The model's split-screen opener
This fourth-round Bundesliga match at BayArena sets the table against the model. Leipzig sit fifth on six points with a 9-3 goal record; Leverkusen are seventh on four points with 8-5. Yet Nutmegly's AI read leans home: 47.4% Leverkusen win, 23.5% draw, 29.0% Leipzig win. That tension is the storyline.
The table is early. Three games played each, so rank and points are a snapshot, not a verdict. Within that snapshot, Leipzig have the better all-around record: two wins, one loss, nine goals scored and three conceded. Leverkusen have one win, one draw, one loss. The model's home lean is not ignoring that; it is weighing something else.
What the stored results say
The newest stored results show Leverkusen's recent run: a 2-2 draw at Augsburg on September 12, a 4-0 home win over Union Berlin on September 5, and a 2-3 loss at SV Elversberg on August 29. The sample also includes older May results: a 1-1 home draw with Hamburger SV and a 1-3 loss at VfB Stuttgart. Across the last eight stored matches, Leverkusen's points-per-game rate is 1.38.
Leipzig's stored sample runs: a 5-0 home win over Hamburger SV on September 13, a 1-4 loss at Como on September 10, a 1-3 loss at Werder Bremen on September 5, and a 3-0 home win over Borussia Mönchengladbach on August 29. An older May result, a 1-4 loss at SC Freiburg, is also in the sample. Their last-eight points-per-game rate is 1.50. In the Bundesliga sample specifically, the key signal gives Leverkusen 1.00 points per game and Leipzig 1.80. That is the strongest counterweight to the model's home lean.
Why the home lean has a case
The home and away splits are where Leverkusen's case lives. In the current table, Leverkusen have played one home game and won it 4-0, with four goals scored and none conceded. Leipzig have played one away game and lost it 1-3, scoring once and conceding three. Those are tiny samples, but they point in the same direction as the model.
The same-venue averages sharpen the point: Leverkusen's home goals average is 3.20, and Leipzig's away goals-conceded average is also 3.20. That symmetry suggests Leverkusen can create against a Leipzig side that has leaked on the road in the stored data. The model also gives over 2.5 goals a 64.5% probability, so the expected script is open rather than cagey.
The model's probabilities are estimates, not guarantees. A 47.4% home win is the leading outcome, but it is not a majority. Leipzig's 29.0% away win is a live possibility, and the draw at 23.5% is not negligible. The lean is home, not home certainty.
The counterevidence and the watchpoint
The strongest supplied counterevidence is Leipzig's overall efficiency. They have more points, a better goal difference, a higher recent points-per-game rate, and a much better Bundesliga sample rate. If those numbers reflect current quality better than the home and away splits do, the model's home lean will look fragile.
The model's own counterpoints add caution: both teams have scoring profiles, and a single defensive error can amplify result volatility. Lineups are not fully confirmed, so tactical and personnel choices could still change the result picture. The injury reports list no absent players, but that does not guarantee full availability, and missing lineup confirmation is not evidence of a settled XI. The stored head-to-head sample is only two matches, too small to settle the argument.
The rest window is a minor home signal: Leverkusen's last stored match was eight days before kickoff, Leipzig's seven. The data cautions that gaps in records do not prove extra rest or absence of other matches, so this should not be overread.
Watchpoint: whether Leverkusen's home scoring average can actually meet Leipzig's away defensive record. If Leipzig keep the away form shown in their stored losses, the model's home lean gains support. If they instead carry their home scoring form into the BayArena, the match tilts toward the open, high-variance game the over 2.5 probability and the model's counterpoints describe. That is the question to watch.