UFC Allen vs Duncan predictions: model picks for all 12 bouts

On Tuesday I wrote that the market and our model had picked different winners in the main event. At the time the line had Brendan Allen at -132. It has kept leaking since. The consensus across 18 sportsbooks is -123 this morning, against +102 for Christian Leroy Duncan, which strips down to 52.7 percent for Allen. Our model still has Duncan at 62.9. The fans are not having any of it: 24 of the 35 people who have picked the fight on our pick'em took Allen.

UFC Fight Night: Allen vs Duncan is Saturday, October 10, at the Meta Apex in Las Vegas. The prelims start at 5 p.m. ET and the main card at 8 p.m. ET; every region and broadcaster is in the viewing guide. Below is the model's number for all 12 bouts, where the market sits, how the fans have voted, and my picks. Lines are quoted as information only, not betting advice.

The card

The model column is the win probability for the model's pick. The market column is the consensus for that same fighter with the bookmakers' margin removed. The fans column is the pick'em vote, the majority side first.

The model never sees a betting line or a pick'em vote. That is the whole point of putting the three side by side.

The main event

Photo: WEIGHING IN / CC BY 4.0

I made the model's case on Tuesday, so the short version: Allen (27-7), No. 4 at middleweight, gets hit more often than Duncan does. He absorbs 3.8 significant strikes a minute to Duncan's 2.4, avoids 48 percent of them to Duncan's 54, and lands 52 percent of his own to Duncan's 58. Duncan (15-2), No. 10, is on a five-fight win streak in the UFC and was unbeaten before he got there. Those are the model's three biggest reasons, and every one of them points at the Englishman.

The line opened on October 1 at -141 on Allen. It is -123 now. Somebody has spent the week agreeing with the model a little at a time.

I am still taking Allen, and I will tell you exactly why, so you can laugh at the right paragraph on Sunday. This is five rounds. Allen stopped Reinier de Ridder after four of them last October, and six of his eight finishes at middleweight in the UFC are submissions. Duncan's last two wins, over Roman Dolidze in March and Jared Cannonier in July, were both 15-minute decisions. The model's case is built on three-round striking rates. Rounds four and five are a different sport, and Allen has been there.

Prediction: Brendan Allen by submission

Frye Jr. vs Harris: 100 seconds against eight days

This is the widest gap on the card that has not had a piece of its own this week. RJ Harris (6-0) is the only unbeaten fighter on the bill, and his UFC career so far is one fight long: a first-round knockout of Alvin Hines at 1:40 in July, 15 significant strikes from 26 thrown and a knockdown. Allen Frye Jr. (6-1) has one UFC fight too, a unanimous decision loss to Guilherme Pat last December. Pat outlanded him 64 to 33, but Frye spent 7 minutes and 55 seconds of that fight in control, which is not nothing for a man every one of whose six pro wins came by knockout.

The market has made up its mind. The first price we recorded, on October 1, had Harris at -220 and Frye at +186. It has moved one way for eight days, with only the odd tick back: -267 by Monday, -300 by Tuesday, -345 this morning. That is 74.1 percent for Harris once the margin comes out.

The model has Harris too, at 60.9 percent, and its biggest reason is his level of competition. But it flags that reason, and two others, as thin samples, and it is right to. Two heavyweights with one UFC fight each is about as little film as a model ever has to work with, and when the evidence is thin it pulls toward a coin flip. The market does the opposite. It has watched 100 seconds of Harris and priced him like a finished product.

Frye is the bigger man, 6-foot-5 with an 80-inch reach against 6-foot-3 and 78. Both are 28. I think Harris wins, and I think the line has gone further than the film does.

Prediction: RJ Harris by KO/TKO

Price vs Shahbazyan: the model's other dissent

Here the model is on the other side of the line. It has Niko Price at 52.1 percent; the market has him at 42.1 and has drifted away from him since the first price on October 2, +115 to +127. The fans have Leon Shahbazyan 5-3.

The case against Price writes itself, and the model writes it first: he is 37, and he has lost four straight. Themba Gorimbo beat him on the cards in October 2024, Jacobe Smith submitted him in the second round in June 2025, Nikolay Veretennikov knocked him out in 1:42 in February and Michael Chiesa submitted him in 1:03 in March. His UFC record is 8-11 with two no contests.

What tips the model to Price is the other man's sample. Shahbazyan (12-5), 30, has one UFC fight, and it lasted 23 seconds: Levan Chokheli knocked him out in June. The model's striking numbers have him absorbing 8.5 significant strikes a minute and avoiding 33 percent, which is what a 23-second knockout does to a rate. I do not put much weight on that figure. I do put some on the fact that Price has had 21 fights in the UFC and Shahbazyan has had one.

Prediction: Niko Price by KO/TKO

Where the model and the market agree

Matheus Camilo vs Jai Herbert. The model's biggest reason is age, and it is a big one: Camilo is 25, Herbert is 38. It also likes Camilo's wrestling, 4.2 takedowns per 15 minutes to Herbert's 0.3, and his defense, 67 percent of significant strikes avoided to 53. What it holds against him is size: Herbert is 6-foot-1 with a 77-inch reach, Camilo 5-foot-10 with 69. Both men are coming off first-round knockouts, Camilo over Nazim Sadykhov in 91 seconds in June and Herbert over Mandel Nallo in April. The line has firmed from -170 to -191. Prediction: Matheus Camilo by decision

Loopy Godinez vs Ketlen Souza. The market is more confident than the model, 67.4 to 58.7, and the line has gone from -210 to -239. Godinez, No. 8 at strawweight, was submitted by Tatiana Suarez at UFC 327 in April. Souza has won two straight, the last a 94-second knockout of Ariane Carnelossi. The model's reasons for Godinez are the grappling numbers, three takedowns per 15 minutes to Souza's 0.5 and a takedown defense of 77 percent to 54. The pick'em is split 4-4. Prediction: Loopy Godinez by decision

Felipe Franco vs Brendson Ribeiro. Ribeiro has lost four straight: Azamat Murzakanov and Oumar Sy by knockout, Abdul Rakhman Yakhyaev by submission and Magomed Tuchalov on the cards in July. Franco is 25 and stopped Levi Rodrigues Jr. in the second round in July. The model and the market sit two points apart. Prediction: Felipe Franco by KO/TKO

Julius Walker vs Gerald Meerschaert. Meerschaert is 38, on a five-fight losing streak, and moving up to light heavyweight for his 60th professional fight. Walker has been knocked out in his last two. This bout has more history in it than its prelim slot suggests. Prediction: Julius Walker by KO/TKO

The rest

Andre Fili vs Kai Kamaka III. The model gives Andre Fili 53.1 percent, the market 56.7. Its reasons against Fili are age, 36 to 31, and level of competition; its reasons for him are his 25 career wins and the size gap, 74 inches of reach to 69. Fili has won five split decisions in the UFC, and I expect a sixth. Prediction: Andre Fili by decision

Malcolm Wellmaker vs Otari Tanzilovi. I wrote about this one on Tuesday, when the model had it level. It now leans Tanzilovi, 51.9 percent, largely on the strength of his 4.5 takedowns per 15 minutes against Wellmaker's 47 percent takedown defense. The line went the other way, from -176 on Wellmaker when it opened to -215 now. I am staying with the wrestler. Prediction: Otari Tanzilovi by decision

Francisco Prado vs Ismael Bonfim. The model has this at exactly 50.0, and the market has spent the week turning it over: Prado opened at +130 and is now a -121 favorite. Prado is 24 and has lost four straight in the UFC, all on the scorecards. Bonfim is 30 and has lost three straight, all inside the distance. Prado is the one whose losses have all lasted 15 minutes. Prediction: Francisco Prado by decision

Alice Pereira vs Daria Zhelezniakova. Pereira turned 20 last December, and age is the model's second-biggest reason for taking her at 58.6 percent; its biggest goes the other way, a fractional edge in level of competition for Zhelezniakova, No. 15 at bantamweight. Pereira knocked out Hailey Cowan in the second round in April. Prediction: Alice Pereira by decision

Ernesta Kareckaite vs Melissa Gatto. All seven people who have picked this fight took Gatto. The line has gone the other way: Gatto opened at -150 and is -117 now. Both women have lost to Dione Barbosa, Kareckaite at UFC 301 in May 2024 and Gatto on a majority decision in April. The model has Gatto at 55.6. Prediction: Melissa Gatto by decision

Picks close at the first bell on the pick'em, and the results will land on the event page as the night goes. If Duncan wins on Saturday, the model gets the main event and I get Sunday morning to explain myself.