AC Milan vs Inter — Serie A forecast, Saturday, 31 October 2026

Serie A·Saturday, 31 October 2026·19:45
AC Milan

vs

19:45

4th vs 1st

Inter

Metis Forecast

Market-anchored distribution

Model updated

AC Milan

30%

Draw

26%

Inter

44%

Expected goals

1.36–1.68

AC Milan · Inter

Inter win 44% of the current Metis Forecast. Most likely scores: 1–1 (12.1%), 1–2 (9.0%), 2–1 (7.5%).More forecast detail +

Both teams score in 61% of simulations.

Those three cover 29% of outcomes; the other 72% is spread across every remaining scoreline.

The most likely Inter win is 1–2 (9.0%).

AC Milan score at least twice in 39%; Inter in 50%.

Scoreline distribution

Match outcomes

Both teams score61%
Over 1.5 goals82%
Over 2.5 goals58%
Over 3.5 goals36%
Over 4.5 goals19%
Team scoring & clean sheets
AC Milan clean sheet19%
Inter clean sheet26%
AC Milan score 2+39%
Inter score 2+50%
AC Milan score 3+16%
Inter score 3+24%

Model vs market

SourceHomeDrawAway
Raw Metis model29%25%45%
Market reference30%26%44%
Metis Forecast30%26%44%

Market reference: pinnacle, de-vigged (shin) · price snapshot .

The forecast combines the model’s scoreline structure with market information where a fresh price exists. The raw model is kept separate and benchmarked independently against the Pinnacle opening price.

Most likely scores

1–112.1%
1–29.0%
2–17.5%
0–17.0%
0–26.7%
2–26.4%

Want the model's read each week? Subscribe for free.

Recent form

AC Milan

23 AugAvs Torino2–1W
28 AugHvs Venezia2–0W
6 SeptAvs Juventus1–1D
12 SeptAvs Lazio2–2D
20 SeptHvs Lecce3–0W

Inter

22 AugHvs Monza4–1W
30 AugAvs Cagliari1–0W
5 SeptHvs Napoli3–2W
14 SeptHvs Udinese5–3W
19 SeptAvs AS Roma2–2D

Head to head

No previous meetings in the Metis results archive for this competition.

Model context

StatAC Milan
(Home · 3 matches)
Inter
(Away · 3 matches)
BTTS rate33.3%66.7%
Over 2.5 rate66.7%66.7%
Clean sheet66.7%33.3%
Avg goals3.333.00

Home stats use AC Milan's home record; away stats use Inter's away record this season. These are completed-match rates over the samples shown above, not model probabilities — early in a season a handful of matches produces 0% and 100% readings.