HomeZ OLD CATEGORIESArsenal (NN)EPL Prediction Model: Week 12 Forecasts | Arsenal, Liverpool to win?

EPL Prediction Model: Week 12 Forecasts | Arsenal, Liverpool to win?

First things first, and the model would like to apologise for last time out… If it wasn’t for Charlie Adam’s penalty, that would have been comfortably the worst week it’s ever had, including running diagnostics right across the past two seasons. The natural volatility of Premier League football results means you’ll get one of those weekends now and again but I still don’t have to like them.

Premier League Predictions Week 12

To my eye the percentages weren’t all that bad and a lot of the model’s deeper game stats were realistic – possession and shot numbers from the simulated games looked pretty good. Just not goal scoring. The odds on Spurs, Man City, Everton and Chelsea all failing to win must have been huge!

Thankfully the first week’s win meant we weren’t completely wiped out for betting and now with the international break putting those results an extra week into the past, let’s never speak of it again.

I’ve been thinking over the past few weeks that sharing only percentage forecasts for a week’s games is a little unsatisfying. I see a lot more detail from the inner workings of the model, but if you’re reading these posts, there’s no hint of whether a prediction is likely to be a good one. It’s just an output number and you can take it or leave it.

To try to counter this a little, for each game I’m going to post some simulated possession and shooting stats, not just the win likelihood. Eventually we might get to the individual player performance predictions, but one step at a time. Sharing this data means for a prediction like Fulham to draw with Swansea (see below) you can make your own judgement on whether my model is behaving sensibly. The model is splitting possession in that game relatively evenly. Is that unrealistic? Now you’ve got a more concrete way to discount the predictions you don’t agree with.

Here are the deeper simulation stats for this weekend’s games…

sim stats 23-11

And those lead to the following percentage chances…

23-11

If you’re betting, there are some really marginal calls on which games could be draws, but I’ll be going for…

Everton v Liverpool – Away win
Arsenal v Southampton – Home win
Fulham v Swansea City – Draw
Hull City v Crystal Palace – Draw
Newcastle United v Norwich City – Home win
Stoke City v Sunderland – Ridiculously close. Home win.
West Ham United v Chelsea – Draw
Manchester City v Tottenham – Home win
Cardiff City v Manchester United – Away win
West Bromwich Albion v Aston Villa – Draw

datamonkey
datamonkeyhttp://www.wallpaperingfog.co.uk
Statistical analyst and econometrician, working for a large marketing agency. Football stats are much more interesting.
More News

LEAVE A REPLY

Please enter your comment!
Please enter your name here