The Model Thinks Usman Has More Left Than Vegas Does

Most fight cards have a main event and a backdrop. This one has a main event and a genuine philosophical argument between our model and the betting market — and the argument is worth having out loud before a single punch gets thrown.

Dricus Du Plessis is a significant favorite to defend his middleweight title against Kamaru Usman. The market has Usman at roughly 33% to win. Our model says 54%. That is a 21-point gap, and on a main event with 19 UFC fights of tracked history on Usman's side, it is the kind of divergence you should at least think about rather than wave off. We will get into the why in a moment.

Beyond the main event, the card is honestly a mixed bag of data quality. Ten of the thirteen bouts carry a low-data flag — the model is working with thin samples or outright debutants in most of those, and we will say so plainly. Where we have conviction, we will tell you. Where we are guessing along with everyone else, we will tell you that too.

Two divergences on trusted bouts cleared the threshold worth discussing. Both involve fighters the market may be ready to move on from before the model is.

Kamaru Usman — Our number: 54%. Market: 33%. Gap: +21 points.

The market is pricing Kamaru Usman like a former champion on his way out the door. At 39, with a recent layoff, that instinct is not crazy. But the model is looking at 19 UFC fights of Usman history and seeing a fighter who, on the numbers, still brings 3.71 significant strikes per minute, 1.61 takedowns per 15, and 1.35 grappling advances per 15 — a complete, multidimensional profile that does not collapse quietly. The market may be heavily weighting age and narrative (the decline arc, the losses to the elite). The model is weighting what the numbers say he actually does inside the cage. Whether that gap reflects a real edge or just a model that has not fully priced in physical erosion at 39 — that is the honest question. Our number says 54, the market says 33. One of them is more wrong than the other.

Jared Cannonier — Our number: 47%. Market: 27%. Gap: +20 points.

Same pattern, different fight. Jared Cannonier is 42 years old with an 18-9 record, and the market has him at 27% against Christian Leroy Duncan. The model says 47%. Cannonier carries 3.87 significant strikes per minute, 0.72 knockdowns per 15 — genuine finishing power — and a 64% career finish rate across the sample. Duncan is a legitimate talent at 30 with a 71% finish rate himself, and the market clearly prefers him at 73%. But the model sees a closer fight than a roughly 3-to-1 underdog price implies. Again: name the tension, not a pick. Our number says near coin-flip. The market disagrees substantially.

These are the fighters the model most believes win, in order of confidence, with the style story behind each number.

1. Brad Tavares over Marc-Andre Barriault — Model: 59%. No market price available for direct comparison, but this is the model's clearest lean on the card. The style read is straightforward: Tavares at 3.56 strikes per minute is the lower-volume man, but Barriault at 4.2 per minute with a 67% finish rate and a 41% model win probability is not the threat his finish rate implies when you look at the broader picture. Tavares has been around forever (21-13, 38 years old) and his 0.13 career finish rate tells you exactly how his wins tend to go — decisions, accumulated over a long night. Barriault will be active, will look for the finish, but the model gives the edge to Tavares to grind it out. Worth noting: neither man is a spring chicken, and both are orthodox with identical reach. This one lives and dies on output and conditioning over three rounds.

2. Kamaru Usman over Dricus Du Plessis — Model: 54%. Already covered the gap above, but the conviction piece is worth naming: the model sees matched volume (Usman 3.71 strikes/min, Du Plessis 3.80), a meaningful wrestling edge for Usman (1.61 takedowns per 15 vs Du Plessis at 0.98), and Usman generating more grappling advances (1.35 vs 0.90). Du Plessis is the bigger finisher at 67% career finish rate vs Usman's 31%, so if Du Plessis lands something clean, he converts. But the model thinks Usman's path — control the wrestling, mix levels, neutralize Du Plessis's finishing opportunities over five rounds — is more viable than a 33% price suggests.

3. Christian Leroy Duncan over Jared Cannonier — Model: 53%. The model does believe Duncan wins — just not as confidently as the 73% market price implies. Duncan at 4.12 strikes per minute with 0.71 knockdowns per 15 and a 71% finish rate across four UFC fights is a live, dangerous middleweight at 30. The model simply thinks Cannonier's power and experience keep this competitive enough that 73% feels steep. Duncan is probably the right side. The model just does not love the price.

Two kinds of traps on this card: market favorites the model is cooler on in trusted bouts, and the low-data fights where anyone claiming a real edge — including us — is mostly bluffing.

Model is cooler on these trusted-bout favorites:

• Dricus Du Plessis — market 67%, model 46%. The finishing ability is real, the title is real. But the wrestling exposure against a fully engaged Usman is the thing the model keeps coming back to. Du Plessis at 0.98 takedowns per 15 and 0.90 grappling advances per 15 is not going to out-wrestle Usman. If Usman makes it a grind, the model thinks Du Plessis's path to a finish narrows considerably.
• Christian Leroy Duncan — market 73%, model 53%. Duncan is talented and the model does think he wins. Twenty points of difference on a three-round fight is just a lot of certainty to assign. The Cannonier power at 0.72 knockdowns per 15 is a real variable at any age.

Low-data bouts — the model is essentially blind here:

Ten fights on this card carry the low-data flag. In several cases we are working from one UFC fight of history, or zero. We will list them plainly: Jose Miguel Delgado vs Austin Bashi, Seokhyeon Ko vs Jean-Paul Lebosnoyani, Felipe Franco vs Levi Rodrigues Jr., Mitch Ramirez vs Chase Hooper, Jacobe Smith vs Kevin Holland, Fatima Kline vs Tabatha Ricci, Stewart Nicoll vs Alden Coria, Alvin Hines vs RJ Harris, Alberto Montes vs Tommy McMillen, and Dione Barbosa vs Anna Melisano.

On all ten, defer to the market price as your baseline. We can describe what little the numbers show, but calling any of those a model-driven lean would be dishonest. We flagged a few divergences in that group that exist in the data, but with one UFC fight — or zero — of sample, those numbers are noise dressed up as signal.

Du Plessis vs. Usman — The Grind vs. The Finisher

Same reach (76 inches each), both switch-stance, nearly identical striking volume (Du Plessis 3.80 sig strikes/min, Usman 3.71). On the feet, this is a dead heat. Where it separates: Usman at 1.61 takedowns per 15 and 1.35 grappling advances per 15 is the significantly more active wrestler. Du Plessis at 0.30 knockdowns per 15 and a 67% career finish rate is the guy who wins by putting you to sleep when the opening comes. The tactical question is whether Usman can convert his wrestling volume into enough control to steer the fight away from Du Plessis's best moments. If this stays on the feet for five rounds, Du Plessis's finishing instinct becomes the dominant factor. If Usman drags it to the mat and makes it ugly, the model thinks the map changes.

Duncan vs. Cannonier — Young Finisher, Old Power

Christian Leroy Duncan is 30, with a 79-inch reach, 4.12 strikes per minute, 0.71 knockdowns per 15, and a 71% career finish rate across four UFC fights. He is long, active, and dangerous. Jared Cannonier is 42, 77-inch reach, 3.87 strikes per minute, 0.72 knockdowns per 15, and 64% finish rate. Here is the thing about Cannonier: his knockdown rate and finish rate do not look like a 42-year-old in decline on paper. Duncan has a reach edge and a meaningful age edge. But Cannonier is the kind of fighter who makes one mistake very expensive, and that variable alone is why the model does not share the market's confidence in Duncan at 73%.

Chase Hooper vs. Mitch Ramirez — The Drag-Down Artist

Chase Hooper at 4.27 takedowns per 15 minutes is one of the more relentless grapplers on the card. His 2.80 strikes per minute tells you he is not trying to win on the feet — he is trying to get you down and keep you there, riding out to a 75% career finish rate. Mitch Ramirez at 4.40 strikes per minute wants nothing to do with that scenario and will need to keep this upright. Low-data caveat applies here — one UFC fight of sample on Ramirez — but the stylistic blueprint is about as clear as it gets on this card: Hooper wants the floor, Ramirez needs the fence. Market has Hooper at 75%, model at 61%. Not a divergence that moves the needle much; both like Hooper.

Jacobe Smith vs. Kevin Holland — Unbeaten vs. Experienced

Jacobe Smith is 12-0, 30 years old, with 6.24 significant strikes per minute — that is genuine volume — 3.73 takedowns per 15, 1.0 knockdowns per 15, and a 100% career finish rate across three UFC fights. That profile is eye-catching. Kevin Holland is 33, 81-inch reach (nine inches on Smith), 4.25 strikes per minute, 0.43 knockdowns per 15, and a 73% finish rate across six UFC fights — a veteran who has seen every kind of fight. Holland's reach advantage is the most tangible physical edge in this matchup. Smith's volume and finishing profile are the counter-argument. Low-data flag applies (three UFC fights for Smith, one meaningful sample fight for his rates), so treat this as a style story, not a firm model lean.

Tavares vs. Barriault — The Decision Machine vs. The Finisher Who Needs a Finish

This is the simplest stylistic read on the prelim slate. Brad Tavares has a 0.13 career finish rate — he essentially never finishes fights. Marc-Andre Barriault has a 67% finish rate and needs to use it, because the model gives him 41% to win. If this goes to the scorecards, Tavares has been grinding out decisions his entire career and you are betting on a known pattern. Barriault at 4.20 strikes per minute is the more active man, but Tavares at 3.56 is controlled and efficient. Same reach, same stance. Barriault needs a finish to beat the model. Tavares needs fifteen minutes.

Kline vs. Ricci — The Striker/Grappler Hybrid vs. The Decision-Maker

Fatima Kline at 5.59 significant strikes per minute and 2.16 takedowns per 15 is a genuinely well-rounded profile for strawweight — high volume, active on the mat, 67% finish rate. Tabatha Ricci at 3.74 strikes per minute and a 0.29 career finish rate is slower-paced and unlikely to finish. The model says 51-49, market says Kline at 76%. Low-data flag on this one (three UFC fights for Kline, two for Ricci), but the stylistic gap — Kline's volume vs. Ricci's measured pace — is real. The model's near-coin-flip and the market's heavy Kline lean is the most interesting low-data divergence on the card at 26.8 points, which is why it deserves a mention even with the data caveat.

DFS on a card this heavy with low-data bouts is a construction puzzle more than a model exercise. Here is where the scoring profiles live.

Volume and Floor

Fatima Kline projects at 67.4 points with a 123.8 ceiling — the highest ceiling on the card. Her 5.59 significant strikes per minute and 2.16 takedowns per 15 produce the kind of multi-dimensional scoring that DFS rewards. Three UFC fights of sample, so there is uncertainty, but the activity profile is the best combination of floor and ceiling on the slate. Chase Hooper at 65.3 projected and 4.27 takedowns per 15 is your grappling-volume play — he scores on the mat, he finishes (75% career finish rate), and he likely sees a lot of output against Ramirez. Jacobe Smith at 64.3 projected and 6.24 strikes per minute with 3.73 takedowns per 15 is the highest raw activity profile on the card, if you trust a three-fight sample.

Ceiling Plays

Kamaru Usman at a 127.3 ceiling is the main event outlier — if he controls and finishes or dominates rounds against a champion, the scoring upside in a five-round fight is substantial. His 57.3 projection is also the highest among the main-card-tier fighters. Dione Barbosa at 61.4 projected with 2.64 takedowns per 15 is a grappling-forward ceiling play in the women's flyweight bout, though with a debutant opposite (Anna Melisano has zero UFC fights of tracked history), the low-data caveat is as loud as it gets.

Profiles to Avoid

Tabatha Ricci's 29.6 projection and 0.29 career finish rate is the lowest upside profile on the card — she is unlikely to finish and the volume rate at 3.74 per minute is not going to generate big DFS numbers unless she controls grappling significantly. Marc-Andre Barriault at 38.7 projected offers ceiling (67% finish rate) but a 41% model win probability means you are paying for upside that the model thinks arrives less than half the time. Kevin Holland's 103.4 ceiling is the lowest on the card by a margin — at 33 with an 81-inch reach he can still output, but the DFS upside just is not there on paper.

Three things worth carrying into Saturday night:

1. The Usman price is the card's conversation. Twenty-one points between our number (54%) and the market (33%) on a main event with nearly 20 UFC fights of data behind it is not a rounding error. The model sees a complete, dangerous fighter whose wrestling volume and grappling advances give him a credible path over five rounds. The market sees a 39-year-old former champion on the back nine. Both arguments are reasonable. Our number says the market has overcorrected. This is not a pick — it is a disagreement worth knowing about before you watch.

2. Cannonier-Duncan is the undercard bout that deserves more attention than it will get. Forty-seven percent from the model, 27% from the market, 42 years old, genuine finishing power — Cannonier is being priced like a punching bag. He may well lose. But that gap is the second-largest trusted-bout divergence on the card and the style profile says he can make it painful for Duncan along the way.

3. Most of this card is a low-data exercise. Ten of thirteen bouts. For DFS and for watching, lean on the activity profiles — Kline, Hooper, Smith — and treat any model number on those fights as a rough sketch, not a blueprint. The market price is your most reliable signal on the undercard. We will own that limitation rather than dress it up.