No bets went out Saturday, and the model's winner picks finished 8-6 — with a method record ugly enough to need its own paragraph.
UFC 333 closed with the model picking winners correctly in 8 of 14 graded fights, good for 57%. That's a coin-flip night dressed up as a card recap, and we're not going to pretend otherwise. The whole point of filing these every Monday is to let the record speak, win or lose, and this one leans lose.
Worth saying up front: zero fights on this card graded A+, A or B. That means zero recommended plays went out, and the unit column reads 0-0, +0.00. Whatever happened Saturday night, it didn't cost or make the desk a dime. That's not a flex — it's just the accounting. The model wasn't confident enough in anything to put a number behind it, and given how the main-card misses stacked up, that caution looks justified in hindsight.
Six misses, and they weren't scattered junior fights — they ran straight through the top of the card. Alexander Volkanovski beat Movsar Evloev in the headliner; the model had Evloev. Petr Yan beat Merab Dvalishvili; the model had Dvalishvili. Those are two of the biggest names on the poster going the other direction from the pick.
It kept going. Dominick Reyes got the nod over Azamat Murzakanov, who the model favored. Abus Magomedov beat Cam Rowston. Grant Dawson beat Nurullo Aliev. Abubakar Vagaev beat Jacob Malkoun. Six wrong calls, six fighters we need to just say plainly: Volkanovski, Yan, Reyes, Magomedov, Dawson and Vagaev all had their hand raised over the model's pick. No spin on any of them.
Here's the number that actually stings more than 8-6: of the eight fights the model called correctly on the winner, it got the method of finish right exactly zero times. Lone'er Kavanagh, Alexander Volkov, Aaron Pico, Abdul-Rakhman Yakhyaev, Colby Thicknesse, Daniel Santos, Kody Steele and Iasmin Lucindo all won the fights the model said they'd win — and every single one of those finishes happened in a way the model didn't call.
That's eight correct winners and eight method whiffs, which tells you the model was reading these fights right in a general sense — favoring the right side — without actually having a handle on how the fights would unfold. A card can go 8-6 on paper and still feel worse than the record suggests when the finish-method read is 0-for-8 on the hits. This one does.
No units moved, no plays got burned, and that's the one thing to hang onto out of a card that otherwise reads like a shrug. The discipline to pass on everything here meant the misses on Volkanovski, Yan and Reyes stayed theoretical instead of expensive.
Next card, the goal is simple: find the fights where the model's conviction and its actual hit rate start lining back up. This one isn't it. We'll say so again next Monday if it isn't.