The Official 2018/19 Premier League Thread

[jts1882]

How can analysis be serious when it doesn't predict draws? There are two ways to do it, which could give different results, but this would even out over a season.

1. Round up or down the goals for each team and that is the result. E.g. Bournemouth 2 -1 Cardiff, Burnley 1-1 Southampton,

2. Round up or down the margin of victory, so a difference of at least 0.5 is needed for a win. This would make Bournemouth-Cardiff a draw (+0.42) and give Burnley a win (+0.53).

 

[Lemonade Money]

https://twitter.com/x/status/1028930401761472512

That looks like they are only doing wins and losses. Surely they need to round off the goals for/against and have draws too?

 

[Baleforce]

That looks like they are only doing wins and losses. Surely they need to round off the goals for/against and have draws too?

I think it’s better if they don’t.

 

[Lemonade Money]

I think it’s better if they don’t.

Why's that?

 

[Baleforce]

Why's that?

Because the difference is still relevant, no matter how marginal. This table is solely to weight the chances, the actual table is normalised.

 

[Lemonade Money]

Because the difference is still relevant, no matter how marginal. This table is solely to weight the chances, the actual table is normalised.

So if a team has got 0.1 better xG, i.e. basically the same for both teams, they should have won the game?

 

[Baleforce]

So if a team has got 0.1 better xG, i.e. basically the same for both teams, they should have won the game?

No, this isn’t about winning the game, it’s about who had the better xG.

 

[Lemonade Money]

No, this isn’t about winning the game, it’s about who had the better xG.

Isn't xG about who should have won the game?

 

[Baleforce]

Isn't xG about who should have won the game?

I think it’s about quality of chances, additional context with which to grade performances.

 

[nayimfromthehalfwayline]

Isn't xG about who should have won the game?

Its about the quality of chances created.

So high XG means a high quality of chance and a higher chance of goals.

Its useful when looking at teams like Utd, continually winning despite low XG is suggestive they are riding their luck rather than "deservedly" winning.

Just as, teams on a bad run with high XG, you could argue its more a blip as they are clearly creating good chances.

As with any stat its only really indicative of one thing, and is in danger of being used as an absolute when it really shouldnt be.

 

[ricky2tricky4city]

Of course they'll win the league, but people on here are all to ready to point out the short comings in our performance on Saturday but let City's sloppy display pass.

Stop conflating the two. You said they were rubbish....I gave you the field.

 

[K.D.D.D.D.Soc]

Stop conflating the two. You said they were rubbish....I gave you the field.

Ok Their keeper played at least 4 suicide passes to defenders in or just outside the box, they gave away possession more than 10 times in their own half of the field, they failed with their final pass to attackers to often and one of the best strikers in the league looked average against two centre backs wearing divers boots add Mendy getting away with the most obvious penalty of the weekend, I would say that was a rubbish performance, but to be critical of City or their coach is forbidden, I accept it wasnt their best XI but they're better than most. They scored a very good goal plus one I would be disappointed to see us concede.

 

[jts1882]

No, this isn’t about winning the game, it’s about who had the better xG.

Then why award points at all? Just use aggregate xG difference over the season.

Surely the point is to ask if the actual results (points awarded in the game) are matched by predictions based on chances (points based on xG differential).

 

[parklane1]

Then why award points at all? Just use aggregate xG difference over the season.

Surely the point is to ask if the actual results (points awarded in the game) are matched by predictions based on chances (points based on xG differential).

Some one has to find a use for all these guys with their different stats, there are so many who have to find a use for what they know. ( allegedly)

 

[Little Luka]

Kevin De Bruyne out for 3 months apparently. Injured in training.

 

[markysimmo]

https://www.bbc.co.uk/news/uk-england-bristol-45194443

 

[Bullet]

https://www.bbc.co.uk/news/uk-england-bristol-45194443

That's not Premier League, pure lower league talent

 

[afsoc4life]

I wonder what impact KdB's injury would have on City's chances of winning the league. They still have Silva, Gundogan, plus a host of wingers. Perhaps an opportunity for Foden to step up?

 

[Daisuk]

https://twitter.com/x/status/1030144658020548608

 

[Jordinho]

https://twitter.com/x/status/1030199932760862720

One of the worst PL referees ever, but hope it's because he wants to do something else and not because of illness or anything.

Probably best to not discuss or mention the rumour going round on twitter.