After outclassing Mourinho Utd, Brighton make the long journey to Anfield. (Unfortunately, the model doesn’t predict that players won’t put in a shift when they get on the pitch). Despite that win, the model gives Brighton slim hopes, with Liverpool fancied to win 85% of the time, and a most likely score of 4-1.
If you are interested in the games which truly matter, look no further than the John Smith’s Stadium on Saturday. In what you feel is already a must-win game if they are to survive the drop, Huddersfield welcome Cardiff City. The model forecasts a close encounter, with a most likely score of 1-0 to The Terriers.
Pressure mounts fast on managers in the Premier League. Unai Emery should not expect the patience given to his predecessor at Arsenal. With West Ham’s poor away and defensive record of late however, he should expect some relief at the Emirates tomorrow. If you listen to phone-ins such as BBC 5Live’s 606, then you know Arsenal fans are especially moany (only rivaled by Liverpool fans). If Arsenal don’t win as predicted (82%), then Savage and company will certainly hear how (for the past 15 years) “Arsenal need a leader”.
The rest of our Premier League forecasts are in the the table below:

  • The Forecast scoreline, with the % chance of that happening
  • The % chance of a win by the Home team, P(H), or Away team, P(A) (with one hundred minus those two numbers giving the % chance of a draw
Forecast Win (%)
Score Pr (%) P(H) P(A)
Liverpool Brighton 4-1 7 85 5
Fulham Burnley 0-1 21 20 46
Huddersfield Cardiff 1-0 19 43 23
Bournemouth Everton 2-1 9 44 33
Southampton Leicester 1-2 9 32 44
Wolves Manchester City 0-1 19 11 64
Arsenal West Ham 3-0 9 82 7
Watford Crystal Palace 1-0 11 46 28
Newcastle Chelsea 1-2 10 26 51
Manchester Utd Spurs 1-0 11 57 20

As discussed before (here, here & here), we continue to tinker with our forecasting method. Our aim has always been to provide a single scoreline forecast to hang our hats on. But it is challenging to balance both plausible result and score forecasts simultaneously, which will perform well against some metric. Therefore, we are now making “Fuzzy Conditional” forecasts, discussed here.