The table says level, the profiles say chaos
Everton and Chelsea arrive at Hill Dickinson Stadium on Oct 17 for a Premier League round 7 match separated only by table position. The current table has both on 10 points from six games: Everton fifth, Chelsea sixth. Everton are unbeaten with two wins and four draws, scoring eight and conceding five. Chelsea have three wins, one draw and two defeats, with 15 goals for and 13 against. That is the headline tension: one side has not lost, the other has not drawn much but has played wilder matches. The kickoff is 11:30 UTC, and the profiles could hardly be more different.
Everton's floor and ceiling
Everton's current table home record is strong: three played, two wins, one draw, no defeats, five goals for and two against. Their stored recent results show the same pattern. They drew 2-2 at Hull City on Oct 11, beat Ipswich 1-0 at home on Sep 19, drew 0-0 at Tottenham on Sep 12, drew 2-2 with Manchester United at home on Sep 6, and drew 1-1 at Bournemouth on Aug 29. One supplied signal puts their stored eight-match points-per-game at 1.00, and their Premier League sample at 1.20. They are hard to beat, but the draws are piling up. A key signal also has their home scoring average at 1.80 goals per game, so they carry a real threat at Hill Dickinson.
Chelsea's away problem
Chelsea's overall numbers are more explosive. In the current table they have 15 goals in six games, but 13 conceded. The away split is the obvious weakness: three played, one win, no draws, two defeats, four goals for and seven against. Their stored recent results include a 5-1 home win over Bournemouth on Oct 10, a 0-3 away defeat at Brentford on Sep 18, a 2-2 home draw with Hull City on Sep 12, a 1-2 away defeat at Arsenal on Sep 6, and a 4-3 home win over Brighton on Aug 30. One supplied signal puts their stored eight-match points-per-game at 1.38, and their Premier League sample at 1.40. The overall profile is better than Everton's, but the road version is leaky: a key signal has Chelsea conceding 2.00 goals per away game.
The AI read and the timestamp
Nutmegly's AI read leans away. It gives Chelsea a 47.1% win probability, the draw 29.7%, and Everton 23.2%. It also puts over 2.5 goals at 32.4%. That is a clear lean, but not a majority, and the over 2.5 figure suggests the model expects a tighter game than Chelsea's recent scorelines might imply. The important caveat is timing: the prediction was generated on Sep 18, before Chelsea's 5-1 win over Bournemouth and Everton's 2-2 draw at Hull City. It therefore did not see the latest stored results. The data also has no odds, no external forecast, no confirmed lineups and no injury reports. The model's own counterpoint notes that lineups are not fully settled and that tactical or personnel changes can still alter the result. This is an estimate, not a guarantee.
Why the model might still be right
The case for Chelsea starts with the stored sample efficiency: 1.38 points per game against Everton's 1.00, and 1.40 against 1.20 in Premier League matches. Chelsea have scored 15 goals in six current-table games, and even with away leaks they have more win equity than Everton's draw-heavy profile. Everton have drawn four of six, which shows a team that struggles to close matches out. If Chelsea's attack travels even partially, the away win probability makes sense. The low over 2.5 probability could also point to a narrow away win rather than a shootout, which would fit the model's lean without requiring Chelsea to reproduce their 5-1 home performance.
Why Everton can break the read
The strongest counterevidence is the home-away split. Everton are unbeaten in the current table, with a home record of two wins, one draw and no defeats, conceding only twice at home. Chelsea's away record is one win, no draws and two defeats, with seven conceded in three. A key signal has Everton scoring 1.80 goals per home game and Chelsea conceding 2.00 per away game. If those patterns hold, Everton can score, and the 29.7% draw becomes very live. The draw is not the model's top outcome, but it is close enough to matter. Chelsea's recent away defeats at Brentford and Arsenal also show that the road version is not the same as the home version. The head-to-head sample is only two matches, so it cannot settle the argument either way.
What to watch
Watch whether Chelsea's away defense can hold closer to its overall level than its road sample. If Chelsea concede early or repeatedly, Everton's home scoring and draw habit could pull the match toward the 29.7% draw or an Everton result. If Chelsea's attack reproduces the Bournemouth performance, the 47.1% away lean looks justified. The specific conditional is whether Everton can turn their unbeaten floor into a win, or whether their draw-heavy profile keeps Chelsea's away lean from becoming a comfortable one. That is the question to carry into kickoff.