The five games that were played on Tuesday and Wednesday this week really didn’t go well for the model as its upset predictions proved somewhat inaccurate. I’ll be hoping that it bounces back this weekend but we do have a few unexpected calls for you once again.
It was surprising to see Manchester United picked by the model to beat Manchester City – even playing at home – and in the end this one wasn’t close. The model’s quite reliant on goals scored to assess strikers and so underestimated Dzeko’s chances as he went on to score twice. Everton also comfortably beat Newcastle, running against the model’s prediction.
We got Liverpool’s home win and should have got Arsenal too, but for a very late own goal. The last game in mid-week was a prediction of Hull to win away at West Ham and the model actually did OK here. An early sending off and an own goal took the game away from Hull, but they still had good chances and West Ham were booed off despite winning.
Keep that mid-week Manchester City call in mind as you look at this week’s percentages…
It’s tight, but the model thinks Arsenal could win at home. We’ve also got a high enough draw percentage to be worth backing Crystal Palace to get a point against Chelsea. Manchester United’s game with Aston Villa is very close, but Villa have the strongest percentage.
Any one of those three would go a long way to repairing the damage from Tuesday night.
As usual, to go with the percentages, here are the more detailed possession and shooting stats:
Finally, the model’s predictions for this week:
Manchester United v Aston Villa – Away win
Crystal Palace v Chelsea – Draw
Southampton v Newcastle United – Home win
Stoke City v Hull City – Home win
Swansea City v Norwich City – Home win
West Bromwich Albion v Cardiff City – Home win
Arsenal v Manchester City – Home win
Fulham v Everton – Draw
Liverpool v Tottenham Hotspur – Home win
Sunderland v West Ham United – Draw