Taken in conjunction with Bure Budgie's post, I think it needs to be pointed out that "interesting" doesn't mean "ammunition for critics". Interest in football doesn't start and stop with Norwich City. And interest in stats doesn't start and stop with our stats. Does posting that table imply I think "18/12 or 25/15 is better than our 17/10" and want to go back to a policy I've consistently argued against? It doesn't, and I don't. The xG table is interesting because it provides insight into the whys as well as the whats of the actual table. That's why coaches are interested in xG even if so many fans apparently are not.
Point taken,but you have to play with the hand we are dealt. Moaning about injuries etc,will not change the situation. We are top for one reason only,we have more points than the other teams. DF and the team have overcome problems all season. Sit back and enjoy the ride
We didn't look very solid after Krul was replaced and let in 2 very soft goals , granted we were down to 10 men but I don't feel as confident that we will nick a goal and keep a clean sheet like most of our wins. I will be thrilled to bits if Mcgovern steps up and we carry on our miraculous run , Pukki playing through injury doesn't fill me with optimism either but we'll have to wait and see.
To be fair to McGovern neither of the goals were down to mistakes from him. The first one he couldn't do much about and the second was down to some rubbish defending by Zimmermann. Stokes come back was as much to do with Buendia getting sent off and us going down to 10 men as it was to Krul's injury. I do have some concerns about McGovern, but blame should be apportioned fairly to those that deserve it on the performance at the time and not on past performance. In my view McGovern should be given the chance to prove himself. We can always try Barden if he's not up to it.
The juxtaposition of the xG table implied an importance at least equal to the league table Robbie, which gave an impression of undermining the achievement of us being top and extending our lead over 5 trailing teams. Taken alone, especially when set out in the form of a table, xG statistics misrepresent the importance of potential to score goals. You often use xG stats to determine whether we 'deserved' to win or lose matches. I don't think that's valid when so many other factors should be considered. At the very least they should be balanced by the potential to concede goals if they are used in that way or put into the form of a table. That's why I found it of limited interest, though I do think that xG analysis does have a role when other factors are taken into account. My point about our style of play was to show the distortion that happens when xG stats are considered on their own.
I appreciate that the stattos might argue the xG and xGa measures are based on lots of data from football, but the way I see this is that, adjusted for the Championship, the measures overestimate the xG and the xGa. That feels to me like either the Championship is low scoring based on their variables, the defending is strong or the finishing is bad (or a combination). That makes me think that these stats are using the top league measures (where the finishing is stronger), with a bias towards Spain/Germany. It needs adjusting for the Championship though I think we reserve judgement until the end of the season. If it’s still hugely overestimating scoring, then it’s not right.
I think xG as a stat is interesting, but I'm not sure I agree with turning it into a table like that. Individual results are highly susceptible to variance - a missed penalty, a wonder goal, and the result on the day is very different to what xG expects. Looking at xG across the whole season shows long-term trends better. Taking the Stoke game for example, we lose that on xG 0.6-2.4, (-3 points on the xG table, we overperformed on the day), yet the xG trend for the season is actually that we've underperformed, scoring 17 goals rather than an "expected" 24 - the highest xG in the league. Which would indicate that when we've been winning by 1 goal margins, our attacking play makes it more likely that those should have been bigger margins, rather than that we've been fortunate to win. The xG table doesn't reflect when a 1-0 win should have been a 3-0 win though. The overperformance on xGA is an interesting one to interpret. We have a higher xGA value than many of our peers, and most of that overperformance comes down to three opposition penalties (two missed, one saved). So both our xGA, and the difference to our actual goals conceded, have been inflated by those. Agree with Rob that something is up if practically every club is underperforming xG. The "average striker" assumed for the calculation might be better than a Championship striker!
Some posts say it all. In fact the juxtaposition resulted simply from the table being published immediately after this week's round of matches were completed. Re. goals conceded, when A plays B, A's xG is B's xGa and vice versa. I don't understand why you think the table focusses on one rather than the other. Going back to your original post, saying "for me the actual table outweighs the statistical one" suggests the two tables are somehow competing with each other. They are not. As I keep saying, describing what is, on the one hand, and making a contribution to understanding why it is, on the other, are two different things. No one pretends that Xg/xGa give the whole story. But they prompt you to think about what the rest of the story is. At least, they prompt some of us, but apparently not Bure Budgie. He may only be interested in league position but that doesn't mean the rest of us have to be too!
@CanaryRob Here's Experimental 3-6-1's Expected Goals Table for the Championship last season: please log in to view this image So 12 teams out-performed xG over the season, 12 teams under-performed. I looked for the equivalent table for the 2018--2019 season, but it appears Ben Mayhew hadn't started producing them at that point.
You obviously love them, Robbie, but I can understand that many people dislike the 'scientification' of football and making tables like this doesn't help. Form tables are meaningful, but like Rob, I think this xG table muddles more than it informs. Used to look at individual games xG can be relevant, but even then it's subject to interpretation, as with the penalty examples Rob cites. Tables are just a step too far for me.
So you think I am the only one interested in league position? 99.9% would take league position over all the useless stats thrown down our throats after every game.
What is the fascination you have xga xgy or whatever robbie. It is the most meaningless load of gobbledygook posted on here. I would guess less than 10% of posters "read" your figures, and few of them understand, or care, about them. Maybe I am old school, and from a post you made a few weeks ago I believe in my dotage I am actually younger than you, you bore the hell out of me with this nonsense.
An Expected Goals Table is simply a tabulation of the league position of each club IF POINTS WERE AWARDED PURELY ON THE BASIS OF QUALITY OF CHANCES CREATED AS MEASURED BY EXPECTED GOALS. What is there about that that is difficult to understand? Everyone, including me and those who create xG tables, know perfectly well that: 1. Real points are awarded on the basis of goals scored/conceded, not expected goals for/against. 2. There is no simple correlation between GF/GA on the one hand, xG/xGA on the other. 3. Discrepancies between the two are explained by a whole range of what Rob calls "other factors" and you refer to as "scoring and conceding potential". The point I keep making is that the existence of these other factors is not a reason for criticising xG and xGa as such. xG and xGa make no claim to measure anything other than quality of chances created, and don't purport to offer any explanation of discrepancies between them and GF/GA. Krul's habit of saving penalties is an example of an "other factor", as also is having the best striker in the league. @DHCanary: Quote: "The xG table doesn't reflect when a 1-0 win should have been a 3-0 win though." On the contrary surely, the xG table does reflect that; it's the real table that doesn't.
No, I don't think that at all, nor did I say it. Compare "Bure Budgie is only interested in league position" and "Only Bure Budgie is interested in league position". If anyone's interested in linguistics, that's a nice example of the semantic importance of word order.
I wish I had been Duncan; I'd probably be enjoying my retirement on some tropical beach rather than in Tier 2 lockdown in Norwich!
