Premiership Game Score Predictions

Sorting the Score Predictions for EPL GW 23 2018

After a hectic festive season, it was nice to have a bit of a respite from FPL, although all EPL teams were in action as FA Cup matches were scheduled.

The results are 10 teams through to the 4th round, 5 teams with replays January 16/17 (Bournemouth, Chelsea, Leicester, Swansea & West Ham), and 5 teams out (Arsenal, Burnley, Crystal Palace, Everton & Stoke).

There were also Carabao Cup matches last night and tonight between Manchester City, Chelsea, and Arsenal. There are bound to be rotations made in these matches. So be on alert for changes in lineups, especially from Chelsea and Arsenal.

Before I dive into the GW 23 data pool let’s review the RMT algorithm results from GW 22. The  Clean Sheet predictions were 3/11, Winning teams predictions at 7/11, and teams scoring 2 or more goals at 8/11. Only the CS predictions appear to have been off the mark. 2 out of 3 ain’t bad, as the song goes.

The RMT algorithm Simulations for Clean Sheets in GW 23 are – 

Home Away Sim% for H CS Sim% for A CS
Crystal Palace Burnley 43.6 28.6
Chelsea Leicester City 42.4 5.6
West Bromwich Brighton & Hove 39.6 31.2
Newcastle United Swansea City 39.0 17.3
Manchester United Stoke City 36.9 1.7
Tottenham Everton 30.2 5.8
Huddersfield Town West Ham United 20.1 22.6
Watford Southampton 22.3 18.7
Liverpool Manchester City 14.9 21.5
Bournemouth Arsenal 10.0 17.8

The teams highlighted in bold are predicted to keep CS’s. As can be seen, the simulations generated for teams in each match can vary. The higher the percentage of simulations related to a team, the better the chance of a Clean Sheet.

When the percentage difference between 2 teams is large that may indicate that the team with the higher value is likely to keep a CS.

If this were the case a CS for Chelsea, Manchester United, Tottenham and possibly Newcastle, could be forecast. The remaining matches’ simulations are grouped close enough together that it is questionable as to whether either team would register a CS.

When the website predictions are taken into account the consensus would seem to support that Chelsea, Manchester United, Tottenham and Newcastle are favored to keep CS’s. (Refer to the Table below on Statistical Analysis of Score Predictions).

If this is the case, GK and defender assets from the 4 teams mentioned could be of value this GW.

Home and away form of Palace v Leicester may be open for a 0-0 punt, as neither has scored or conceded many goals in the last 6 matches.

The RMT  algorithm Simulations for Winning Teams in GW 23 are – 

Home Away Sim% for HW/AW/Draw Sim% & Projected Score
Chelsea Leicester City 61.2 / 8.3 / 18.0 16.0 / 2-0
Manchester United Stoke City 57.1 / 2.3 / 9.5 13.8 / 2-0
Newcastle United Swansea City 50.6 / 19.9 / 25.5 18.6 / 1-0
Tottenham Everton 50.1 / 11.3 / 21.3 12.2 / 1-1
Crystal Palace Burnley 43.9 / 24.6 / 29.7 19.6 / 1-0
Bournemouth Arsenal 23.3 / 39.0 / 24.9 13.8 / 1-1
Liverpool Manchester City 25.4 / 38.4 / 28.1 16.2 / 1-1
West Bromwich Brighton & Hove 37.8 / 28.4 / 32.5 19.2 / 1-1
Huddersfield Town West Ham United 30.8 / 36.0 / 26.8 15.4 / 1-1
Watford Southamptom 35.8 / 31.9 / 26.5 15.7 / 1-1

Again, the teams in bold are predicted to win. The simulations in 6 of the matches seem to go against the grain with 1-1 draws being predicted.

As shown, 4 of the 5 teams at the top of the Table are predicted to win. The only questionable result is the 1-1 between Spurs & Everton. If the final score simulations for this match were to be lowered to the 10.2% level this would predict a final score of 2-0 or 2-1, rather than the 1-1 score predicted by 12.2% of simulations. The choice of a winning result, or a CS, comes down to individual instincts.

The bottom 5 matches have simulations that are tightly grouped, and open to either team winning or to a draw. But in my opinion, the Liverpool v Manchester City match could turn out to be a goal fest.

Bournemouth v Arsenal is predicted as a draw, but the website predictors tend to favor an Arsenal win. “You pays yer money, and you takes yer choice”.

The RMT algorithm Simulations for Teams to Score 2 or More Goals in GW 23 are –

Home Away % Sim for H 2+ G % Sim for A 2+ G
Chelsea Leicester City 51.6 11.6
Manchester United Stoke City 51.3 6.8
Tottenham Everton 49.2 20.2
Bournemouth Arsenal 33.8 45.8
Liverpool Manchester City 31.1 41.0
Newcastle United Swansea City 36.9 17.7
Watford Southampton 35.7 33.5
Huddersfield Town West Ham United 31.4 35.6
Crystal Palace Burnley 27.8 15.8
West Bromwich Brighton & Hove 22.8 18.0

Again, the teams in bold are predicted to score 2 or more goals in this week’s matches.

But based on the simulation percentages of final scores, Chelsea and Manchester United are the only teams predicted to score at least 2 goals.

The remaining teams listed in bold are predicted to score 2 or more goals only in 10% – 11% of the final score simulations otherwise, the simulations that are shown in the above Winning Teams Table predict these matches to end in 1-1 draws.

If you were to look for a Captain choice the best place would be to scan the assets of Chelsea or Manchester United. For the punters in the crowd, Liverpool hosting Manchester City could turn into a goal fest, as mentioned previously.

And if the Gunners could score on the road they may surprise and put 2 on the board given that Bournemouth has conceded 1.8 goals/gm in the last 6 at home.

Finally, the 1-1 score predicted for Tottenham v Everton is surprising. In my humble opinion, the Spurs have the capacity to put 2 past the Everton backline. Their attacking potential is present in midfield and at striker.

The Statistical Score Predictions for GW 23 are –

Home v Away Mean Mode Stats. Prediction
Chelsea v Leicester City 2.3-0.6 2-0 2-1 or 2-0
Crystal Palace v Burnley 1.2-0.7 1-1 1-1
Huddersfield Town v West Ham United 0.9-0.7 1-1 1-1
Newcastle United v Swansea City 1.3-0.3 1-0 1-0
Watford v Southampton 1.2-0.9 1-1 1-1
West Bromwich v Brighton & Hove 0.6-0.6 0-0 1-1 or 0-0
Tottenham v Everton 2.0-0.1 2-0 2-0
Bournemouth v Arsenal 0.7-1.8 1-2 1-2
Liverpool v Manchester City 1.4-1.4 2-1 1-1 or 2-1
Manchester United v Stoke City 2.5-0.3 3-0 3-0

Chelsea, Newcastle, Tottenham, Arsenal and Manchester United are predicted to win by a majority of the websites reviewed. Manchester City’s statistical analysis yields a choice between a draw and a City win.

The statistical predictions compare well with the simulations generated by the RMT algorithm as to wins. Three of the remaining 4 matches are predicted as draws, with only the Palace v Burnley result differing between a statistical draw and an RMT algorithm Palace win.

Overall, the simulation results are quite similar to those generated by statistical analysis. This may imply that great mind think alike, lol.

The Top Sites from GW 21 & GW 22 were –

Site # of Predictions # of Correct Wins/Draws # Correct Exact Scores – merson 20 11 2 20 10 3 20 9 4 20 9 1 20 9 1 20 9 0 20 8 5 20 8 2 20 8 1
RMT algorithm 20 8 1 – LAWRO 20 8 1

This Table has the combined results fro the past 2 GW’s that weren’t included in the previous article due to time constraints of the overlapping matches. The RMT algorithm was among the leaders during the match-packed schedule.

Hopefully, the schedule will be a little less congested over the next few weeks.


  • Bio
  • Latest posts


79 yr old retired water & wastewater Chemist from Philadelphia, PA; retired to Florida. Enjoying the game more through association with all on site.
Continuing my FPL educations with the help of quite a few friends.

5 thoughts on “Sorting the Score Predictions for EPL GW 23 2018

  1. Thanks Mosc.

    Glad the hectic festive scheduling is out of the way.
    We can now get back to more settled starting eleven’s and hopefully higher prediction/results accuracy :thanks:

Comments are closed.