Comebacks & Blown Leads
Every match has a moment where one team was closest to losing. We take the winner's win probability at the end of every round and find the lowest it ever got. That number is the comeback; the same number from the other side is the blown lead.
Miami fell to 7.3% against Las Vegas on 8/22/2026 — down 7 after 3 rounds — and won by 4. That's the biggest comeback of the season.
Only the states after each round count. The opening bell is always 50-50, so including it would floor every match at 50% and make wire-to-wire wins read as comebacks. See the full ranking →
A team is credited with a comeback win when they were once below 25%. That happened 12 times in 54 matches.
| Winner fell | Matches |
|---|---|
| below 5% | 0 of 54 |
| below 10% | 4 of 54 |
| below 15% | 7 of 54 |
| below 25% | 12 of 54 |
| below 35% | 21 of 54 |
The median low point for a winning team is 0.409 — most winners are never in real danger. Draws are excluded: with no winner there is no comeback.
Worth calling out on its own. Los Angeles blew 4 leads — the largest from 92.4%. No other team blew more than 2.
It's descriptive, not a skill. A team with comeback wins isn't better at comebacks — they're a team that fell behind and recovered, and falling behind isn't a virtue. Read the leaderboard as a record of what happened, not a ranking of grit.
Built on the same win-probability model as WPA — no separate model and no extra assumptions. Scheduled-round differences (two 2026 matches ran 21 rounds instead of 24) are already baked into those probabilities.
† Match 14: Win-probability note: this match's final entry (round 25) was an administrative points award recorded after the match ended and is excluded from WPA. Win probability here reflects the competitive scoreboard (San Antonio 15–13); the official result stands in the standings.