (edited) 8 minutes ago, Genie said: How do they decide how likely it is that Saka scores from that position? It is completely flawed and I’m in the camp of waste of time too. It's based on analysis of thousands of similar shots - counting in factors of how many players are in the way etc. You can argue all you want about how this is interpreted, go crazy. I will join you. But arguing against it as a metric of something more/less likely to happen is simply not statistically sound. In Gareth's case, England have incredible players. So the likelihood of a small xG situation turning into a goal is more likely. Simple as. Messi running through the entire pitch and dribbling through 5 opposition players is a very low xG. But we have seen it over and over again. Because he turns LESS LIKELY situations into goals. Edited July 16 by Mic09 Link to comment Share on other sites More sharing options...
Stevo985 Posted July 16 VT Supporter Share Posted July 16 Just now, Davkaus said: But those low percentage chances are, for a team of the quality of England, probably far less likely than the stat would suggest. I wouldn't say "far less", but yes that's true. But it doesn't change the point of the tweet Link to comment Share on other sites More sharing options...
villan95 Posted July 16 Share Posted July 16 7 minutes ago, Stevo985 said: They log thousands and thousands of shots that have been taken from that position before and work out how many of them went in. That give you your probability. Who is taking the shot is irrelevant Tonev must have dragged the average down a ton to be fair Link to comment Share on other sites More sharing options...
HanoiVillan Posted July 16 Share Posted July 16 16 minutes ago, Davkaus said: I also don't think you can say that things are sustainable, a good approach, or the likely outcome just because they happened, but I'm just arguing the toss at this point Where I think xG does fall down in this particular regard is that it's a measure of how likely a chance is to go in taken from that position, by the average player. Saka, Bellingham, and Watkins aren't the average player, and I'd wager if you put Ollie in that spot he's scoring it more than 6 times out of a hundred. Well I can say it was sustainable, because it literally was sustained throughout the entire tournament. Luckily for me I'm not saying and haven't said that it was 'a good approach' or 'the likely outcome', just that it's a problem for the model if things it says are extremely unlikely keep happening over and over again, and that even if you do chalk this up to a truly extraordinary level of luck it makes no sense to claim it's 'unsustainable' *after the tournament is over* when it has been sustained! I agree with your second paragraph, though that's probably far from the only problem with it. Link to comment Share on other sites More sharing options...
Stevo985 Posted July 16 VT Supporter Share Posted July 16 Just now, HanoiVillan said: Well I can say it was sustainable, because it literally was sustained throughout the entire tournament. Luckily for me I'm not saying and haven't said that it was 'a good approach' or 'the likely outcome', just that it's a problem for the model if things it says are extremely unlikely keep happening over and over again, and that even if you do chalk this up to a truly extraordinary level of luck it makes no sense to claim it's 'unsustainable' *after the tournament is over* when it has been sustained! I agree with your second paragraph, though that's probably far from the only problem with it. It hasn't been sustained though. The tweet is talking about sustaining results while relying on low xG chances We won 2 out of 7 games in normal time. It's not sustainable It would be like like rolling a 6 on a dice four times in a row and saying "That's sustainable because it's happened 4 times" Low probability things happen. It doesn't mean they'll keep happening 3 Link to comment Share on other sites More sharing options...
HanoiVillan Posted July 16 Share Posted July 16 26 minutes ago, Stevo985 said: That's not what it's saying though. Yes it is. *Your* argument is that xG is not a prediction. *His* argument is that it is 'unsustainable'. That's a prediction. Your argument might be more defensible than his, but don't be confused that you're making the same argument. 1 Link to comment Share on other sites More sharing options...
Stevo985 Posted July 16 VT Supporter Share Posted July 16 2 minutes ago, HanoiVillan said: Yes it is. *Your* argument is that xG is not a prediction. *His* argument is that it is 'unsustainable'. That's a prediction. Your argument might be more defensible than his, but don't be confused that you're making the same argument. You've misinterpreted the tweet. Imo of course Link to comment Share on other sites More sharing options...
Recommended Posts