The table and the model disagree
This is the most interesting kind of Champions League preview: the current league-phase table points one way, the AI read points another. Arsenal have won both of their matches, scored four and conceded none, and sit seventh with six points. Bayern have one win and one draw, scored eight and conceded three, and sit tenth with four points. Yet Nutmegly's AI read makes Bayern the clear home lean: 80.1% for a home win, 12.1% for a draw and 7.8% for an Arsenal win. The same read puts the over 2.5 goals probability at 76.5%. That is a strong forecast, not an observed result, and it was generated on 22 September, before the latest results in the stored sample.
Why Bayern's home case is real
The home lean is not built on nothing. Bayern's one home match in this league phase was a 5-0 win, and their stored home sample averages 4.60 goals per home match. The latest stored result is a 6-1 home win over RB Leipzig on 17 October; the sample also includes a 7-0 home win over Union Berlin on 18 September and a 2-1 away win at SV Elversberg on 13 September. The draws are part of the picture too: 3-3 away to Viking on 13 October and 2-2 away to FC Augsburg on 10 October. That is a team whose recent games have produced plenty of goals, and the model's 76.5% over 2.5 probability fits that pattern. In the broader stored recent sample, Bayern's points-per-game rate is 2.50 against Arsenal's 2.38, a small additional edge for the home side.
Arsenal's countercase is stronger than 7.8%
Arsenal's current league-phase record is the strongest counterevidence. Two wins, four goals scored, none conceded, six points. Their latest stored result is a 1-0 away win at Nottingham Forest on 18 October. The stored sample also includes a 3-0 home win over Lille on 13 October, a 2-1 home win over Leeds on 10 October and a 2-0 away win at Sunderland on 12 September; the exception is a 0-3 away defeat at Brighton on 19 September. Across the stored away sample, Arsenal concede 0.80 per away match. The AI read still gives Arsenal only a 7.8% win probability. That looks harsh against a two-match record with zero goals against, even allowing for the small sample. It is also why the 12.1% draw should not be treated as noise.
The missing pieces and the watchpoint
The data does not include confirmed lineups, injury reports, an odds summary or an external forecast. That does not mean both squads are fully available; it means this preview cannot lean on confirmed team news. The supplied counterpoint is exactly that: starting lineups are not fully settled, and late personnel or tactical choices could change the result picture. The model's age matters too. A read generated on 22 September predates Bayern's 6-1 win over Leipzig and Arsenal's 1-0 win at Forest, so the probabilities are a pre-match estimate, not a verdict on the latest form.
The match at the Allianz Arena on 21 October therefore has a clear watchpoint. If Bayern score first and the game opens up, the 76.5% over 2.5 probability becomes the more useful guide than the 80.1% home-win number. If Arsenal keep the game level deep into the second half, their zero-goals-against league-phase record and the 12.1% draw deserve more weight than the headline home lean. That is the conditional question: can Arsenal's current defensive record survive Bayern's home scoring sample, or does the model's home read get the last word?