Last week saw some satisfying predictions on Sunday, as Everton ran out comfortable winners at home to Arsenal – in line with the model’s forecasts – and Liverpool won away at West Ham. Saturday was close to a great result, with the three clear favourites winning, but Hull vs. Swansea and Norwich vs. West Brom both finished 1-0 and scuppered our calls of draws.
This week, we’ve got nine games up to Tuesday night and then another couple on Wednesday. There are predictions to Tuesday in this post and I’ll publish Wednesday’s after Everton, Palace, Man City and Sunderland have played at the weekend, in case of any injuries.
Here are this week’s percentages:
Obviously you jumped to Sunday’s big game first and without a doubt, the model’s gone slightly wild! The scoreline feels very unlikely but is sometimes a feature of this type of simulation model. It’s happened because Liverpool have scored freely this season at home – including against good teams – so that free scoring doesn’t get moderated downwards, even against Manchester City.
It’s worth mentioning that the model doesn’t do ‘form’ (yet) and so uses stats for the whole season, including Man City’s less impressive early away games. The bookies are also predicting a Liverpool win though and we’ll be backing Rodgers’ men this week.
Here are the usual deeper possession and shooting stats:
And finally, the model’s predictions:
Crystal Palace v Aston Villa – Draw
Fulham v Norwich City – Home win
Southampton v Cardiff City – Home win
Stoke City v Newcastle United – Home win
Sunderland v Everton – Draw
West Bromwich Albion v Tottenham Hotspur – Away win
Liverpool v Manchester City – Home win
Swansea City v Chelsea – Away win
Arsenal v West Ham United – Home win (but don’t be surprised by a draw)