A one-point gap with a soft home edge
Newcastle and Everton meet in round nine at St James' Park on 2 November separated by a single point in the Premier League table: Newcastle are 10th with 11 points from eight games, Everton 13th with 10. Nutmegly's AI read makes Newcastle the lean, with a 40.6% home-win probability against 32.3% for Everton and 27.1% for a draw. The important word is lean. The home win is the leading outcome, not a majority, and the model is not describing a clear favourite.
The table also hides the more interesting tension. Newcastle's overall record is three wins, two draws and three defeats, with 12 goals for and 14 against. Everton are two wins, four draws and two defeats, with 8 for and 9 against. Newcastle have won more; Everton have drawn more. At home, Newcastle are one win, two draws and one defeat, with six goals for and six against in four games. Away, Everton are no wins, three draws and one defeat, with three goals for and five against in four. So the home lean is not built on a fortress. It is built on a marginal points edge and the possibility that Newcastle's home scoring can outweigh their home concessions.
Newcastle's sample: a good latest result, not a settled home case
The stored recent sample for Newcastle runs from a 2-0 win at Crystal Palace on 25 October back to a 1-4 defeat at Leeds on 14 September. In between there is a 0-1 home loss to Aston Villa, a 1-4 away defeat at Coventry and a 2-1 home win over Hull City. That is a mixed sequence: two wins, three defeats, and two of those defeats by three-goal margins. The latest result is positive, but the sample does not support the idea that Newcastle have turned St James' Park into a reliable points machine. Their four home league games have produced one win, two draws and one defeat, with the goals split evenly at six each.
That even home goal split is the detail that matters for the model. Newcastle's home attack has produced six goals in four, but their home defence has conceded the same number. A 40.6% home win probability is therefore a reasonable expression of advantage, not a statement of control. If the game becomes open, Newcastle's home scoring can justify the lean. If it becomes a sequence of equal exchanges, the draw and away-win probabilities have more room than the headline number suggests.
Everton's away draws are the strongest counterargument
Everton's case is not built on away wins. It is built on away draws. In four away league games they have drawn three and lost one, scoring three and conceding five. Overall they have drawn four of eight. The stored recent sample shows the same pattern: a 0-2 defeat at Arsenal on 24 October, a 0-2 home defeat to Chelsea on 17 October, a 2-2 draw at Hull on 11 October, a 1-0 home win over Ipswich on 19 September and a 0-0 draw at Tottenham on 12 September. The last two results are defeats, which is a real warning. But the away profile before that was low-margin and draw-heavy.
This is where the model's home lean deserves scrutiny. Everton's away record of three draws in four is exactly the kind of sample that can pull a 40.6% home win toward a draw. The model still gives Everton a 32.3% win probability, higher than the draw's 27.1%, which suggests the away win is a live possibility in the model, not just a draw hedge. An external forecast supplied alongside the model points in a different direction, and the data notes flag differences in sample definitions. That does not settle the argument, but it is a fair reason to treat the home lean as soft rather than commanding.
The prediction, the absences and the question to watch
Nutmegly's AI read is plain: Newcastle are the lean at 40.6%, with the draw at 27.1% and Everton at 32.3%. The total-goals read is almost a split verdict, with the over 2.5 goals probability at 50.45%. That fits the table evidence more than it contradicts it. Newcastle have scored 12 in eight; Everton have scored 8 in eight. Newcastle's home games have produced 12 goals in four, six for and six against. Everton's away games have produced eight goals in four, three for and five against. The model is not forecasting a shootout; it is describing a match with a slight home edge and a near-even goals line.
The absence picture adds uncertainty. Neither starting XI is confirmed, and the data flags six potential absentees without a confirmed team-by-team breakdown. That is not enough to say one side is weakened, and it is not enough to assume both are at full strength. The stored head-to-head sample is two matches, so it cannot carry much weight either. The match may be decided by whether Newcastle's home scoring edge is real enough to break Everton's away draw pattern.
The question to watch is simple: if the game is level late, can Newcastle find the home goal that separates the lean from the draw, or does Everton's three-draws-in-four away profile hold? That is the match within the match. Newcastle's 40.6% is the model's answer before kickoff. The 90 minutes will test whether their one-win home record is a warning or just a small sample.