Rosas Jr.vsBarcelos

CLAUDEThis is a genuine coin-flip main event where the two priors point in opposite directions. Elo gives Barcelos a 54-46 edge (1613 vs 1584) and that rating is earned: his current five-fight run over Talbott, Garbrandt, Simon and a split over Montel Jackson is the strongest strength-of-schedule streak of any unranked bantamweight, and he did it by neutralizing wrestlers (Simon) and out-grappling a hyped prospect (Talbott). His 85% takedown defense and 5.22 SLPM at 51% accuracy are the exact tools that historically trouble Rosas, whose 1.55 SLPM is by far the lowest on this card and confirms he cannot win rounds standing. The case for Rosas is the variable Elo does not encode: age. Barcelos turned 39 in May, Rosas is 21, and this is scheduled for 25 minutes, a distance neither man has fought in the UFC. Rosas attempts 4.02 takedowns per 15 minutes; extrapolated to five rounds that is roughly 12-14 attempts, and even at a 30% conversion against elite defense that yields four or more takedowns with control time in a fight where Barcelos will also be absorbing chain-wrestling pressure on legs that have already seen 27 pro fights. Rosas's unanimous decision over Rob Font in March showed he can impose grappling on an experienced, defensively sound opponent and stay disciplined for 15 minutes without being drawn into a firefight. Rounds 1-2 likely go to Barcelos on volume; the fight turns on whether Rosas's pace degrades Barcelos's takedown defense in rounds 3-5. I disagree with Elo here by a modest margin and take Rosas, but this is a low-confidence pick where a Barcelos decision is entirely live.

GPTBarcelos' 85% takedown defense could blunt Rosas' primary offensive route, while his higher striking output, accuracy and defense favor him in standing exchanges. Rosas' takedown activity and submission threat keep this competitive. Round 5 assumes a five-round main event; the supplied data does not confirm the scheduled distance. Method probabilities are conditional on Barcelos winning.

CLAUDEElo slightly favors Dumont (1580 vs 1569, 52-48) but the recent trend and common-opponent evidence both point to Perez. Perez's +55 Elo trend is the steepest on the card; Dumont's is +11 and she is coming off a unanimous decision loss to Joselyne Edwards, the same fighter Perez beat by unanimous decision in June 2024. Both women beat Karol Rosa and Ashlee Evans-Smith, so the Edwards result is the only differentiator among three shared opponents and it favors Perez. Stylistically this is Dumont's technical striking (3.86 SLPM, 65% defense, only 2.16 absorbed) against Perez's relentless wrestling at 4.9 takedowns per 15 minutes on 49% accuracy. Dumont's 72% takedown defense is respectable, but Perez just took down and controlled Macy Chiasson, a much bigger and stronger grappler than Dumont, for three rounds in February. Dumont has the size (5'7 vs 5'5, former featherweight) and the cleaner striking to win rounds at range, and if she stuffs the first two or three takedown attempts in each round the fight tilts her way. The historical concern with Perez is her round-three fade, visible in the Rosa fight, which gives Dumont a late window. On balance Perez's takedown volume against a 72% defender should produce roughly one to two takedowns per round with top control, enough to win two of three on judges' cards that reward control. Dumont at 35 with a loss in her last outing, versus Perez at 31 on a five-fight streak with the shared-opponent edge, is enough to override the marginal Elo gap.

GPTPerez earns a narrow edge through her substantially higher takedown rate and favorable supplied common-opponent results. Dumont's 72% takedown defense, higher striking volume and stronger striking differential make this competitive. Perez needs successful wrestling exchanges to offset Dumont's striking advantage. Low submission activity supports a decision; method percentages are conditional on a Perez victory.

CLAUDEHernandez is a UFC debutant with no UFCStats data, so every statistical field for him is null rather than a measured skill and this analysis leans on his 8-0 record with a rear-naked-choke win on DWCS 10.6 only eleven days ago, which makes this an extremely short-notice booking. That turnaround cuts both ways: no camp specific to Dumas, but he is fresh, weight-managed and just won. The reason to favor him anyway is how specifically Dumas's failure mode matches Hernandez's profile. Dumas has lost four of his last five completed fights, all by finish inside two rounds: D'Arce choke (R1) to McVey, guillotine (R2) to Donte Johnson, KO (R1) to Oleksiejczuk, KO (R1) to Ruziboev. His 40% takedown defense is the worst on this card, and the two most recent losses were front-headlock chokes surrendered when he shot or got scrambled on. Hernandez is a submission grappler with a 100% finish rate in his most recent bout and the stylistic tag that has beaten Dumas twice in a row. Dumas's physical tools remain legitimate at 6'2 with a 79-inch reach and he has early-round power and athleticism; a fresh Dumas landing a big right hand in the first three minutes on an undersized debutant is the realistic upset path, and the move up to light heavyweight from middleweight may mean he is not draining himself. But his 2.15 SLPM shows he does not use that reach to build volume, and once the fight hits the mat his history says he does not get up. Confidence is capped at 60 given the debutant uncertainty and the eleven-day turnaround.

GPTHernandez's substantially higher striking output and positive strike differential favor him in sustained exchanges. His takedown activity also threatens Dumas's 40% takedown defense, creating opportunities for ground strikes or submissions. Dumas's reach advantage and stronger striking defense offer resistance, while Hernandez's high absorption rate creates countering opportunities. Unspecified sample sizes limit confidence in Hernandez's extreme statistics.

CLAUDEThis is a double UFC debut at bantamweight with no UFCStats history, no Elo rating, and no market or community data supplied, so the projection rests almost entirely on the two men's pre-UFC records and regional pedigree. Osmanli arrives at 13-1, a 93% career win rate that reflects an unusually clean run through the Caucasus/Eastern European regional circuit; Akylbek is 11-3, a 79% clip, with three losses that suggest he has been beaten by comparable-tier opposition more than once. Both fighters come from grappling-heavy fight cultures (Azerbaijan and Kyrgyzstan respectively), which points toward a wrestling-and-clinch-centric bout where positional control, not knockdown power, decides rounds. In such a matchup the fighter with the fewer career losses and the more consistent record is the marginally safer pick, but the edge is small: regional records are inflated by matchmaking, and neither man has faced UFC-calibre depth. Debutant nerves, an unfamiliar Apex octagon, and a full camp cut to 135 for a first UFC weigh-in add variance for both. I project Osmanli to win more of the grappling exchanges and take a competitive decision, with a meaningful chance the bout goes the other way or ends in a scramble submission. Confidence is deliberately capped given zero UFC-derived data on either side.

GPTThe supplied data establishes no advantage for either fighter. Osmanli is an arbitrary tiebreak pick, with confidence set to the requested minimum rather than an evidence-supported edge. Decision assumes a three-round bout; method probabilities are illustrative estimates.