How the Stats Work
Saw a column you didn't recognise? Here they all are in a sentence each. The full explanation of any one of them is further down the page.
- WAR (Wins Above Replacement)
- How many wins a fighter added compared to the kind of fighter a team could easily find to replace them. Built from their scoring rate, how many rounds they fought, and what a win costs in points. How WAR works →
- Stakes Faced (Leverage Index)
- How much was riding on a round before it started. It belongs to the moment, not the fighter, so both fighters in a round have the same number. 1.00 is an ordinary round; 6.63 is the maximum — a tied match with one round to go. A fighter's average tells you how big their moments were, not how they did in them. How Leverage works →
- Timing (Clutch)
- The part of a fighter's WPA that came from when their results landed. We compare what they actually did to what the same wins and losses would have been worth in ordinary rounds; the difference is Timing, in wins. Positive means their wins landed in the big moments. Disqualifications do not count. Biggest moments →
- WPA (Win Probability Added)
- How much a round moved the team's chance of winning, credited to the fighter who caused it. A round won with the match on the line is worth far more than one won in a blowout. How WPA works → · Leaderboard →
- SOS (Strength of Schedule)
- How good the opponents were that a fighter had to face. Above zero means a harder schedule than most. Rounds an opponent fought against this fighter are thrown out first — otherwise beating someone four times would drag their numbers down and make your own schedule look easy. How it works → · Leaderboard →
- aNPPR (Adjusted NPPR)
- Net points per round, once you account for who a fighter actually fought. Every fighter is rated against every other at the same time, across the whole season. Fighters with few rounds get pulled toward average on purpose. Each rating shows a range; gaps under 0.20 do not mean anything. How it works →
- Comeback Win / Blown Lead
- The lowest the eventual winner's chances ever fell during a match. Dip under 25% and it counts as a comeback win — and, from the other side, a blown lead for the team that let it slip. Draws do not count. How it works → · Biggest comebacks →
See them all in action on the advanced stats page, and on every fighter profile.
Win Probability Added
Every round changes your team's chance of winning. WPA is that change, credited to the fighter who caused it.
Every round changes your team's chance of winning the match. WPA — Win Probability Added — is that change, credited to the fighter who caused it. Win a round and your team's chances go up; you get that gain. Lose it and your opponent gets the gain while you get the equal-and-opposite loss.
The point of the stat is context. In the box score, a round won with the match on the line looks identical to one won in a blowout. WPA is the stat that knows the difference.
See it in action on the advanced leaderboard, on every fighter profile, and round by round on every match page. Two stats build directly on this model: Leverage Index and Clutch and Comebacks and Blown Leads.
Match 6, Phoenix vs Boston, round 16. Phoenix trailed by 2 points with 8 rounds left to fight. In that spot, teams win about 25% of the time — they need to claw back two points against a team that only has to hold serve.
Manuel Villalobos then scored a TKO — worth 4 points. One round later Phoenix wasn't down 2; they were up 2 with 7 rounds left, a spot teams convert about 75% of the time.
Before the round: 25% chance. After the round: 75% chance. Villalobos personally moved his team 50 percentage points toward a win. That's WPA +0.500 — the biggest single round of the 2026 season. His opponent is charged the mirror image, −0.500.
Take the most ordinary result there is: a 1-point decision win while the match is tied. Watch what it's worth as the match runs out of road:
| Situation (match tied) | WPA for the win |
|---|---|
| 24 rounds left | +0.08 |
| 20 rounds left | +0.09 |
| 12 rounds left | +0.12 |
| 8 rounds left | +0.14 |
| 4 rounds left | +0.20 |
| 1 round left | +0.50 |
Same punch, same scorecard, wildly different stakes. Fans understand this intuitively from watching — a tied final round is everything, a tied 4th round is a skirmish. WPA just gives that intuition a number.
Holding a lead while rounds tick away raises your win probability even when the score doesn't move. Fewer remaining rounds means fewer chances for the trailing team to catch up — so the clock quietly works for whoever is ahead.
A real case: match 8, round 23. Philadelphia lost the round, but their win probability still went up — Las Vegas simply no longer had enough rounds left to close the gap. WPA reflects this: a round can carry value even at 0-0, and a lost round late can still be a "good" round for the team protecting a big lead. This is intended behavior, not a quirk.
Disqualification rounds are credited to neither fighter. Both fighters receive exactly zero WPA for a DQ round.
The reasoning: a disqualification is a referee's decision, not a boxing performance. Beyond that, the source data records only who was awarded the points — not who was disqualified. We can usually infer the offender from the scoring, but "usually" isn't a standard a public stat should rest on: a DQ can arise from an accidental foul, a cut, or a fighter unable to continue, and the data cannot distinguish those cases. Rather than credit or charge the wrong fighter, WPA attributes disqualification rounds to nobody.
The points still count. They're on the scoreboard, they shift the match differential, and they change every following round's win probability. Only the credit is withheld. This affects 8 rounds in the 2026 season. The largest came in match 13, where a DQ in the final round of a tied match decided it — worth 0.500 of win probability, now attributed to nobody.
The model uses exactly two inputs: the score differential and how many rounds are left. Nothing about who the fighters are.
That is deliberate. How good the fighters are, and how good their teams are, is left out entirely. WPA measures what happened to the team's chances — not how impressive it was against that particular opponent.
It was built from all 55 matches and 1,314 rounds of the 2026 season. And it balances exactly: setting aside disqualification rounds, every fighter on a winning team splits exactly +0.500 of WPA between them, and every fighter on a losing team splits −0.500. Nothing is invented and nothing goes missing.
Read this table as: "when the model said 62%, teams actually won about 65% of the time." Across every round of the season, the model's stated chances track what actually happened:
| Model said | Times | Actually won |
|---|---|---|
| 0–5% | 422 | 0% |
| 5–15% | 165 | 12.7% |
| 15–30% | 245 | 24.9% |
| 30–45% | 313 | 34.7% |
| 45–55% | 338 | 50.0% |
| 55–70% | 313 | 65.3% |
| 70–85% | 245 | 75.1% |
| 85–95% | 165 | 87.3% |
| 95–100% | 422 | 100% |
- It does not account for opponent difficulty. Beating the best fighter in the league counts the same as beating the worst.
- It is built on a single season of data. Expect refinements as more seasons accumulate.
- Fighters with very few rounds can post extreme values. That's why the leaderboard defaults to a 10-round minimum.
- It correlates about 0.78 with raw round win rate. WPA is round wins adjusted for when they happened — it is not a completely independent measure of skill.
- Disqualification rounds are attributed to nobody, so a small amount of win probability in a handful of matches is unassigned.
Technical Details — the formal spec
Two-stage model (version 2026.1, fit on the 2026 season). Stage 1: the per-round point margin follows a fixed empirical distribution (symmetrized so neither side has a built-in edge). The chance of winning from differential d with r rounds remaining is obtained by convolving that distribution r times, shifting by d, and taking P(final > 0) + ½·P(final = 0) — the half credit because TBL matches can end in a draw. Stage 2: the baseline probability is sharpened through a logit transform with a single fitted parameter γ = 1.4164008922 — real TBL leads are more decisive than a pure random walk predicts, because teams that lead are often genuinely better. r = 0 is deterministic (1 / 0 / ½), never routed through the formula.
Margin distribution (probability of each per-round point margin):
| Margin | -4 | -3 | -2 | -1 | 0 | +1 | +2 | +3 | +4 |
|---|---|---|---|---|---|---|---|---|---|
| P | 0.025875 | 0.003044 | 0.038432 | 0.431507 | 0.002283 | 0.431507 | 0.038432 | 0.003044 | 0.025875 |
Per-round WPA. For a match with N scheduled rounds (24 for most 2026 matches; matches 7 and 48 were 21), round t compares WP(diff after round t, N−t rounds left) against WP(diff after round t−1, N−t+1 left). The winning side's fighter is credited the change; the other fighter gets the exact negative.
Rules. DQ rounds: zero credit to both fighters; points still count toward the differential (8 rounds in 2026). Match 14's round 25 was a post-match administrative award and is excluded entirely — WPA uses the competitive scoreboard (San Antonio 15-13) while the official result stands in the standings. Match 25 was an official draw; both teams' WPA sums to 0.000. Rounds scored 0-0 still elapse and still move win probability for the leading team; one such round (match 25, round 19) has no fighter data and is assigned to no fighter.
Validation. Before the DQ adjustment: every round is zero-sum between the teams; each match's round WPA telescopes to (outcome − ½) within 1e-9; the season sums to 0.000; and the table satisfies WP(d,r) + WP(−d,r) = 1 for every cell. After the adjustment, all 8 DQ rounds carry exactly 0.0 on both sides. The shipped lookup table is verified against fixed checksums (e.g. WP(+1,1) = 0.788514, WP(+7,2) = 0.999963) on every regeneration and in the test suite.
Win probability grid — the team's chance of winning at score differential d (rows) with r rounds remaining (columns):
| d \ r | 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| +20 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 |
| +19 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 |
| +18 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 99.9 |
| +17 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 99.9 | 99.9 | 99.9 | 99.9 |
| +16 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 99.9 | 99.9 | 99.9 | 99.9 | 99.8 | 99.8 |
| +15 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 99.9 | 99.9 | 99.9 | 99.9 | 99.9 | 99.8 | 99.8 | 99.7 | 99.7 |
| +14 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 99.9 | 99.9 | 99.9 | 99.9 | 99.8 | 99.8 | 99.7 | 99.7 | 99.6 | 99.6 | 99.5 |
| +13 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 99.9 | 99.9 | 99.9 | 99.8 | 99.8 | 99.8 | 99.7 | 99.6 | 99.5 | 99.5 | 99.4 | 99.3 | 99.2 |
| +12 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 99.9 | 99.9 | 99.8 | 99.8 | 99.7 | 99.6 | 99.6 | 99.5 | 99.4 | 99.2 | 99.1 | 99.0 | 98.9 | 98.7 |
| +11 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 99.9 | 99.9 | 99.8 | 99.8 | 99.7 | 99.6 | 99.5 | 99.3 | 99.2 | 99.1 | 98.9 | 98.8 | 98.6 | 98.4 | 98.3 | 98.1 |
| +10 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 99.9 | 99.9 | 99.8 | 99.7 | 99.6 | 99.5 | 99.4 | 99.2 | 99.0 | 98.9 | 98.7 | 98.5 | 98.3 | 98.0 | 97.8 | 97.6 | 97.3 | 97.1 |
| +9 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 99.9 | 99.9 | 99.8 | 99.6 | 99.5 | 99.3 | 99.1 | 98.9 | 98.6 | 98.4 | 98.1 | 97.8 | 97.5 | 97.2 | 96.9 | 96.6 | 96.3 | 96.0 | 95.7 |
| +8 | 100.0 | 100.0 | 100.0 | 100.0 | 100.0 | 99.9 | 99.8 | 99.6 | 99.4 | 99.2 | 98.9 | 98.6 | 98.3 | 98.0 | 97.6 | 97.2 | 96.9 | 96.5 | 96.1 | 95.7 | 95.4 | 95.0 | 94.6 | 94.2 | 93.9 |
| +7 | 100.0 | 100.0 | 100.0 | 100.0 | 99.9 | 99.7 | 99.4 | 99.2 | 98.8 | 98.4 | 97.9 | 97.5 | 97.0 | 96.5 | 96.0 | 95.5 | 95.0 | 94.5 | 94.1 | 93.6 | 93.1 | 92.7 | 92.2 | 91.8 | 91.4 |
| +6 | 100.0 | 100.0 | 100.0 | 99.8 | 99.6 | 99.2 | 98.7 | 98.1 | 97.5 | 96.9 | 96.2 | 95.5 | 94.9 | 94.2 | 93.6 | 92.9 | 92.3 | 91.7 | 91.2 | 90.6 | 90.1 | 89.5 | 89.0 | 88.5 | 88.1 |
| +5 | 100.0 | 100.0 | 99.8 | 99.5 | 98.8 | 98.0 | 97.1 | 96.1 | 95.2 | 94.2 | 93.3 | 92.4 | 91.6 | 90.8 | 90.0 | 89.2 | 88.5 | 87.9 | 87.2 | 86.6 | 86.0 | 85.4 | 84.9 | 84.4 | 83.9 |
| +4 | 100.0 | 99.8 | 99.4 | 98.4 | 96.7 | 95.5 | 93.8 | 92.6 | 91.1 | 90.0 | 88.8 | 87.8 | 86.7 | 85.9 | 85.0 | 84.2 | 83.4 | 82.7 | 82.0 | 81.4 | 80.8 | 80.3 | 79.7 | 79.2 | 78.7 |
| +3 | 100.0 | 99.4 | 98.1 | 94.3 | 92.9 | 89.7 | 88.5 | 86.2 | 85.0 | 83.3 | 82.3 | 81.1 | 80.2 | 79.2 | 78.4 | 77.6 | 76.9 | 76.2 | 75.6 | 75.0 | 74.5 | 74.0 | 73.5 | 73.1 | 72.6 |
| +2 | 100.0 | 98.6 | 89.5 | 88.1 | 82.7 | 81.6 | 78.4 | 77.4 | 75.5 | 74.6 | 73.2 | 72.5 | 71.5 | 70.8 | 70.1 | 69.5 | 68.9 | 68.4 | 67.9 | 67.5 | 67.0 | 66.7 | 66.3 | 66.0 | 65.6 |
| +1 | 100.0 | 78.9 | 77.6 | 70.5 | 69.8 | 66.7 | 66.1 | 64.4 | 63.9 | 62.9 | 62.4 | 61.7 | 61.3 | 60.8 | 60.5 | 60.1 | 59.8 | 59.5 | 59.2 | 59.0 | 58.8 | 58.6 | 58.4 | 58.2 | 58.0 |
| 0 | 50.0 | 50.0 | 50.0 | 50.0 | 50.0 | 50.0 | 50.0 | 50.0 | 50.0 | 50.0 | 50.0 | 50.0 | 50.0 | 50.0 | 50.0 | 50.0 | 50.0 | 50.0 | 50.0 | 50.0 | 50.0 | 50.0 | 50.0 | 50.0 | 50.0 |
| -1 | 0.0 | 21.1 | 22.4 | 29.5 | 30.2 | 33.3 | 33.9 | 35.6 | 36.1 | 37.1 | 37.6 | 38.3 | 38.7 | 39.2 | 39.5 | 39.9 | 40.2 | 40.5 | 40.8 | 41.0 | 41.2 | 41.4 | 41.6 | 41.8 | 42.0 |
| -2 | 0.0 | 1.4 | 10.5 | 11.9 | 17.3 | 18.4 | 21.6 | 22.6 | 24.5 | 25.4 | 26.8 | 27.5 | 28.5 | 29.2 | 29.9 | 30.5 | 31.1 | 31.6 | 32.1 | 32.5 | 33.0 | 33.3 | 33.7 | 34.0 | 34.4 |
| -3 | 0.0 | 0.6 | 1.9 | 5.7 | 7.1 | 10.3 | 11.5 | 13.8 | 15.0 | 16.7 | 17.7 | 18.9 | 19.8 | 20.8 | 21.6 | 22.4 | 23.1 | 23.8 | 24.4 | 25.0 | 25.5 | 26.0 | 26.5 | 26.9 | 27.4 |
| -4 | 0.0 | 0.2 | 0.6 | 1.6 | 3.3 | 4.5 | 6.2 | 7.4 | 8.9 | 10.0 | 11.2 | 12.2 | 13.3 | 14.1 | 15.0 | 15.8 | 16.6 | 17.3 | 18.0 | 18.6 | 19.2 | 19.7 | 20.3 | 20.8 | 21.3 |
| -5 | 0.0 | 0.0 | 0.2 | 0.5 | 1.2 | 2.0 | 2.9 | 3.9 | 4.8 | 5.8 | 6.7 | 7.6 | 8.4 | 9.2 | 10.0 | 10.8 | 11.5 | 12.1 | 12.8 | 13.4 | 14.0 | 14.6 | 15.1 | 15.6 | 16.1 |
| -6 | 0.0 | 0.0 | 0.0 | 0.2 | 0.4 | 0.8 | 1.3 | 1.9 | 2.5 | 3.1 | 3.8 | 4.5 | 5.1 | 5.8 | 6.4 | 7.1 | 7.7 | 8.3 | 8.8 | 9.4 | 9.9 | 10.5 | 11.0 | 11.5 | 11.9 |
| -7 | 0.0 | 0.0 | 0.0 | 0.0 | 0.1 | 0.3 | 0.6 | 0.8 | 1.2 | 1.6 | 2.1 | 2.5 | 3.0 | 3.5 | 4.0 | 4.5 | 5.0 | 5.5 | 5.9 | 6.4 | 6.9 | 7.3 | 7.8 | 8.2 | 8.6 |
| -8 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.1 | 0.2 | 0.4 | 0.6 | 0.8 | 1.1 | 1.4 | 1.7 | 2.0 | 2.4 | 2.8 | 3.1 | 3.5 | 3.9 | 4.3 | 4.6 | 5.0 | 5.4 | 5.8 | 6.1 |
| -9 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.1 | 0.1 | 0.2 | 0.4 | 0.5 | 0.7 | 0.9 | 1.1 | 1.4 | 1.6 | 1.9 | 2.2 | 2.5 | 2.8 | 3.1 | 3.4 | 3.7 | 4.0 | 4.3 |
| -10 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.1 | 0.1 | 0.2 | 0.3 | 0.4 | 0.5 | 0.6 | 0.8 | 1.0 | 1.1 | 1.3 | 1.5 | 1.7 | 2.0 | 2.2 | 2.4 | 2.7 | 2.9 |
| -11 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.1 | 0.1 | 0.2 | 0.2 | 0.3 | 0.4 | 0.5 | 0.7 | 0.8 | 0.9 | 1.1 | 1.2 | 1.4 | 1.6 | 1.7 | 1.9 |
| -12 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.1 | 0.1 | 0.2 | 0.2 | 0.3 | 0.4 | 0.4 | 0.5 | 0.6 | 0.8 | 0.9 | 1.0 | 1.1 | 1.3 |
| -13 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.1 | 0.1 | 0.1 | 0.2 | 0.2 | 0.2 | 0.3 | 0.4 | 0.5 | 0.5 | 0.6 | 0.7 | 0.8 |
| -14 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.1 | 0.1 | 0.1 | 0.1 | 0.2 | 0.2 | 0.3 | 0.3 | 0.4 | 0.4 | 0.5 |
| -15 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.1 | 0.1 | 0.1 | 0.1 | 0.1 | 0.2 | 0.2 | 0.3 | 0.3 |
| -16 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.1 | 0.1 | 0.1 | 0.1 | 0.2 | 0.2 |
| -17 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.1 | 0.1 | 0.1 | 0.1 |
| -18 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.1 |
| -19 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| -20 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
Stakes & Timing
How much was riding on a round before it started, and whether a fighter delivered in the ones that mattered.
Not every round matters equally. Stakes Faced — Leverage Index, in the technical literature — is how much win probability is on the table before a round starts — how big a swing the round is capable of producing, no matter who ends up winning it.
1.00 is an average TBL round. 2.00 means twice the normal stakes. 0.10 means the round is nearly meaningless. The most important situation possible in TBL is a tied match with one round to go, at 6.63 — more than six times an average round.
LI belongs to the situation, not the fighter. Both fighters in a round face exactly the same leverage.
See the highest-leverage rounds of the season on the Biggest Moments board.
| Situation | LI |
|---|---|
| Tied, 1 round left | 6.62 |
| Tied, 2 rounds left | 4.18 |
| Tied, 4 rounds left | 3.10 |
| Tied, 12 rounds left | 1.86 |
| Tied, opening bell | 1.32 |
| Up 3, 12 rounds left | 1.26 |
| Up 5, 8 rounds left | 0.51 |
| Up 10, 8 rounds left | 0.02 |
| Up 15, 4 rounds left | 0.00 |
Leverage climbs as rounds run out while the score stays close, and collapses toward zero once a lead is out of reach. Up 15 with 4 rounds left, the match is over in all but name — LI 0.00.
TBL matches run in three phases of equal length. Sort every round by phase and something counterintuitive falls out:
| Phase | Rounds | Avg LI | Median LI | Max LI |
|---|---|---|---|---|
| Launch | 438 | 1.21 | 1.32 | 1.57 |
| Middle | 438 | 1.01 | 1.20 | 2.16 |
| Money | 438 | 0.78 | 0.09 | 6.62 |
Money rounds have the lowest average importance of any phase — and by far the highest ceiling. The median Money round has an LI of just 0.09, meaning the match is usually already settled by the time it arrives. But every one of the most important rounds in TBL history is a Money round.
Money rounds aren't consistently big. They're either nothing or everything.
Zoom out and the same pattern holds leaguewide: 22% of all TBL rounds have an LI below 0.10 — roughly one round in 5 is competitively meaningless — while 6.8% carry more than double average importance.
Did a fighter's results come in the moments that mattered? Timing — Clutch, in the technical literature — is the part of a fighter's WPA that came from when their results landed rather than what those results were. It is measured in wins, the same unit as WPA itself, and it is simply the gap between the two numbers below.
Positive Timing means a fighter's wins landed in bigger spots than their losses. Negative means they piled up round wins in garbage time and lost the ones that counted.
Take Money Powell: 19 rounds, 12 of them won — but Stakes Faced of 0.41×, the lowest of any qualified fighter. His WPA is just +0.04 wins. The same results in ordinary rounds would have been worth +0.59. Timing: −0.55 wins. He won plenty of rounds, but almost none of them mattered.
Now take Erika Sanchez. Her WPA of +1.54 wins is well clear of the +1.09 the same results would have been worth in ordinary rounds — Timing of +0.45 wins. Same league, opposite story.
Both are on the advanced leaderboard, sortable by Timing and by Stakes Faced.
- LI is descriptive, not evaluative. A high average leverage means a fighter was used in big spots. It says nothing about whether they were good in them.
- Clutch over one season is noisy. In every sport that computes it, clutch performance has proven hard to sustain year over year. Treat it as a description of what happened, not a prediction of what comes next.
- Neither adjusts for opponent quality.
- Disqualification rounds are excluded from both.
Technical Details — the formal spec
Leverage Index. Before a round, the state is a score differential d and rounds remaining r — where r INCLUDES the round about to be fought. Take every possible round outcome, weight each by how often it occurs, and average the absolute win-probability change:rawLI(d, r) = Σ over margins v of [ P(v) · |WP(d + v, r − 1) − WP(d, r)| ]LI(d, r) = rawLI(d, r) / 0.0753028919P(v) is the same per-round margin distribution the win-probability model uses (see How WPA Works) — it is not redefined here.
The normalizer is frozen. 0.0753028919 is the mean rawLI across all 1,314 competitive rounds of 2026. It is deliberately never recomputed per season: if it were, an LI of 1.20 in 2026 and 1.20 in 2027 would mean different things and the stat would stop being comparable across years. LI is exactly symmetric — LI(d, r) = LI(−d, r) — because both teams face identical stakes; the shipped table is asserted against that on every regeneration.
Clutch = WPA − cnWPA. Context-neutral WPA depends only on the round's point margin, derived as (WP(v, 23) − 0.5) × 0.7579047422 — a tied match at the opening bell as the reference state, rescaled to exactly average leverage. Symmetric by construction: winning by 2 is worth exactly the negative of losing by 2.
| Round margin | -4 | -3 | -2 | -1 | 0 | +1 | +2 | +3 | +4 |
|---|---|---|---|---|---|---|---|---|---|
| cnWPA | -0.221511 | -0.174726 | -0.121034 | -0.061960 | +0.000000 | +0.061960 | +0.121034 | +0.174726 | +0.221511 |
Edge cases. Scheduled rounds vary by match (24 for most 2026 matches; matches 7 and 48 were 21), and r is measured against that match's own count. Disqualification rounds are excluded from fighter-level LI and Clutch — they already produce zero WPA, so counting their leverage in the denominator would distort both stats; the round still carries an LI for display on the match page. This affects 8 rounds, so a fighter's LI round count can be lower than their WPA round count. Match 14's post-match round 25 is excluded entirely. Rounds scored 0-0 count normally: they carry real leverage and a small, real WPA. Leaguewide, Σ cnWPA and Σ Clutch are both 0.000.
Comebacks & Blown Leads
How close the winner came to losing — and how close the loser came to winning.
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.
Adjusted NPPR & Schedule
Beating the best fighter in the league is harder than beating the worst. These two stats are the only ones here that know the difference.
Beating the best fighter in the league is harder than beating the worst. Every other stat on this site scores them the same.
Net points per round counts what you did. It says nothing about who you did it to. Two fighters can post the same number against completely different opposition, and the box score will never tell you which one had the harder season.
These two stats fix that from opposite ends. Strength of Schedule tells you who a fighter had to face. Adjusted NPPR re-scores the fighter with that already taken into account.
See them live on the adjusted ratings leaderboard and on every fighter profile.
Take everyone a fighter faced. Average how good they were, using their net points per round. That is the fighter's Strength of Schedule.
Above zero means a harder schedule than most. Below zero means an easier one. Opponents you met four times count four times as much as one you met once.
This season the hardest schedule went to Stacia Suttles at +0.772. The easiest went to Michael Bracamontes at -1.010. That is a gap of 1.78 points a round between two fighters in the same league.
This is the part that makes the number trustworthy, so it is worth spelling out.
Beat someone four times and you drag their numbers down. Now average your opponents' numbers and your own beatings come back to bite you: your schedule looks weak because you were good. The better you fight, the easier your season would look.
So before we measure how good your opponent was, we throw out every round they fought against you. We only look at how they did against everybody else.
This matters more in TBL than in most leagues. About half of all match-ups repeat inside the same match, so a big chunk of your opponent's season is rounds against you.
Here is the proof it works. How good a fighter is should have nothing to do with how hard their schedule looks. So this number should sit at zero:
| How we measure the schedule | How much the fighter's own skill leaks in |
|---|---|
| Just average the opponents | −0.569 |
| Throw out one round | −0.344 |
| Throw out every shared round | −0.034 |
Zero would be perfect. The bottom row is worked out live from this season, not typed in. If it ever slides back toward −0.5, something has broken.
Adjusted NPPR takes a different route. Instead of scoring fighters and then patching the result, it works out everybody's rating together.
Think of it this way. Every round tells you something about both fighters at the same time. Winning by three says you were good and says your opponent was not. Do that across all 1,305 rounds of the season at once, and you get one set of ratings that fits the whole year.
It does not blow up the leaderboard. Most fighters land close to where their raw numbers put them. The ones who move are the interesting part.
Zaire Jefferson is the clearest example. His raw net points per round is -0.615, which reads like a losing fighter. Account for who he was in there with and he sits at -0.107 — right around league average. The box score was describing his schedule, not him.
Most sites give you a rating and let you assume it is exact. Here is how solid ours actually is.
We rebuilt the whole thing 200 times, each time on a slightly different version of the season, and watched how much each fighter's rating moved. A typical rating wobbles by about 0.20 either way. The ratings themselves are spread about 0.39 apart.
So there is roughly 2.0 times as much real signal as wobble. That is genuinely useful — but it is not precise, and we would rather say so.
The rule that follows: gaps smaller than about 0.20 do not mean anything. A fighter three spots up the leaderboard is often not actually better. Every rating on the site comes with a range; where two fighters' ranges overlap, treat them as tied.
On the leaderboard, a ⚠ marks any fighter whose rating wobbles more than 0.25 — those are the softest numbers on the page.
We publish a pound-for-pound list as well as one per division. That sounds like the shakier of the two, and we expected it to be. It is not.
Comparing across weight classes leans on the fighters who changed class during the season — they are what ties the divisions to a common scale. When we rebuilt the ratings over and over, though, the pound-for-pound order held up better than the within-division order: 0.87 against 0.83, where 1.00 means it never moved.
The reason is simple. Fighters inside one division are closer in quality, so it takes less to flip two of them.
One warning we are not going to bury. Rebuilding the season over and over catches random luck. It does not catch a whole division being mis-set. If fighters who move up in weight tend to be the better ones, an entire division could sit too high or too low and none of our checks would notice. One season of data cannot settle that.
Field size matters too, which is why the division menu shows how many fighters qualified. Lightweight has 19. Bantamweight has 3. Leading a field of 3 should be read as exactly that.
- They ignore when a round happened. A round won with the match on the line counts the same as one won in a blowout. That is what Stakes and Timing are for.
- They pull thin records toward the middle on purpose. A fighter with 11 rounds will read closer to average than their raw numbers suggest. With that little evidence, we would rather under-claim than over-claim.
- A hard schedule is not an achievement. Strength of Schedule describes what happened to a fighter, not how good they are. Mostly it is just who the fixture list put in front of them.
- This season only. Nothing carries over from any other year.
Technical Details — the formal spec
Strength of Schedule. For fighter A, over each opponent X faced k times:X_nppr_excl_A = (X's total net points − X's net points in rounds vs A) ÷ (X's total rounds − k)SOS(A) = Σ k · X_nppr_excl_A ÷ Σ k
An opponent whose round count falls to zero after exclusion is dropped from both the numerator and the weight total. Totals are the published NPPR numerator and denominator, so the stat is defined against the NPPR readers actually see.
Adjusted NPPR. Ridge regression over every round with both fighters identified:net points in the round = (fighter's rating) − (opponent's rating) + errortheta = solve(XᵀX + λI, Xᵀy) with λ = 5. X carries one row per round — not one per fighter-round; duplicating each round from both perspectives would double-count every observation and push the penalty tuning to a degenerate value — with +1 in the fighter's column, −1 in the opponent's, and y the margin from the first fighter's side. Row orientation is irrelevant: flipping a row negates both the ±1 pattern and y, leaving XᵀX and Xᵀy unchanged.
On λ = 5. Ten-fold CV on held-out rounds actually minimizes near LAMBDA 2.0 (MSE 1.677 vs 1.694 at 5.0). 5.0 is chosen deliberately: at 2.0 fighters with fewer than 5 rounds show a WIDER rating spread (0.539) than fighters with 15+ (0.424), which is backwards — small samples should shrink harder. At 5.0 the ordering corrects (0.302 vs 0.356) for about 1% worse prediction error. Worth it for a published leaderboard. Do not re-tune per season without telling the operator.
Solver. XᵀX + λI is a graph Laplacian plus λI — sparse, and strictly positive definite for λ > 0. It is solved by conjugate gradient without ever forming the matrix: each product costs one pass over the round list. The point solve converged in 31 iterations this season. A dense Cholesky factorization is kept as an independent cross-check in the unit tests.
Uncertainty. 200 refits, resampling rounds with replacement, reported as a standard deviation and a 90% interval. The resampler is seeded (20260001) — ratings are recomputed on every cache revalidation, and an unseeded bootstrap would make every published interval flicker between page loads. Rank stability is the mean Spearman correlation between the point-estimate ranking and each refit's ranking; the within-division figure averages over divisions with at least five qualified fighters.
Round universe. Both stats read the same bouts NPPR reads (data.fighterHistory), so DQ rounds are already absent. The ridge additionally requires both fighters to be identified, which drops rounds recorded against an N/A opponent — 1,313 of them qualify in 2026. Those rounds still count toward NPPR and SOS weighting. This season: 1,305 paired rounds across 334 fighters, 124 of them qualified at 10+ rounds.
Why SOS is not aNPPR − NPPR. SOS is NOT derived as aNPPR - NPPR. Ridge shrinks extreme NPPRs toward zero, so that difference is mechanically anti-correlated with NPPR regardless of schedule — measured -0.848, worse than the naive version it would replace. It conflates regression to the mean with schedule strength.
Validation. The pure functions are unit-tested against synthetic leagues (CG against Cholesky, orientation invariance, shrinkage ordering, the head-to-head exclusion firing, bootstrap determinism). The season figures — every reference fighter's SOS and aNPPR, the correlation bound above, the paired-round count — are asserted against live data by the admin validation route. Point estimates are exactly reproducible; bootstrap spreads are stochastic and checked as ranges.
Wins Above Replacement
How many wins a fighter added, compared to the kind of fighter a team could easily find to replace them.
WAR asks a simple question: how many wins did this fighter add, compared to the kind of fighter a team could easily find to replace them?
A WAR of 2.0 means their rounds were worth about two extra wins for the team. An easily-replaced fighter, in exactly those same rounds, would have delivered two fewer.
WAR appears in the Advanced section of every fighter profile.
Three ingredients, all derived from the round-by-round results:
- Scoring rate. A fighter's net points per round (NP/R) — points scored minus points conceded, divided by rounds fought.
- The replacement bar. The scoring rate of a fighter a team could find without trying. We fix it where a whole team of them would win 29.4% of their matches — see below. Score above the bar and you add value; below it and you cost your team.
- Points per win. How many net points it takes to buy one extra win. This is what turns a pile of points into a number of wins.
Put it together: take how far above the bar a fighter scores, multiply by how many rounds they fought, then convert that into wins. In formula form: WAR = (NP/R − Replacement NP/R) × Rounds ÷ Points Per Win.
The obvious answer to “how many points is a win worth?” is the average margin a match is won by — about 8.5 points this season. That answer is wrong, for two reasons.
Flipping a match takes twice the gap. If your team lost by 3, adding 3 points gets you a draw, not a win. You need about 6 to come out the other side.
Most points don’t decide anything. Up by 15, an extra point is worth nothing. Down by 15, the same. Points only turn into wins in matches that are close, and a fighter’s season points land in blowouts as well as nail-biters — so the average point is worth far less than a point in a tight match.
So we use the answer our own win-probability model already gives. It puts the value of a one-point round at average stakes at 0.062 of a win, which makes a win worth 16.1 net points. That is the same number Win Probability Added is built on, so a fighter’s WAR and their WPA are finally denominated in the same wins — and should land within a few tenths of each other, the gap being the replacement cushion.
This changed during the 2026 season. WAR previously divided by the average match margin of 8.5 points, which made every figure about a quarter larger than it should have been. If you have an older WAR number written down, it is not comparable to the one on this site now.
The tempting way to set it is to line every fighter up by scoring rate and take the one a quarter of the way from the bottom. We used to do exactly that, and it put the bar at -1.00 net points per round.
That number is impossible. Feed it to the win-probability model: a whole team of fighters scoring at that rate loses every match by roughly 24 points and wins none of them. Replacement level is meant to describe a fighter a team can find easily, not one who cannot win a single match. The bar landed there because a fighter who appeared in a single round and was knocked out scores −4.00, and enough of those one-off appearances filled the bottom quarter to drag it down.
It also broke the rankings. The credit for being above the bar grows with every round you fight, so with the bar that low, an ordinary fighter who scored no net points at all could out-earn a genuinely productive one simply by fighting more rounds.
So we set it the way baseball does. Baseball does not take a percentile either: it fixes replacement as a league-wide total — 1,000 WAR across 2,430 team-games, which works out to a team of replacement players winning about 29.4% of the time — and spreads it across playing time. We ask our win-probability model the same question: what scoring rate would leave a team winning 29.4% of its matches? The answer is a margin of about 2.7 points over 24 rounds, or -0.1129 net points per round. That is the bar.
The bar and the divisor are used everywhere on the site and do not change between the regular season and the playoffs — only a fighter's own scoring rate and round count do. The other two are shown for comparison and are not used in the formula:
Both numbers are computed live from every decided match and every fighter's bout log — nothing is hand-entered. Disqualification rounds are excluded from fighter statkeeping (and therefore from WAR), though their points still count in match results.
- It doesn't know when points were scored. A garbage-time knockout counts the same as a match-winning one — that's what WPA is for.
- It doesn't adjust for opponent quality — beating the best fighter in the league counts the same as beating the worst. That's what Adjusted NPPR and Strength of Schedule are for.
- Fighters with few rounds can post noisy values, up or down.
Every Stat, Defined
Every number on the site in one place, including the basics.
- Record
- Rounds won and lost (W-L). Each round of a match is a bout between one fighter from each team; the round winner takes the W.
- Win%
- Round wins divided by decided rounds (wins + losses). Draws and disqualification rounds are not counted as wins or losses.
- Rounds
- Total rounds fought. Disqualification rounds are excluded from a fighter’s statkeeping.
- Streak / Form
- Consecutive results, most recent first — W3 means the last three rounds were wins.
- Round scoring
- Winning a round earns points on a fixed scale: a Decision is worth 1 point, a Knockdown (KD) 2, a Double Knockdown (2x KD) 3, and a KO/TKO 4. The match is decided by total points, not round count.
- Points For / Points Against
- Total points a fighter scored across their rounds, and total points their opponents scored on them.
- Net Points
- Points For minus Points Against — a fighter’s total scoring margin.
- NP/R (Net Points Per Round)
- Net Points divided by rounds fought — the scoring-rate stat, comparable across fighters with different workloads.
- Extra Points / Extra Points Allowed
- Points above the 1-point decision baseline, by finish: Decision 0, KD +1, 2x KD +2, KO/TKO +3. Extra Points sums over rounds won; Extra Points Allowed over rounds lost — a measure of finishing power and finishing vulnerability.
- Knockdowns / Double KDs / KO-TKO
- Counts of rounds won by each finishing method.
- KO%
- KO/TKO wins divided by total wins — the share of a fighter’s victories that came by stoppage.
- WAR (Wins Above Replacement)
- How many wins a fighter added compared to the kind of fighter a team could easily find to replace them. Built from their scoring rate, how many rounds they fought, and what a win costs in points. How WAR works →
- Stakes Faced (Leverage Index)
- How much was riding on a round before it started. It belongs to the moment, not the fighter, so both fighters in a round have the same number. 1.00 is an ordinary round; 6.63 is the maximum — a tied match with one round to go. A fighter's average tells you how big their moments were, not how they did in them. How Leverage works →
- Timing (Clutch)
- The part of a fighter's WPA that came from when their results landed. We compare what they actually did to what the same wins and losses would have been worth in ordinary rounds; the difference is Timing, in wins. Positive means their wins landed in the big moments. Disqualifications do not count. Biggest moments →
- WPA (Win Probability Added)
- How much a round moved the team's chance of winning, credited to the fighter who caused it. A round won with the match on the line is worth far more than one won in a blowout. How WPA works → · Leaderboard →
- SOS (Strength of Schedule)
- How good the opponents were that a fighter had to face. Above zero means a harder schedule than most. Rounds an opponent fought against this fighter are thrown out first — otherwise beating someone four times would drag their numbers down and make your own schedule look easy. How it works → · Leaderboard →
- aNPPR (Adjusted NPPR)
- Net points per round, once you account for who a fighter actually fought. Every fighter is rated against every other at the same time, across the whole season. Fighters with few rounds get pulled toward average on purpose. Each rating shows a range; gaps under 0.20 do not mean anything. How it works →
- Comeback Win / Blown Lead
- The lowest the eventual winner's chances ever fell during a match. Dip under 25% and it counts as a comeback win — and, from the other side, a blown lead for the team that let it slip. Draws do not count. How it works → · Biggest comebacks →