A model pick that cuts against the table
Bayern München arrive at the WWK Arena second in the Bundesliga, unbeaten in four matches, with 14 goals scored and two conceded. Augsburg sit fourth on seven points from four games, having won two, drawn one and lost one. On that evidence the visitors look like the obvious pick. Nutmegly's AI read disagrees. The model leans home, handing Augsburg a 49.1% win probability, the draw 21.3% and Bayern 29.5%. That gap between the current table and the model's read is the central storyline of this fifth-round fixture.
What the two form lines actually show
Augsburg's stored results, newest first, read: a 2-3 defeat away at Werder Bremen on 19 September, a 2-2 home draw with Bayer Leverkusen on 12 September, a 4-1 win away at Eintracht Frankfurt on 6 September, a 3-0 home win over FC Schalke 04 on 30 August, and a 0-4 defeat away at Union Berlin back on 16 May. The four most recent of those line up exactly with the current table: two wins, one draw, one defeat, 11 goals for and six against.
Bayern's stored run is louder: a 7-0 home win over Union Berlin on 18 September, a 2-1 away win at SV Elversberg on 13 September, a 5-0 home win over Bodo/Glimt on 10 September, a 0-0 away draw at Schalke on 5 September, and a 5-1 home win over VfB Stuttgart on 28 August. Their four league matches match the table too: three wins and a draw, 14 goals for, two against. But the split matters. Bayern have scored 12 goals in two home league games and just two in two away games, conceding one on the road. Their away record is solid rather than spectacular.
The case for the home lean
Augsburg's home numbers are the strongest support for the model. They are unbeaten at the WWK Arena in the current league sample, with five goals scored and two conceded across two home matches. Bayern's away output of two goals in two games is the softest part of their profile, and it is the kind of split that can keep a home side in a match. Across the stored eight-match samples, which span older seasons and should not be read as current-season totals, Augsburg's points-per-game rate edges Bayern's. The model also puts the over-2.5 goals probability at 70.1%, so it is not forecasting a cagey stalemate.
Where the read could break
The counterevidence is real. The model's own counterpoints flag that both teams carry a scoring threat and that a single defensive error can swing the result, which is exactly the risk in a match where the win probabilities are split rather than lopsided. The read was also generated on 11 September, before Augsburg's 2-3 loss at Werder Bremen and before Bayern's 7-0 demolition of Union Berlin, so it does not fully account for the most recent stored results. There is no odds summary and no external forecast in the data to calibrate against, and the head-to-head sample is only two matches, too thin to lean on. Bayern's ceiling is also visible: 7-0 and 5-1 home wins show what happens when they convert early.
Lineups and injuries are not confirmed in the available data, so nothing here should be read as evidence that either squad is fully available or that a selection decision has already been made. The stored calendar shows a long gap since each side's most recent recorded match, 33 days for Augsburg and 29 for Bayern, but gaps of that length are not treated as a rest advantage and should not be framed as freshness. They simply mean the sample is what it is.
The watchpoint
The match turns on whether Bayern's away scoring pattern holds or reverts. If their road output stays as modest as two goals in two league away games, Augsburg's unbeaten home record and the model's 49.1% home lean have a genuine path. If Bayern instead carry their home ruthlessness onto the road and score early, the 70.1% over-2.5 read becomes the more relevant number and the home lean looks exposed. Watch the first quarter of the match for which version of Bayern's away game shows up, because that is the single thread connecting the table, the form lines and the model's surprising pick.