

Sources complete · 3 of 3 models
What the models saw→
Goals against expected, both sides
Each provider runs its own shot model over the same shots. They weight chance quality differently, so they disagree. Gaps under 0.3 xG are noise.
Models that reported put LA Galaxy in a 2.18 xG band. A very wide spread. Treat any single figure here with caution.
3 of 3 models complete. Values sit side by side. They are not averaged. How we handle this
Who made the match→
16 players · measured against their own per-90 baselines
2 goals, 3 shots, 3 on target, 3 key passes
1 shot, 4 key passes
98 minutes
Component ratings out of ten. An average night is 6.5. Above is better than that player's own baseline, below is worse. Ratings need 300 minutes of history.
What it changes
Got a take on that one?
I check it against the numbers and show the working.
Coming soon
The take tester is being rebuilt.