Kompany's Bayern and Kim Min-jae, big matches. Big matches get a full animated analysis; every match gets its score, xG and a graded prediction.
Tap a highlighted match to open its analysis. The dashed card at the end is the next match's preview.
Augsburg are fourth and scoring freely, but their underlying numbers are thin. They are also the only team to beat Bayern in the league last season.
Bayern win 82% · draw 11% · Augsburg win 7%. Expected goals 3.32 – 0.97.
Bayern have created 20.69 xG and scored 25 in 7 games. Opponents: 4.36 xG, 4 goals. In the league, 10 points from 4 games against 10.7 expected.
| # | Team | P | GD | Pts |
|---|---|---|---|---|
| 1 | Dortmund | 4 | +7 | 12 |
| 2 | Bayern | 4 | +12 | 10 |
| 3 | Freiburg | 4 | +9 | 10 |
| 4 | Augsburg | 4 | +5 | 7 |
| 5 | Leverkusen | 4 | +5 | 7 |
| 6 | Mainz | 4 | +4 | 7 |
Bars: expected goals (xG) created. Rings: actual goals. ■ Bayern ■ Opponent
Solid: actual goals. Dashed: xG. Bayern in the team colour, opponents in the contrast colour.
Expected points: the points the chances deserved, from each Bundesliga match's xG.
| Date | Comp | Opponent | Score | xG | W / D / L from the chances | Verdict | Pre-match model |
|---|---|---|---|---|---|---|---|
| 22 Aug | DSC | Dortmund | 2–1 | 2.37–1.65 | 54% / 20% / 27% | Deserved | – |
| 28 Aug | BL | Stuttgart | 5–1 | 4.06–0.79 | 91% / 6% / 3% | Deserved | – |
| 2 Sep | DFB | Osnabrück | 4–1 | 3.11–0.13 | 94% / 6% / 1% | Deserved | – |
| 5 Sep | BL | Schalke | 0–0 | 1.94–0.21 | 80% / 17% / 3% | Points dropped | – |
| 10 Sep | UCL | Bodø/Glimt | 5–0 | 2.61–0.30 | 87% / 11% / 3% | Deserved | – |
| 13 Sep | BL | Elversberg | 2–1 | 2.83–0.94 | 76% / 14% / 10% | Deserved | – |
| 18 Sep | BL | Union Berlin | 7–0 | 3.77–0.34 | 94% / 5% / 1% | Deserved | – |
“W / D / L from the chances” replays each match with the same xG using a Poisson model. xG from FotMob (Opta data) match pages. For Union, FotMob's report and FC Bayern both give 3.77; FotMob's match page shows 3.71. Kick-offs confirmed by bundesliga.com, the clubs and DFB/UEFA listings. Europe switches to CET on 25 Oct.
Seven competitive games, six wins and a draw, 25 goals scored and four conceded. The chance numbers back it up: Bayern have created 20.7 expected goals and allowed 4.4 (FotMob), about 3.0 per game for and 0.6 against. That is roughly the 2025/26 attack (2.9 xG per game in the league, the season Bayern scored a record 122) with a clearly tighter defence so far. Only Dortmund (1.65 xG in the Supercup) have created more than one expected goal against them.
The attack runs through Michael Olise more than ever: 8 goals and 4 assists in seven games, including a hat-trick against Union. Kane has six and became the fastest player to 100 Bundesliga goals. The one blank was telling, though. At Schalke Bayern had 83% of the ball and 1.94 xG but did not score. A deep, disciplined block can still make them wait, and that draw is why they are second, two points behind a Dortmund side that has won all four league games.
For Korean readers: Kim Min-jae, in his fourth season at Munich, is the third-choice centre-back behind Upamecano and Tah. Kompany uses him when he rotates. He played all 90 minutes at Elversberg, where Abendzeitung rated him above Upamecano, and the second half against Union after Upamecano was booked. Kompany's high line suits him, because it rewards defenders who step in early and win the ball back quickly, but his minutes depend on rotation. In October that works in his favour.
Every diagram plays on a loop like a short GIF: the dots are players moving, the white ball travels with the play, and lines trace where the ball and the runners went. Goal diagrams show the scoring team attacking up the page. Use Pause to freeze any moment. Diagrams are reconstructions from match reports and analysis, not tracking data, so positions are approximate.
Facts (line-ups, minutes, stats, quotes) are cross-checked across at least two outlets where possible. xG sources are named in the notes above and in each match. Tactical claims are credited to the outlet that made them. YouTube analyses are linked, not transcribed. Everything under Column is our own analysis.