UFC 333 predictions: fans back both champions, our model doesn't

UFC 333 is Saturday at Etihad Arena in Abu Dhabi, and it starts early for most of the world: early prelims at 10:00 a.m. ET, prelims at noon, and the main card at 2:00 p.m. ET, which is 10:00 p.m. in the arena. Two title fights sit on top. Alexander Volkanovski defends the featherweight belt against Movsar Evloev, and Petr Yan defends the bantamweight title against Merab Dvalishvili in the third fight between them. Broadcasters by country are in our UFC 333 viewing guide.

This week I've written about these two fights from the sportsbooks' side and from our model's side. This is the last preview before the bell, so it adds a third voice. By Friday morning, 364 people had made picks on the UFC 333 pick'em, 275 of them on the main event. They have picked a lane: they are backing both champions. Our AI model is backing both challengers.

Three opinions, 14 fights

Here is the whole card. Fans is the share of pick'em entries on the first-named fighter, model is our model's win probability for that fighter, and the line is the consensus across 18 sportsbooks as of this morning.

Fans and model land on the same winner in six fights. They split in six more: the two title fights, Murzakanov vs Reyes, Thicknesse vs Chopurov, Vagaev vs Malkoun and Oki vs Santos. In five of those six, the fans are standing next to the betting favorite. The odd one out is Oki, picked by 60 percent of the crowd while the books lean slightly to Santos. The other two fights have no split to speak of. The model has Dawson vs Aliev at 49.9 percent, and the pick'em is exactly 8 to 8 on Steele vs Gugnon.

The prelim numbers come with small samples, 14 to 55 entries a fight, so treat them as a mood and not a poll. The main event's 275 entries are a different matter.

Volkanovski vs Evloev

Volkanovski (28-4) turned 38 on September 29 and comes in off two unanimous decisions over Diego Lopes. Evloev (20-0) is 32, ranked No. 1 at featherweight, and has won all ten of his UFC fights on the scorecards. The fans have Volkanovski at 62.5 percent. The books have it at a coin flip and have barely moved since August. Our model has Evloev at 66.5 percent, the widest gap between the crowd and the model on the main card.

I went through Evloev's 30 UFC scorecards on Tuesday and the champion's path to a win on Thursday, and both pieces ended in the same place: the fight is decided by whether Evloev starts shooting in the first round or waits until the third, the way he did against Lerone Murphy in March. The crowd is betting on the Volkanovski who outlanded Lopes 158 to 63 in their first fight. I think the crowd is underrating how few men have even tried to wrestle him. The one who did it seriously, Islam Makhachev, won three rounds on every card.

Prediction: Evloev by decision.

Yan vs Dvalishvili

Photo: Onlyfans / CC BY 3.0

The series is 1-1. Dvalishvili (21-5) won the first fight 50-45 on all three cards in March 2023. Yan (20-5) took the belt from him at UFC 323 in December, stopping 27 of 29 takedown attempts on the way to a unanimous decision. The fans and the books both read that December fight as the true one: 63.8 percent of entries are on Yan, and he is -155. Our model reads the career instead and gives Dvalishvili 57.1 percent.

The fans' case is easy to make. Yan is the younger man, 33 to 35, and he has won four straight. The model's case rests on something the fans may be forgetting. December was Dvalishvili's fourth five-round fight of 2025, 63 days after he shot 37 takedowns at Cory Sandhagen. By Saturday he will have had 322 days off. I made the full case on Tuesday, and I haven't talked myself out of it.

I'll say this for the crowd: Yan stopped all eight of Dvalishvili's first-round attempts in December, on fresh legs, before fatigue could be the excuse. If the first round goes that way again, the 64 percent will look smart.

Prediction: Dvalishvili by decision.

The rest of the main card

Lone'er Kavanagh vs Ramazan Temirov. Everyone shrugs here. Fans have Temirov at 56 percent, the model at 52.6, and the line is a pick'em. Kavanagh is No. 6 at flyweight and is coming off a third-round submission loss to Brandon Royval at UFC 329 in July. Temirov is No. 10, 3-0 in the UFC, and two of those three ended in the first round, the latest against Steve Erceg in July. Prediction: Temirov by knockout.

Alexander Volkov vs Rizvan Kuniev. This is the fans' strongest main-card opinion: 80 percent on Volkov. The model agrees, at 63.8. The books are the outlier, with Volkov at -120 after opening at -130. Volkov (40-11) is No. 2 at heavyweight and has won two straight decisions, over Jailton Almeida and Waldo Cortes Acosta. Kuniev (14-3-1, one no contest) is No. 7 and also beat Almeida, in February. Prediction: Volkov by decision.

Aaron Pico vs Losene Keita. This fight has had the biggest line move on the card. It opened at a pick'em on October 14, and Pico is now -143. The model and the pick'em have both stayed on Keita, at 59.6 and 59 percent. Pico lands 6.54 takedowns per 15 minutes in his short UFC run, and that is the reason I can't follow the model this time. Prediction: Pico by decision.

Prelims and early prelims

Azamat Murzakanov vs Dominick Reyes. Murzakanov is -209 and has 56 percent of the crowd. The model has Reyes at 58.3, a gap of 23 points against the market, and I took that apart earlier in the week. Murzakanov is coming off a third-round knockout loss to Paulo Costa in April, the only loss of his 17-fight career. Prediction: Murzakanov by knockout.

Nikita Krylov vs Abdul Rakhman Yakhyaev. Here, and only here, the fans are more certain than the books: 87.3 percent on Yakhyaev, against a line that implies about 81. Yakhyaev (10-0) has finished all three of his UFC opponents in the first round, two of them by submission. Prediction: Yakhyaev by submission.

Abus Magomedov vs Cam Rowston. Fans 76 percent on Magomedov, model 51.4, the line -125. That makes the crowd 23 points more confident than the people setting the price. Rowston is 3-0 in the UFC with two knockouts. Prediction: Rowston by knockout.

Grant Dawson vs Nurullo Aliev. The model can't separate them, the books barely can, and 70 percent of the crowd is on Aliev, who is 12-0 and has won all four UFC fights on the cards. Prediction: Aliev by decision.

The early prelims I'll take quickly. Asaf Chopurov (11-0) makes his UFC debut as a -609 favorite over Colby Thicknesse, and the model is the only one of the three on Thicknesse. Abubakar Vagaev has 73 percent of the pick'em against Jacob Malkoun; the model has Malkoun at 58.2. Bolaji Oki and Daniel Santos both arrive off knockout losses. Kody Steele vs Noah Gugnon is the one fight the crowd can't decide, and Iasmin Lucindo, at 80.2 percent, is the model's most confident pick of the night. Predictions: Chopurov by decision, Vagaev by decision, Santos by knockout, Steele by submission, Lucindo by decision.

Who's right on Saturday

Count up my picks and you'll find I side with the crowd in eight fights and with the model in eight, and in two, Pico and Rowston, I'm against both. That isn't me being diplomatic. It's what happens when you read each fight and ignore the scoreboard. The two that matter are at the top, and on those I'm with the model, which means against 62.5 and 63.8 percent of the people who bothered to pick. If both champions keep their belts, the crowd gets to say it told me so. If both lose, Abu Dhabi watches two belts change hands in one night, and the model gets the last word.

The pick'em stays open until the first bell. If you think I've got it backwards, add your picks and move the numbers.