Fight Genie logo Fight Genie
Back to the card

Tai Tuivasa vs Robelis Despaigne

Heavyweight · Prelims · UFC 331: Van vs. Pantoja 2

Tai Tuivasa
Tai Tuivasa
Age 33 · Elo 1508
LLLLL
231 days since last fight
vs
Robelis Despaigne
Robelis Despaigne
Age 38 · Elo 1484*
LWL
875 days since last fight

Who wins, by source

Tai TuivasaRobelis Despaigne
Claude Claude: Tai Tuivasa 42% Claude: Robelis Despaigne 58% 42% 58% GPT GPT: Tai Tuivasa 44% GPT: Robelis Despaigne 56% 44% 56% Market Market: Tai Tuivasa 19% Market: Robelis Despaigne 81% 19% 81% Elo Elo: Tai Tuivasa 53% Elo: Robelis Despaigne 47% 53% 47%

Each row splits one source's view of the fight. The dashed line is even odds. Where the models sit further from the market, they are claiming an edge.

Tale of the tape

Tai TuivasaRobelis Despaigne
Strikes landed / min Tai Tuivasa Strikes landed / min: 3.37 3.37 Robelis Despaigne Strikes landed / min: 2.11 2.11 Strikes absorbed / min Tai Tuivasa Strikes absorbed / min: 4.61 4.61 Robelis Despaigne Strikes absorbed / min: 2.54 2.54 Striking accuracy Tai Tuivasa Striking accuracy: 48% 48% Robelis Despaigne Striking accuracy: 59% 59% Striking defense Tai Tuivasa Striking defense: 42% 42% Robelis Despaigne Striking defense: 46% 46% Takedowns / 15 min Tai Tuivasa Takedowns / 15 min: 0.00 0.00 Robelis Despaigne Takedowns / 15 min: 0.00 0.00 Takedown defense Tai Tuivasa Takedown defense: 58% 58% Robelis Despaigne Takedown defense: 50% 50% Sub attempts / 15 min Tai Tuivasa Sub attempts / 15 min: 0.0 0.0 Robelis Despaigne Sub attempts / 15 min: 0.0 0.0

Career UFC averages. Bars are scaled within each row, so compare left against right, not row against row. The stronger side is highlighted, and for strikes absorbed the lower number is the better one.

How it ends

KO/TKOSubmissionDecision
Claude Claude: KO/TKO 55% 55% Claude: Submission 5% Claude: Decision 40% 40%GPT GPT: KO/TKO 62% 62% GPT: Submission 2% GPT: Decision 36% 36%

Each model's split across the ways this fight can finish, out of 100.

The analysis

Claude
Robelis Despaigne by KO/TKO, 58% confidence

The market prices Despaigne at 79% and I disagree by roughly 19 points; my read is closer to 60-40. The factor the market appears to be underweighting is Despaigne's 687-day layoff, which lands squarely beyond the 450-day risk threshold, stacked on top of a documented cardio problem. In his two UFC losses (Cortes-Acosta, Austen Lane) Despaigne failed to land the early bomb and then faded badly, landing just 2.11 significant strikes per minute across his UFC run, the lowest output of any fighter on this card. His 5-2 record is essentially a first-minute finish rate: when the fight passes the 5-minute mark he is 0-2 in the UFC. Tuivasa is on a five-fight losing skid and his 42% striking defense is a genuine liability against an 84-inch reach (a 9-inch disadvantage for Tuivasa), which is why Despaigne remains the pick. But Tuivasa's last two losses were decisions to Sutherland and Teixeira where he went the distance, not knockouts; his last KO loss was to Pavlovich in 2022. His 58% takedown defense is irrelevant here since Despaigne has zero UFC takedown attempts. Southpaw-versus-orthodox with Despaigne's length favors the taller man early, and Tuivasa's habit of eating the first clean shot is what gives Despaigne a real KO window in round one. If the fight leaves round one, it flips: Tuivasa's body work and pressure against a gassed opponent is how Cortes-Acosta and Lane won. The Elo (Tuivasa 53%) is provisional on Despaigne's side so I do not lean on it heavily, but it directionally supports the view that -530 is too steep. Tuivasa at +360 is the value play on the card regardless of who is favored to win outright.

GPT
Robelis Despaigne by KO/TKO, 56% confidence

Despaigne's nine-inch reach advantage and higher striking accuracy favor him against Tuivasa's 42% striking defense and 4.61 significant strikes absorbed per minute. Tuivasa's higher output and supplied win rate make this a narrow lean. Neither fighter records offensive grappling production, suggesting a striking-led contest. Missing meaningful finishing data limits confidence in the method and round.