UFC Fight Night 289 odds: three underdogs our model backs

Our model and the betting market usually agree. They are both looking at the same fighters, the same records and mostly the same tape, and on a normal card you get eleven bouts where the model's number lands within a few points of the implied probability and one where it does not. UFC Fight Night 289 on Saturday is not a normal card. There are three fights on it where the disagreement is large enough to be worth writing down, and in all three the model is standing on the underdog.

I am not going to tell you the model is right. Twenty-three sportsbooks priced these fights and they are not run by idiots. But a twenty-point gap is not noise, and the reasons behind each of these three are legible enough that you can decide for yourself which side is missing something.

Vanessa Demopoulos at +602

The consensus has Vanessa Demopoulos at +602 against Yazmin Jauregui, which is a market saying she wins about one time in seven. Our model gives her 34.3 percent. That is the widest gap on the card by a distance, and the line has been moving away from her: she opened around +470 and has drifted out to +602 across the week.

The case against Demopoulos writes itself. She is 11-8, she has lost three in a row (Jamey-Lyn Horth, Talita Alencar, and a first-round submission to Jaqueline Amorim), she absorbs 5.39 significant strikes a minute, she defends 41 percent of what comes at her and 26 percent of takedowns. She turns 38 today. A 38 year old on a three-fight skid with those defensive numbers is not a fighter the market is going to protect.

The case the model is making is about the other side. Jauregui is 11-2 and she has not fought since September 15, 2024, when Ketlen Souza submitted her in the first round. That is a two-year layoff, and the fight before it she was knocked out in twenty seconds by Denise Gomes. Her striking volume is enormous, 6.64 significant strikes landed per minute, but her accuracy is 39 percent and she takes 5.25 back, which is the profile of someone who wins firefights and loses them in equal measure. Two years out of the cage, at 27, coming off a submission loss, against an opponent who has been in nineteen professional fights and keeps showing up. Seven-to-one is a lot to lay on that.

Brady Hiestand at +277

Brady Hiestand is +277 against Rinya Nakamura and the model has him at 44.4 percent, which is nearly a coin flip. This is the gap I find easiest to argue with the market about, because both fighters are carrying the same problem and only one of them is being charged for it.

Hiestand is 8-2 with three straight wins, the most recent a third-round submission of Garrett Armfield in June 2024. That is twenty-seven months ago. Nakamura is 10-1, and his last appearance was a 62-second knockout of Nathan Fletcher in August 2025, which is thirteen months ago. Both men are rusty; Nakamura is half as rusty and is being priced as though he is three times the fighter.

Nakamura's numbers are genuinely good, and this is where I have to argue against myself: 81 percent takedown accuracy and a perfect takedown defense record in the UFC, 3.27 significant strikes landed against 1.72 absorbed. That is a controlling, low-damage profile and it is why the market likes him. But his one loss came the last time somebody dragged him into a full fifteen minutes, a unanimous decision to Muin Gafurov in January 2025, and Hiestand attempts 4.80 takedowns per fifteen minutes with a 1.1 submission average. The model is not saying Hiestand wins. It is saying that a fighter who wants exactly the fight Nakamura's one loss came in should not be a three-to-one dog.

Elves Brener as a dog at all

The third one is smaller in price and larger in implication. Josiah Harrell is a -134 favourite over Elves Brener, which makes Brener a +111 underdog. The model has Brener at 55.3 percent, so it has the fight inverted: not a big edge, but the wrong man is favoured.

Brener is 16-6 and lost his last three, so I understand the price. Esteban Ribovics outpointed him in August 2025, Joel Alvarez stopped him in 2024, Myktybek Orolbai outpointed him before that. Those are three real lightweights and he was competitive with two of them. He lands 4.76 significant strikes a minute and defends 76 percent of takedowns.

Harrell is 11-1 and that single loss is his entire UFC career: Jacobe Smith knocked him out in three minutes in February 2026. In that appearance he absorbed 10.94 significant strikes per minute and defended 29 percent of what was thrown at him. One fight is a small sample and I will say so plainly. It is also the only octagon-level evidence that exists about him, and the market has decided to discount it in favour of the eleven wins he built before he got here. The model has not.

What the market is confident about

Photo: Profigrappling / CC BY-SA 4.0

For balance, the biggest line move on this card runs the other way. Rodolfo Vieira opened at -107 against Robert Bryczek and is now -152, a move of forty-five points with no announced change to the fight. Our model has Vieira at 56 percent, which is roughly where he opened and well short of where he has been bet to.

Norma Dumont has drifted the same way, from +70 to +116 against Ailin Perez, and that one at least has a story attached: Perez has won five straight and Dumont lost in April. When a line moves that far on a fight where nothing has changed, somebody knows something or somebody is guessing loudly, and from outside you cannot tell which.

Model numbers for every bout are on the event page, and you can log your own calls at the pick'em before Saturday.