xgEdge

Premier League results 2026/2027

Every Premier League match played this season, newest first, with the xG each side created and the home, draw and away probabilities the model held before kickoff.

Model record

Most likely outcome landed in 33 of 80 matches (41%). The probabilities implied about 35 (44%). Brier score 0.63 (reference 0.66; lower is better). How the model works →

  • Home
  • Draw
  • Away
What happened Most likely outcome Upset

Sunday, 08/11/2026

5 matches
  • xG 3.55 – 2.88 Draw with an xG gap
    Pre-match probabilities
    →
  • xG 2.38 – 0.66 Followed the xG
    Pre-match probabilities
    The outcome the model rated most likely landed
    →
  • xG 1.26 – 2.23 Followed the xG
    Pre-match probabilities
    The outcome the model rated most likely landed
    →
  • xG 2.60 – 2.65 Level on xG
    Pre-match probabilities
    The outcome the model rated most likely landed
    →
  • xG 2.11 – 0.59 Followed the xG
    Pre-match probabilities
    The outcome the model rated most likely landed
    →

Saturday, 07/11/2026

4 matches
  • xG 1.06 – 1.76 Lower-xG side won
    Pre-match probabilities
    →
  • xG 0.70 – 1.14 Level on xG
    Pre-match probabilities
    →
  • xG 1.26 – 0.36 Followed the xG
    Pre-match probabilities
    The outcome the model rated most likely landed
    →
  • xG 1.96 – 0.72 Followed the xG
    Pre-match probabilities
    The outcome the model rated least likely landed
    →

Friday, 06/11/2026

1 match
  • xG 1.43 – 1.21 Level on xG
    Pre-match probabilities
    The outcome the model rated most likely landed
    →

Monday, 02/11/2026

1 match

Sunday, 01/11/2026

1 match
  • xG 1.05 – 0.96 Level on xG
    Pre-match probabilities
    The outcome the model rated most likely landed
    →

Saturday, 31/10/2026

8 matches
  • xG 1.83 – 2.24 Level on xG
    Pre-match probabilities
    The outcome the model rated least likely landed
    →
  • xG 3.47 – 0.14 Followed the xG
    Pre-match probabilities
    The outcome the model rated most likely landed
    →
  • xG 1.64 – 1.00 Lower-xG side won
    Pre-match probabilities
    →
  • xG 0.78 – 1.06 Level on xG
    Pre-match probabilities
    →
  • xG 0.69 – 1.82 Lower-xG side won
    Pre-match probabilities
    →
  • xG 0.46 – 0.86 Level on xG
    Pre-match probabilities
    The outcome the model rated most likely landed
    →
  • xG 1.29 – 0.71 Followed the xG
    Pre-match probabilities
    The outcome the model rated most likely landed
    →
  • xG 3.00 – 1.24 Followed the xG
    Pre-match probabilities
    →

Sunday, 25/10/2026

5 matches
  • xG 0.83 – 0.76 Level on xG
    Pre-match probabilities
    The outcome the model rated least likely landed
    →
  • xG 1.17 – 2.15 Followed the xG
    Pre-match probabilities
    →
  • xG 1.20 – 1.32 Level on xG
    Pre-match probabilities
    The outcome the model rated most likely landed
    →
  • xG 1.57 – 1.74 Level on xG
    Pre-match probabilities
    The outcome the model rated least likely landed
    →
  • xG 1.59 – 0.95 Lower-xG side won
    Pre-match probabilities
    →

Saturday, 24/10/2026

4 matches
  • xG 0.47 – 0.74 Level on xG
    Pre-match probabilities
    The outcome the model rated least likely landed
    →
  • xG 0.51 – 0.55 Level on xG
    Pre-match probabilities
    The outcome the model rated most likely landed
    →
  • xG 1.96 – 0.63 Followed the xG
    Pre-match probabilities
    The outcome the model rated most likely landed
    →
  • xG 0.61 – 1.56 Lower-xG side won
    Pre-match probabilities
    →

Friday, 23/10/2026

1 match
  • xG 0.46 – 2.29 Draw with an xG gap
    Pre-match probabilities
    The outcome the model rated least likely landed
    →

Monday, 19/10/2026

1 match
  • xG 0.86 – 0.77 Level on xG
    Pre-match probabilities
    The outcome the model rated most likely landed
    →

Sunday, 18/10/2026

4 matches
  • xG 0.90 – 1.14 Level on xG
    Pre-match probabilities
    The outcome the model rated most likely landed
    →
  • xG 1.41 – 0.92 Level on xG
    Pre-match probabilities
    The outcome the model rated least likely landed
    →
  • xG 2.00 – 0.77 Followed the xG
    Pre-match probabilities
    The outcome the model rated most likely landed
    →
  • xG 4.23 – 0.07 Followed the xG
    Pre-match probabilities
    The outcome the model rated most likely landed
    →

Saturday, 17/10/2026

5 matches
  • xG 0.99 – 1.57 Followed the xG
    Pre-match probabilities
    →
  • xG 3.53 – 1.66 Followed the xG
    Pre-match probabilities
    The outcome the model rated most likely landed
    →
  • xG 1.38 – 0.16 Followed the xG
    Pre-match probabilities
    The outcome the model rated most likely landed
    →
  • xG 0.88 – 1.64 Followed the xG
    Pre-match probabilities
    →
  • xG 0.43 – 1.47 Followed the xG
    Pre-match probabilities
    →

Monday, 12/10/2026

1 match
  • xG 1.47 – 1.84 Level on xG
    Pre-match probabilities
    The outcome the model rated most likely landed
    →

Sunday, 11/10/2026

3 matches
  • xG 0.94 – 1.28 Level on xG
    Pre-match probabilities
    →
  • xG 1.33 – 1.68 Level on xG
    Pre-match probabilities
    The outcome the model rated least likely landed
    →
  • xG 2.41 – 3.34 Lower-xG side won
    Pre-match probabilities
    The outcome the model rated most likely landed
    →

Saturday, 10/10/2026

6 matches
  • xG 0.80 – 1.62 Draw with an xG gap
    Pre-match probabilities
    →
  • xG 0.72 – 1.41 Followed the xG
    Pre-match probabilities
    →
  • xG 1.15 – 1.81 Lower-xG side won
    Pre-match probabilities
    The outcome the model rated most likely landed
    →
  • xG 1.10 – 1.38 Level on xG
    Pre-match probabilities
    The outcome the model rated most likely landed
    →
  • xG 2.28 – 1.06 Draw with an xG gap
    Pre-match probabilities
    The outcome the model rated least likely landed
    →
  • xG 2.33 – 0.62 Followed the xG
    Pre-match probabilities
    The outcome the model rated most likely landed
    →

xG is a model estimate of the chances each side created. "Went with the xG" means the side with more xG won; "Went against the xG" means the side with less xG won or the match was drawn across a gap of at least 0.5. The probabilities are the ones the model published before kickoff, rounded so the three add to 100; the outline marks what happened.