The table says comfortable; the sample says closer
AC Milan arrive at this Serie A fixture unbeaten in the current league table: two wins, two draws, eight points, sixth place. Lecce sit 12th on six points, with two wins and two defeats. Nutmegly's AI read leans firmly home, giving Milan a 74.6% chance of winning, a draw 17.1% and Lecce 8.3%. That is a heavy split for a round-five meeting between teams separated by two points.
The storyline is that gap. Milan are the unbeaten side with the better defensive record. Lecce are the mid-table team whose stored form sample scores better per game than Milan's. Both things are true, and they pull in opposite directions.
Why the model likes Milan anyway
The case for the favourite is structural rather than purely statistical. Milan's eight points come with a plus-three goal difference, seven scored and four conceded, and no defeats. Lecce's six points come with a minus-two, five scored and seven conceded. Milan are also the higher-ranked side and have the venue.
The match-up data hardens the logic. In the stored same-venue sample, Milan's home scoring rate is 1.00 goals per game while Lecce's away concession rate is 0.80. Milan have played one home league game this season and won it 2-0. The rest gap is negligible: eight days since Milan's most recent recorded match, seven for Lecce.
The counterargument: Lecce's away numbers travel well
Here is where the favourite's case gets uncomfortable. Across the stored eight-match form samples, Lecce average 1.88 points per game and Milan average 1.38. Narrow the lens to Serie A fixtures within the sample and Lecce still lead, 1.80 to 1.60. Lecce's two away games in the current table produced one win, one defeat, two goals scored and one conceded, a tidy return that contrasts with the six goals they have conceded across two home fixtures.
Milan's own recent record is a study in draws rather than dominance. Their last two recorded league outings were away draws, 2-2 at Lazio and 1-1 at Juventus, following a 2-0 home win over Venezia and a 2-1 win at Torino; the stored sample also carries a 1-2 home defeat to Cagliari from May. Lecce's swings are wider: a 3-2 home win over Monza, a 0-1 loss at Cagliari, a 0-4 home defeat to Roma, a 2-0 win at Venezia and a 1-0 home win over Genoa, also from May.
Reading the sample honestly
Two cautions apply to both sides. The stored results span seasons, so the May fixtures are prior-season data rather than current form, and the same-venue and Serie A samples run to five matches each. Points-per-game averages over a handful of games describe a mood more than a mechanism. The stored head-to-head record contains two matches, which is not a pattern.
On availability, the picture is incomplete. Lineups are not confirmed, and while no absentees are reported, a quiet injury feed is not evidence of a fully fit squad. Nutmegly's AI prediction, generated on 17 September, also gives an over-2.5 probability of 54.3%, a lean toward goals rather than a conviction, and the model's own counterpoint notes that unsettled starting XIs and late selection or shape changes could move the picture. Forecasts are estimates, not results: 74.6% still leaves roughly a one-in-four chance this does not end in a Milan win.
The watchpoint
The question is not whether Milan are favourites but whether their home scoring runs above the stored one-goal-per-game sample. If Lecce keep their away concession near 0.80 and Milan's home output stays near one goal, the 74.6% will look generous and a single moment could settle it. If Lecce's wider volatility travels with them, the sample that contains both a 0-4 defeat and a 3-2 win, Milan's unbeaten start has room to extend. The evidence to trust is the shape of Milan's home attack, not the table.