Contender Series 91 results: Pasley starches Berisha in the main

Five fights, five finishes or decisions that made somebody in a suit sweat, and a card that spent the whole night telling anyone with a model or a betting slip to sit down. I had the numbers. The numbers were wrong. Let's get that out of the way first.

Dana White's Contender Series 91 ran Tuesday at the Meta Apex in Las Vegas with seven bouts announced and five that survived to the cage. The sportsbooks had a real opinion in three of the five and got two of them right. Our own model had an opinion in all five and got one right. The only fight it called correctly was the one where a 24-year-old Slovak was priced above 88 percent and had already been through two opponents to get there.

On to the rest of it.

Reginaldo Geraldo Jr. scored a knockdown, four takedowns and 4:51 of control, and lost

This is the one I keep coming back to, and it is not a robbery accusation. It is a criteria question.

Reginaldo Geraldo Jr. (11-2), out of the Fighting Nerds, put Isaac Moreno down once, landed four of six takedown attempts, threatened three submissions and held 291 seconds of control across fifteen minutes. Moreno (9-0) did none of that. What Moreno did was throw, and land, and keep throwing: 66 significant strikes from 147 attempts, split 37 to the head, 13 to the body, 16 to the legs. Geraldo landed 22 of 45.

Three judges gave it to Moreno unanimously. Read the two stat lines cold and you can build either case. Volume plus damage, or knockdown plus control plus submission threat. The scoring criteria say effective striking and grappling first, then aggression, then cage control, and they do not tell a judge how to weigh three minutes of a man lying on top of another man against forty-four more strikes landed.

I would have scored it for Geraldo. I also understand exactly why nobody on the panel did, and that gap is the whole problem with the current criteria rather than a problem with these three judges. When a fight is close, the criteria hand the round to whoever looked busier, and control time only counts when the man on top is doing something with it. Geraldo, to be blunt about it, was not.

Moreno is 9-0 and unbeaten as a pro. Geraldo takes his second career loss on the night he arguably did more work.

Christian Natividad needed 66 seconds against a man who had lost his opponent

Colton Loud was supposed to fight Frank da Silva Castro. That bout came off the card, as we wrote on Sunday, with nobody announcing much of anything. Christian Natividad stepped in, and the market moved to somewhere around minus 320 on him.

The market was right, and it was right fast. Natividad dropped Loud, landed eight of eleven significant strikes and had it finished at 1:06 of the first round. Loud, who came in 7-1 and left 7-2, threw nineteen and landed nine.

Our model had this one at 52 percent for Loud, which is the sort of number a model produces when it has almost nothing to work with on either fighter and splits the difference. Natividad is now 10-0 with five amateur wins behind it. He deserved better than a coin flip and did not need it.

Apollo Gomes got knocked down twice and still won a clean decision

Kwon Won Il is the recognizable name here, the 31-year-old former ONE bantamweight who came in at 14-6 with twenty tracked pro bouts behind him. He put Apollo Gomes on the floor twice. He lost every round anyway.

Gomes, 26, out of Chute Boxe with Diego Lima, landed 76 of 159 significant strikes to Kwon's 48 of 117, went three for thirteen on takedowns and still banked 345 seconds of control. That is 5:45 of the fifteen with Kwon underneath him. Two knockdowns are dramatic. Six minutes of control and a 28-strike edge are the fight.

This is the bout our model-vs-market piece hung its whole argument on. The model liked Kwon at roughly 63 percent. Twenty-one sportsbooks, averaged out, had the fight at 50.1 to 49.9 and were therefore saying, politely, that they had no idea. Refusing to have an opinion turned out to be the better opinion. Gomes goes to 13-2 with a no-contest, Kwon to 14-7.

Martin Kozák was the only sure thing, and he had every reason not to be

Martin Kozák lost Juan Cabrera off this card in the same ten-day stretch that took Loud's opponent. He got Christian Echols (8-5) instead, on short notice, and the line still opened in the minus 300 to minus 600 range depending on the shop and finished north of minus 700 at eighteen of the twenty-three books we track.

He then went out and landed 63 of 88 significant strikes. Seventy-two percent accuracy. Two knockdowns. Forty-one of those strikes to the head, and a TKO at 1:34 of the second.

Twenty-four years old, 7-0, and he is the only man on the card who did what the price said he would do. Note what that means, though. The books had him at 88 percent. Our model had him at 71.6. The one fight we got right, we were the timid ones on.

Quentin Pasley finished the main event and nobody had him

Arlind Berisha came in 5-0 with a kickboxing base out of Norway and closed at roughly minus 170. Quentin Pasley closed at plus 140 or so, and our model made it worse for him, at 46 percent.

Pasley then took the fight apart from the outside. Twenty-eight of forty-five significant strikes landed, and here is the interesting part of the breakdown: only six of them went to the head. Twelve to the body, ten to the legs. He chopped Berisha down rather than hunting the finish upstairs, then got it anyway at 4:46 of the first.

Berisha landed fourteen, nine of them leg kicks of his own, and took his first professional loss. Pasley is 4-0.

Getty's wire captions from the Apex have Pasley, Kozák, Natividad and Isaac Moreno each photographed reacting to an earned UFC contract. Four from five fights would be a heavy night by Contender Series standards, and the promotion has not put out a numbered list, so treat that as what the photographers saw rather than an official tally.

The receipts, since I opened with them

The pattern in that table is not really the one-for-five. It is the second column against the third. On the two fights where the market declined to pick a side, both landing inside half a percentage point of a coin flip, our model was 74 percent and 63 percent confident. It lost both. On the two fights where the market was confident, our model agreed once and picked the other way once. The agreement won. The contrarian call lost in sixty-six seconds.

That is a specific failure mode and it is worth naming: a model trained mostly on UFC data will manufacture certainty about fighters who have never been in the UFC. Every man on this card walked in with twenty-one professional bouts or fewer, most of them under fifteen, and a large share of those on regional shows that publish no statistics at all. The books know they are guessing there and price it. We evidently do not, yet.

Contender Series 92 is Tuesday, and it already lost a bout and swapped an opponent since we wrote the card up. Same building, same fifteen-minute job interviews, same ten men who will be told something about their future by a man in a golf shirt. I will have picks. Take them however you like.