Premier League 20/21: Page 168 - The Arsenal discussion forum and community: One More in the Tolly

Premier League 20/21

Big Willie

goon wrote:
So Everton lose one of their games in hand, 0 points gained.

They'll have that advantage back at the weekend if they win.

Gazza M

Claudius wrote:
Yeah. How do you play this team? We have to hope that we can keep them busy as well. I think we just have to believe that our players can be confident and build good attacks, and keep possession high up the field.

I think we see Luiz this weekend. He was instrumental in getting the ball past the press and high up the field versus Leeds. So even if he did end up losing some long balls, we could press’s upfield. What we cannot afford is to be chasing shadows around our penalty box because we will wear our mentally and physically

aggressively get behind their fullbacks with direct balls and direct running. that's been the only way to beat peak pep teams down the years.

don't even attempt to try and outpass them or tiki taka through the middle. they will catch you out at some point, and punish you

goon

Big Willie wrote:
goon wrote:
So Everton lose one of their games in hand, 0 points gained.

They'll have that advantage back at the weekend if they win.

They're at Anfield. We'll draw level with them by the end of the weekend.

In all seriousness, they play Liverpool, West ham play Tottenham and Villa play Leicester. It wouldn't be the worst weekend of us to lose or draw but an amazing one to win.

Big Willie

True. Didn't realise they all had such hard games this weekend.

Sicario

Sounds like a perfect weekend for Chelsea ffs

TambourineMan

And United.

banduan

the advantage for us is that Citeh will have an eye on CL.

goon

The good thing about this is that it suggests, as I expect, that the likes of West Ham, Everton and Villa should drop away. Tottenham have been even shitter than their points tally suggests too. The bad news is it also suggests Chelsea are more likely to pull away further to get the last CL spot.

LOL at Fulham and Brighton.

Claudius

Even by that measure we are still only the 9th best team in the league.
Spurs need to get their attack fixed soon. Hope we can pass them

Klaus

Brighton on 25 points with an expected 39 is a bit incredible and says something about their wastefulness in front of goal. If they find a good striker in the summer they could be the big surprise team next season.

Qwiss

Who'd have thought signing Danny Welbeck wasn't the answer to their striker issues.

Claudius

Qwiss! wrote:
Who'd have thought signing Danny Welbeck wasn't the answer to their striker issues.

😆

Claudius

I always find this expected points such a weird measure. My understanding is what they do is they take each game’s expected goals statistics, run a Monte Carlo simulation and award the most likely results based on the simulations. But what you effectively end up with is a discrete xP for each game based on a highly variable measure. And then you add all of these up and end up with a league table. So our simulation vs Leeds would’ve taken the mean of 2,5 xGA vs 1,1 xGA.

I know the Monte Carlo helps - I’ve used them for work, but it’s helpful to know that you are taking 24 very volatile data points and adding them up.

banduan

I thought it was a straightforward xG comparison? i.e. you have higher xG by a certain amount you win.

QuincyAbeyie

I thought so as well.

goon

The description under the chart seems to suggest so.

Either way it's probably quite crude but I do actually think it confirms what I've seen with my eyes, especially in relation to the likes of Fulham, Brighton, Everton, Spurs, United and Leicester.

Claudius

These guys typically run a simulation. Taking a straight xG is even worse because of the inherent variance in the measure. We outsourced our xG by 50% in the Leeds game (2,4 vs 4 actual goals). Using clean xG on an individual game basis is deadly

Big Willie

Villa are underperforming in terms of expected points Vs actual points.

flobaba

The expected points tally does not take into consideration the effects of game to game momentum and confidence (and inertia and the lack of confidence).

Those intangibles are also contributors

banduan

Claudius wrote:
These guys typically run a simulation. Taking a straight xG is even worse because of the inherent variance in the measure. We outsourced our xG by 50% in the Leeds game (2,4 vs 4 actual goals). Using clean xG on an individual game basis is deadly

But that's what the description implies?

Arsenal v Leeds was Arsenal xG 2.5 vs Leeds xG 0.77.

So this could either just be pure rounding up to 3-1
or it could be some other formula to arrive at an expected scoreline based on that xG.
Regardless a result is drawn from this and it would end up with Arsenal getting 3 points. Then just repeat for every other match.

I don't think any simulation is necessary here for what they're attempting to do?

« Previous Page Next Page »