The model went 6 for 13 at UFC Fight Night 272, a losing card by our own standard, and it missed on the biggest fight of the night. Renato Moicano submitted Chris Duncan in the second round to close the April 4 card, and the pick had leaned the other way.

That headliner ended inside two rounds, with Moicano finishing by submission. The undercard supplied the cleaner work. Abdul Rakhman Yakhyaev needed one round, taking a first-round submission. Tommy McMillen matched the pace from the other side of the cage, scoring a first-round knockout. Two fights, two quick finishes, both on the board before the second horn.

The model autopsy\n\nThis is where we grade ourselves, and the grade for this card is below the line. The model went 6 for 13 across the night, 46 percent. That is a losing card by our own standard, and there is no soft way to frame a sub-50 number.\n\nThe flagged picks held up better than the field. Among the LOCK and HIGH-confidence calls, the model went 3 for 4. Yakhyaev was tagged at 82 percent and delivered, finishing in the first. McMillen carried the same 82 percent confidence and also came through, again inside a round. Those are the kinds of calls the flagged tier exists for, and on this card they did their job.\n\nThe fourth flagged pick is the one we have to own, and it was the biggest fight on the card. The model put Chris Duncan as the side at LOW confidence, 43 percent, in the main event. Moicano won by second-round submission. The pick was wrong. A 43 percent read means the model was already closer to a coin flip than a conviction call, so this was not a confident miss so much as a fight the model never had a handle on. It still goes in the loss column, and it sat at the top of the marquee.\n\nNo upsets were flagged as called on this card, so there is no contrarian win to point to here. The two first-round finishes from the flagged tier are what kept the grade off the floor.\n\nSo the honest version: a 46 percent card with a main-event miss, and a flagged tier that went 3 for 4 on the strength of two quick undercard finishes.\n\n### One honest takeaway\n\nCards like this are why we publish the grade instead of the highlight reel. Our canonical live 2026 record sits at 67.5 percent across all picks and 82.8 percent on the LOCK and HIGH tier over 64 graded selections. Fight Night 272 ran under both of those marks, and that is the point of tracking it. A single card swings hard on a handful of fights, and one night at 46 percent does not move a season-long number much, but it does not get hidden either.\n\nThe flagged tier going 3 for 4 here is a more useful signal than the raw 6-of-13. The high-confidence calls were mostly right, including both first-round finishes. The main-event pick, at 43 percent, was wrong, and a near-coin-flip read on a headliner is the model signaling it had no edge. We would rather show you that than dress it up.