Damn two extra points per week!?! mixedknuts take note!

Sports bettors use data to handicap games. Simplyfying a lot, back in the day it started with just going by results/gd -> shots allowed/conceded -> everything including QoC, xG stuff, possession numbers etc. Will you be able to beat asia in the big leagues using Caley's model? No, but for the public it's probably the best predictor out there.

In for Foxes zomg

The annoying thing is when people use a predictor that is (slightly) better than other very simple metrics like past results, total shots, chances created/allowed etc... and use it as the holy truth.

Past results depend a lot on variance so people rightly try to turn towards other metrics to have better idea of team's performance and once they've found that one metric that is better but still far from being perfect and has a lot of obvious flaws then all of a sudden that metric cancels any possible variance and that metric dictates perfectly what the result should have been.

Like it's unfair criticism because a lot of people in this thread are sensible enough to acknowledge xG is an useful metric that is better or slightly better than other very simple metrics namely past results but it's also a flawed metric.
But then you'll see hundreds of posts along the line of "the xG said we should have won so we were unlucky not to win", "look at the xG we deserved to win", "Arsenal is the best team in the league just check out the xG" etc... type posts

Because a metric is maybe the best available out there (it's definitely not the case but since it's free and tweeted by some dude on twitter it's available to everybody) doesn't make it perfect. If anything it just says a lot about how far behind football is in analytics that a fairly simple metric with tons of flaw is the best we got

Payet's free kick

Tough one today.

It's more that football is resistant to metrics which is what makes it fascinating. That baseball is basically solved with stats and football isn't is to do with the nature of the games, not that football needs more work done on it.

Lets 1-0 the crap out of them!

Good post. I think the problem in football is you can't even use simple statistics to predict anything with any degree of accuracy, never mind "advanced" statistics.

xG is therefore just an exercise in turd polishing.

Are these the same sensible people who refuse to believe in any concept that is not supported by stats?

here to play panto villains today

Leicester's epic downfall starts today.


#ArsenalBPLChampions2016

Not sure if this was mentioned.

Leicester City fans are being treated to a free beer and a doughnut at the sell-out game against Southampton.

So like 2 quid times 30k people in the stadium. Big PR move that costs 60 grands for a club about to get hundreds of millions ££. Decent

Ban them all for winning silverware
People are not (Or only rarely) handing out what their model predicts. None of the xG guys do it really, the only one I can think of is Goalimpact, which is a different type of model. I think the few that openly share their models do not use them for betting or ikts. That said, there are absolutely a lot of people who do use them for those purposes that understandably don't just give away their predictions.

That said, even a tool that allowed you to more objectively evaluate past results would be useful to someone who doesn't want to base whether they were lucky/unlucky on merely hot takes.

Haha

The chairman got a big cheesy banner and happy birthday wishes on the hoarding all around the ground.

Probably a fair deal.

vardy is such a ****ing pest

Wow Mahrez just owned that guy

incoming

foxes really starting to threaten now

soton dont really look up for it which is understandable

lol everything kante touches turns to solid gold

n'golo'd

Kante and Clasie/Wanyama going to battle today

How on earth have they not scored?

Ffs

That was all skill

jesus, thank you god

lol yesssss