The Market Has Ankalaev. Our Model Has Questions.

This is a card that looks straightforward on the surface — a light heavyweight main event between a well-regarded contender and a live opponent, a few interesting prelim matchups, and a lot of short-sample noise underneath. Reasonable enough.

Except our model keeps poking at two things. First: the market has Magomed Ankalaev as an 81 percent shot. We have him at 61. That's a 20-point gap, and it's the biggest trusted divergence on the card. We're not saying Ankalaev loses — our number still has him winning — but the price implies something close to a sure thing, and the model is nowhere near that comfortable. Second: Santiago Ponzinibbio is getting about 23 percent implied odds against Sam Patterson. We have him at 53. That's a 30-plus point gap on a two-fight sample, and it's the kind of number that earns a long, honest look.

That's the tension. A main event where the market is more confident than we are, and a co-main-card welterweight bout where we're considerably more bullish on the guy the market has buried. Both of those situations deserve explanation, not a sales pitch. Let's get into it.

Two fights where the gap between our number and the market price is wide enough to talk about honestly. This is a v0 model — 56.9 percent on holdout fights, not yet validated against the market — so treat these as conversations, not conclusions.

Santiago Ponzinibbio — Our number: 53%. Market: 23%.

Santiago Ponzinibbio is 39 years old with a 31-9 record, which is not a résumé that inspires confidence at first glance. The market clearly sees age and recent results and prices him accordingly. What the model sees is a guy posting 4.76 significant strikes per minute across his last two UFC fights — the highest volume among any fighter on this card with at least two fights of UFC data — with 0.61 knockdowns per 15 minutes and a 58 percent career finish rate. His opponent, Sam Patterson, throws 3.05 strikes per minute, has a takedown rate of 0.34 per 15 minutes, and has a career finish rate of 100 percent, which sounds good until you realize it's built on three fights. The model sees a volume striker with real pop against a slower opponent on a limited sample. The market sees Ponzinibbio's age and history. One of them is overweighting something. We just can't tell you which.

Bogdan Guskov — Our number: 39%. Market: 19%.

Bogdan Guskov is a 20-point divergence in a main event, which is notable. The market has him at roughly 19 percent implied. We have him at 39. This is based on three UFC fights of data, so it's not a deep sample, but what's there is legitimately interesting: 5.14 significant strikes per minute — more volume than Ankalaev's 4.06 — and 0.96 knockdowns per 15 minutes, which is the highest knockdown rate among any fighter with UFC data on this card. A 100 percent career finish rate across those fights. The model isn't saying Guskov wins this fight. It's saying the gap between 19 and 39 is the kind of gap that warrants an honest look at whether name value and Ankalaev's overall résumé are doing more work in that market price than the recent numbers justify.

These are the fighters our model most believes win — ranked by the combination of model confidence and how the style rates support the outcome. Most believed first.

1. Magomed Ankalaev vs. Bogdan Guskov — Model: 61%. As noted above, the market is at 81% and we're at 61%, so we're less certain than the price. But 61 is still a meaningful lean. Ankalaev posts 4.06 sig strikes per minute with 0.9 grappling advances per 15 — a more complete fighter than Guskov on paper. Guskov's volume (5.14 per min) and knockout rate (0.96 per 15) make him genuinely dangerous in a stand-up exchange, but Ankalaev's ability to mix in the grappling dimension and control distance across five rounds is the edge the model respects. It's a five-round fight, and that matters. We believe Ankalaev wins. We just don't believe it at 81%.

2. Valter Walker vs. Thomas Petersen — Model: 58%. This is the smaller divergence on the card (we're at 58, market at 62, so a 3.5-point gap in the market's direction), and our conviction is moderate rather than strong. But the style rates earn Walker the top of the trusted board. He posts 4.4 significant strikes per minute against Petersen's 2.34 — nearly double the volume. He has 0.6 knockdowns per 15 minutes and a 100 percent career finish rate across three UFC fights. Petersen brings a southpaw stance and 2.59 takedowns per 15 minutes, which is a real wrinkle — if he can get Walker to the mat, the fight changes. But Walker's superior output and finishing ability make him the more compelling profile here.

3. Santiago Ponzinibbio vs. Sam Patterson — Model: 53%. The divergence is enormous (30.6 points) but the model confidence is modest — 53 percent. This is the honest reality of a big gap: the model sees more than the market does, but not a lot more. The argument is volume and striking rate, as outlined above. The concern is sample size on both sides and Ponzinibbio's age. If this is live, it's not because we're certain. It's because the price implies near-certainty in the other direction, and the numbers don't support that.

Two categories of caution here: trusted fights where the market favorite's profile worries the model, and low-data bouts where the model is essentially flying blind.

Market favorites the model doesn't fully trust

Sam Patterson is priced at 77 percent implied in our data. Our model has him at 47. That's a 30.6-point negative divergence — the largest on the card. Patterson's UFC sample is three fights, his striking volume is the lowest among any welterweight on this card at 3.05 per minute, and while his 100 percent finish rate sounds impressive, it's a small-sample number. The market appears to be doing a lot of work based on his record and perhaps recent performances that may not fully show up in these rates. Whether that's right or wrong, the model flags it.

Magomed Ankalaev at 81 percent market implied versus our 61 percent is covered above. Worth repeating as a caution: the model isn't fading him, it's just not comfortable with that price.

Low-data bouts — the model is guessing, and so is everyone else

Eleven fights on this card fall into the low-data category, meaning at least one fighter has zero or near-zero UFC-level data in our system. The model's numbers on these are not meaningfully different from priors. Anyone claiming a strong edge here — including us — is working on incomplete information. These fights are: Muhammad Said vs. Dustin Jacoby, Ramazan Temirov vs. Steve Erceg, Wellington Turman vs. Islam Dulatov, David Martinez vs. Umar Nurmagomedov, Damian Rzepecki vs. Magomed Zaynukov, Axel Sola vs. Ismael Bonfim, Brendson Ribeiro vs. Magomed Tuchalov, Tyrell Fortune vs. Rizvan Kuniev, Saygid Izagakhmaev vs. Abubakar Vagaev, Mike Davis vs. Nurullo Aliev, and Abdul Hussein vs. Cody Gibson. On these, defer to the market. It has more signal than we do.

The fights that are most interesting to watch as style stories — using the actual rate numbers to explain how each one is likely to play out.

Ankalaev vs. Guskov — Volume vs. Control (5 rounds)

Guskov brings the higher volume at 5.14 significant strikes per minute versus Ankalaev's 4.06, and his 0.96 knockdowns per 15 minutes is real — he finishes people on the feet. But this is a five-round fight, and Ankalaev's profile is more complete. Ankalaev mixes 0.9 grappling advances per 15 with his striking, giving him a second dimension Guskov essentially lacks (0.45 per 15). Guskov needs to make it a firefight and land something clean early. Ankalaev needs to use his grappling as a control tool and manage the pace. The longer this stays upright and frantic, the better Guskov's chances. The more Ankalaev varies the attack, the more the fight plays to his strengths.

Walker vs. Petersen — Volume Striker Meets Southpaw Wrestler

Walker at 4.4 strikes per minute with a 100 percent finish rate is a real problem for most opponents. Petersen's counter is the wrestling: 2.59 takedowns per 15 minutes is a genuine threat, and the southpaw stance creates angles that can disrupt Walker's rhythm. Petersen's own striking volume of 2.34 per minute won't win a stand-up exchange with Walker — he needs the level changes. Walker's 0.6 knockdowns per 15 minutes means he doesn't need many clean looks to change the fight. This one likely hinges on whether Petersen can establish the wrestling early or if Walker puts him in danger before that game plan takes hold.

Ponzinibbio vs. Patterson — High Volume vs. Pure Finisher

Ponzinibbio at 4.76 strikes per minute with 0.61 knockdowns per 15 is the aggressor here. Patterson's 3.05 per minute means he's the slower, more selective striker — but his 100 percent finish rate suggests he's dangerous when he lands. The reach advantage goes to Patterson at 78 inches versus Ponzinibbio's 73, which matters in a striking-dominant fight. Ponzinibbio's volume means Patterson will be busy all night; Patterson needs to find his spots and make them count. This is the kind of fight where pace and output often decide more than style.

Umar Nurmagomedov vs. David Martinez — Elite Wrestler Meets Striker

This is a low-data bout, so treat these numbers with appropriate skepticism — Nurmagomedov has three UFC fights of data, Martinez has three. That said: Nurmagomedov's 3.41 takedowns per 15 minutes is the highest grappling rate among any fighter with meaningful data on this card. At 3.42 strikes per minute he's not a slouch on the feet either. Martinez posts 4.56 strikes per minute and 0.47 knockdowns per 15 — a more active striker. If this stays standing, Martinez has the output edge. If Nurmagomedov gets the fight to the mat, he controls the rate and the round. The central question is where the fight lives.

Said vs. Jacoby — The Debutant Problem

Said is 9-0 and making his UFC debut — the model has zero UFC-level data on him. Jacoby at 38 has the experience edge and two fights of data in our system. Said's record and age (30) suggest he's not a raw prospect, but we genuinely cannot model what he brings at this level. Jacoby's 70 percent career finish rate is noted. Beyond that, defer to people who've watched the tape.

Temirov vs. Erceg — Orthodox Mirror, Different Finisher Profiles

Both fighters are 30, both orthodox, and the model splits them 50-50. Erceg at 4.81 strikes per minute is the higher-volume guy, while Temirov posts 4.22. Erceg's 0.15 knockdowns per 15 and 25 percent finish rate paint a picture of a fighter who wins on output and decisions. Temirov's 0.45 knockdowns per 15 and 45 percent finish rate suggest slightly more danger. Neither profile screams obvious path to victory. This one may well go to the cards.

A few honest notes on where scoring volume concentrates on this card, using the projection and ceiling numbers from the model.

Ceiling plays

The fighters with the highest DFS ceilings on the card among those with meaningful data: Bogdan Guskov at 116.9, Umar Nurmagomedov at 120.1, and Valter Walker at 112.6. Ceiling in DFS is driven by finishes — Guskov's 100 percent finish rate and elite knockdown rate (0.96 per 15) make him the obvious high-variance play in the main event. He either gets stopped or he stops someone; there's not much decision energy in that profile. Nurmagomedov's ceiling comes from a different direction — 3.41 takedowns per 15 minutes generates DFS points through control and ground activity, not just finishes.

Volume and projection

Walker leads the card's trusted-data fighters in DFS projection at 55.7, supported by his 4.4 strikes per minute and grappling activity (1.8 advances per 15). Ankalaev projects at 54.5 — solid but not explosive, consistent with a fighter who wins rounds rather than hunts finishes (50 percent career finish rate). Guskov at 45.6 is the lower projection with the higher ceiling — that's the classic boom-or-bust DFS profile, and it fits the fight.

Low-data ceiling caveats

Several low-data fighters post enormous ceilings in the 119-122 range — Magomed Zaynukov, Nurullo Aliev, Islam Dulatov, and others. These numbers are driven by model priors rather than observed rates. High ceiling on zero data is noise, not signal. Build around the fighters whose ceilings are grounded in actual UFC performance.

Projected scoring leaders worth noting

Dustin Jacoby (55.8 projection, 116.6 ceiling) and Axel Sola (55.5 projection, 118.1 ceiling) lead the low-data contingent on raw projections. Both have small UFC samples backing the numbers up partially. Sola's 4.23 strikes per minute with a southpaw stance and 1.78 takedowns per 15 makes for an active profile; Jacoby's 70 percent finish rate provides ceiling support. Both are low-data caveats, but they're not operating purely on priors either.

Three things worth carrying into Friday night.

1. Ankalaev is probably winning, but the price is doing a lot of work. Our number says 61 percent. The market says 81. That 20-point gap is the most meaningful disagreement between the model and the market on a trusted, data-grounded fight. The honest read is that Ankalaev is the right side — the grappling dimension, the experience (14 UFC fights versus Guskov's five), the five-round format. But Guskov's 5.14 strikes per minute and 0.96 knockdowns per 15 minutes are not decorative statistics. He hits hard and he hits often. If Ankalaev lets this turn into a firefight, Guskov is genuinely capable of making something happen. The price implies that almost can't occur. The model is less sure.

2. The Ponzinibbio situation deserves a real look, not a dismissal. A 30.6-point gap between our number (53%) and the market (23%) is the largest trusted divergence on the card. The model is not highly confident — 53 percent is a coin flip with a slight lean — but the market is pricing near-certainty in Patterson's direction on a three-fight UFC sample for a guy posting 3.05 strikes per minute. Ponzinibbio at 4.76 per minute with real knockdown rate is not a corpse. Age is a real factor and the sample on both sides is small. This is a conversation, not a conclusion. But it's the most interesting one on the card.

3. Eleven fights are low-data. Treat them as such. More than half this card is operating below the threshold where the model has meaningful conviction. That's not unusual for a Fight Night card, but it's worth being clear-eyed about. On those fights, the model defers to the market because the market has seen the tape and we haven't. The three fights with real data — Ankalaev-Guskov, Walker-Petersen, Patterson-Ponzinibbio — are where the analysis is grounded. Everything else is working on incomplete information, and anyone who tells you otherwise is overselling their edge.