Didn't mean for the call to prove anything, I just like watching it. I'm pretty sure I've heard him say he's an Arsenal fan, but I don't have any link to it in print.
1044600419148877825 is not a valid tweet id Interesting datapoint here: Arsenal's high press has gotten worse under Emery (Arsenal were actually good at winning the ball via the press over the last few seasons). Well, that or Arsenal are pressing less in order to defend against the counter, but that hasn't worked at all.
That seems like a pretty big statistical oversite doesn't it? Or is the assumption of individual talent built into the stat. A bad pass, that needs a elite finish, technically falls within the realm of expected outcome. If the only measure is player and ball GPS location. There are times I can specifically recall not making a pass because the strikers body position was off, or his angle was wrong. I'm curious how that is built into the stat, if it is at all.
Are you talking about XG? It is based on a league average striker from that location The whole point is to isolate the value of the shooting locations
Thinking about it, because locations is a key component of striker quality, over a season the quality of your forwards will be reflected in the XG for
I remember looking in years past and finding nothing. But just now I googled and this was the only hit I read: https://www.mirror.co.uk/sport/football/news/exclusive-martin-tyler-opens-up-11835667 Near the bottom he reveals his club... Woking! "CAN YOU BELIEVE THAT?!!"
I seem to notice he is so much more into a match on Sky's domestic telecasts for the UK as opposed to the international feeds, in which we get here in the States, here is the Martin Tyler that is SO more into it on Sky:
Interesting Cech repeated the same stories about Arsenal not working hard enough under Wumger One upside now is with the likes of Mhki, Laca, etc floating about - there is genuine competition for places
The other problem with XG is its accuracy. For example, last season it underpredicted the league goals total by about 30 goals i think.
Less than a goal every match week doesn't seems within reason imo. But I'm more interested in its variance when measuring a "big chance".
The 'big chance' stat is purely subjective, though. That's just OPTA trackers making a call on a case by case basis. xG is a completely different animal. Now, some guys with xG models may refer to a specific shot as a 'big chance' based on how their model rates it, but it's not the 'official' stat.
I guess it depends on how you use the data. The year prior it underestimated by 80 goals. Haven’t looked at other leagues/seasons yet. Also half of the teams were off by more than 10 percent with Chelsea a whopping 23 goals off (27%). Could possibly be just a variation that will regress to the mean over time or could also mean the models are systemically missing something. I suspect the latter. Not to say it isn’t useful at identifying things, but just don’t put too much into it—yet.