I think Wissa would fit in well. He is a super finisher and has great movement that would suit our game. But, given his age, I would not want to pay too much and I think Brentford are unlikely to let him go.
Isakly what we need
- Thread starter Hansern
- Start date
- Latest activity Latest activity:
So one goal from open play in last 20 shots. Some were shit shots and several were very good saves. And, that is sometimes the problem with xG. Although, one could argue that 20 shots is not a big enough dataset to account for goalkeepers saving shots that maybe they should not.
Shame they want 50 mill for him
Alternatively, he tears a groin in pre-season and Newcastle fans have to suffer a toothless front line for 3 months whilst Ekitike hits the ground running in style…
He will have a back injury & miss the fixture - he doesn't want to jeopardise the proposed swap deal with Ektititktktk next summer
I was going to ask exactly this, then saw you had, and then realised @MomoWASright hadn't responded.
@MomoWASright, I think you may be misunderstanding the metric.
The metric should be he scores 19/20. To say he should only be scoring 16 of those is kind of arbitrary. He’s just fucking good at them based on real data. In a hypothetical game, it’s drab, zero chances for either side, Milner gets a pen. What is the xG? 0.8 - 0.00 or 0.95 - 0.00? Looking at the real data, it’s 0.95.
xG utilises, I think, 5 metrics. Where on the pitch the ball is, distance, angle (could really be rolled into the first), shot or header and defensive set up. Like I said in the post, there are way too many variables. Like with the pen, is the chances of scoring a pen 80% in the 84th minute whether 7-0 up or at 0-0 in the 38th game of the season to win you the league? Is it away? Which supporters are you shooting towards? What are the weather and pitch conditions? It applies the same scoring when all things are not equal.
xG tries to over simplify what is a technical thing.
So, you're suggesting that it actually needs to take in more metrics to be more accurate? Maybe it does, but at least it's not as simple as all chances are equal. There's probably trade offs with adding in things that are more subjective.
But, it absolutely needs to be taken into consideration with other things, it shouldnt be used in isolation, it's just a better metric than chance conversion as it attempts to put in the difficulty.
This I agree with. It shouldn’t be the be all and end all. Over the years it has been given too much prominence. Saying how did city lose 2-1 when they won xG 4.85 to 0.55 when if you were to look at the actual game you’d get better answers through other metrics.
Yeah, but you could be pretty sure that they were pretty shit in front of goal, it wasn't a lack of creating good chances, it was not putting them away. Maybe something else too.
All that stuff should be washed out if you have a big enough dataset.
I don't really understand your stance on the Milner thing. An average pen taker would score 80% based on the stats. Milner being way better than average has no bearing on that, surely?
If every shot in each zone of the pitch was roughly equal then yes, but there are too many variables including a shit load of luck.
Presenting a Milner penalty as 80% is wrong because we have a better source of data to help make a more complete reflection on what happened.
I'm 100% in that boat. Really really hope I'm a terrible judge of potential.
Yes I think that was obvious from his 'Last 60 Shots' YT video that put the shits up many of us! Something the coaching team can work on though.
Oh I see what you're getting at - ie that it'd be wrong to say we had 0.8xG when really it was 0.95 in our particular case, etc?
I think that sort of shortcoming is just inherent in the metric isn't it? It's not saying these chances will convert to x goals for this team, but as an average across all teams. It's not really there to make an actual prediction - it's not meant to be intelligent in that way. It's basically just a chance creation measurement. And then how tightly chances created will correlate with goals will vary by team based on different factors which xG isn't designed to assess.
Except in the case when Arsenal xG is better than PSG - then it clearly and, indisputably, demonstrates that Arsenal are the best team in Europe, but conversely, second only to Chelsea in the world.
