Some UFC cards are made for analysis.

This is not really one of them.

UFC Fight Night: Rosas Jr. vs Barcelos has a handful of genuinely interesting matchups at the top, but once you move through the full 12-fight card, the amount of reliable information starts thinning out quickly. There are relative unknowns, fighters with shallow UFC samples, long periods of inactivity and matchups where neither side has done enough at this level to justify pretending there is a strong read.

The model reflects that uncertainty.

Nine of the 12 fights came back with low separation between the two fighters. Two were effectively dead even at 50%, and only John Castaneda and Yazmin Jauregui reached the model’s higher-confidence band.

That does not make it a bad fight card.

It does make it a pretty bad card to analyse with much certainty.

So rather than manufacturing confidence where there isn’t any, this week’s Fight Lab concentrates on the five fights where there is enough substance to actually dig into the matchup.

At the top, Raul Rosas Jr. gets the biggest test of his career against Raoni Barcelos. Ailin Perez faces another significant step towards the top of the women’s bantamweight division. Brady Hiestand and Rinya Nakamura could produce one of the more interesting wrestling battles on the card.

And further down, Castaneda and Jauregui provide two of the clearest positions from the model.

Raul Rosas Jr. vs Raoni Barcelos

This is exactly the sort of fight Raul Rosas Jr. needs at this stage of his career.

The talent has never really been the question.

Rosas is still only 21, already sits inside the bantamweight rankings and continues to build a game around relentless wrestling pressure, back takes and the ability to force opponents into prolonged grappling exchanges.

His unanimous decision victory over Rob Font earlier this year was another important step.

Barcelos presents a very different problem.

At 39, he is almost the opposite proposition. There is very little Rosas can show him that Barcelos has not encountered somewhere during a long career.

He is experienced, physically strong, difficult to control and capable in every phase of the fight.

More importantly, he is currently winning.

Barcelos arrives having built a five-fight UFC winning streak, proving that whatever the birth certificate says, he remains an extremely awkward man to beat at bantamweight.

The Rosas Jr. path

There should be no mystery about what Rosas wants.

Pressure Barcelos backwards, get onto the hips and keep chaining attempts together until the fight reaches the floor.

The first takedown does not necessarily have to work.

Neither does the second.

What makes Rosas difficult is his willingness to remain attached to opponents and continue working until something eventually opens.

Over five rounds, that persistence could become increasingly important.

The challenge is efficiency.

This is Rosas’ first UFC main event, and repeatedly forcing wrestling exchanges against someone as experienced and defensively sound as Barcelos could become expensive if he is unable to establish control.

Rosas needs to make Barcelos work without emptying his own tank in the process.

The Barcelos path

Barcelos needs to make Rosas start over.

Defend the entry, create separation and force the younger fighter to cross distance again.

Every clean reset matters.

The more time this fight spends at kickboxing range rather than with Rosas attached to a hip or hunting the back, the better it becomes for Barcelos.

He does not need to overwhelm Rosas anywhere.

He needs to frustrate him.

There is also a question here that Rosas has not really had to answer yet. What happens when the wrestling works, but not easily?

If Barcelos makes every takedown difficult and every control position temporary, Rosas may have to solve a much more complicated fight than he has encountered so far.

Fight Lab verdict

This is close.

Very close.

The model has Barcelos at just 52.6%, with decision as his most likely route to victory.

Fred’s read lands almost exactly in the same place.

Rosas has absolutely reached the point where beating Barcelos would not be surprising. His pressure, wrestling and ability to keep creating grappling exchanges could be enough to carry rounds.

But Barcelos is probably the more complete fighter right now and, crucially, one of the better-equipped opponents Rosas has faced when it comes to making his preferred game difficult.

There is barely anything between them.

Model: Raoni Barcelos, 52.6%

Fred’s Pick: Raoni Barcelos by decision

Norma Dumont vs Ailin Perez

This might be the most meaningful fight outside the main event.

Norma Dumont enters ranked No. 4 at women’s bantamweight, with Ailin Perez sitting directly behind her at No. 5.

Perez has now won six straight.

And there is very little confusion about how she has done it.

Wrestling.

Pressure.

Top position.

Repeat.

Perez has increasingly developed into a fighter who understands exactly where her advantages are and has little interest in abandoning them.

She does not need to produce spectacular offence if she can consistently put opponents on their backs and spend large portions of rounds controlling the action.

Dumont is a serious test of whether that approach continues working against the division’s better fighters.

The Dumont path

Dumont needs this fight to stay clean.

Space matters.

If she can keep Perez outside, deny the entries and force extended striking exchanges, she should have opportunities to build rounds behind the more measured kickboxing game.

The danger is allowing Perez to turn every minute into a wrestling battle.

Even defending takedowns can become a losing proposition if Dumont spends long stretches pinned against the fence and unable to establish her own offence.

She needs separation rather than simply defence.

The Perez path

Perez should not overcomplicate this.

Get close.

Make contact.

Wrestle.

Her current six-fight winning streak has been built around repeatedly forcing opponents to deal with that exact problem.

Dumont should be harder to control than many of the women Perez has already beaten, but Perez does not necessarily need extended domination.

Two competitive rounds with enough takedowns, control and positional pressure could be enough.

This has 29-28 written all over it.

Fight Lab verdict

The model gives Perez 57.8% and strongly leans towards a decision if she wins.

Fred agrees with the winner and the likely shape of the fight.

Dumont is good enough to make this competitive and quite possibly take a round. She may even have the cleaner moments at range.

The difference could simply be Perez’s ability to decide where enough of the fight takes place.

Expect something tight rather than dominant.

Model: Ailin Perez, 57.8%, with decision the preferred method

Fred’s Pick: Ailin Perez by decision, likely 29-28

Brady Hiestand vs Rinya Nakamura

This could be one of the most interesting fights on the card.

It is also one where the model and Fred agree on the fighter but not necessarily the finish.

Brady Hiestand’s game is based around activity.

He wants to wrestle, pressure and turn fights into prolonged physical exchanges.

Rinya Nakamura brings a much stronger wrestling pedigree into the matchup and has looked capable of beating opponents at their own game.

That immediately creates an interesting question.

What happens when Hiestand cannot reliably dictate the grappling?

The Hiestand path

Hiestand needs to make this exhausting.

Constant entries, clinch pressure, scrambles and repeat attempts.

Even if he cannot consistently complete takedowns, making Nakamura defend them over three rounds could still change the fight.

There is also an uncertainty surrounding Nakamura because of inactivity.

He has been away for more than a year, and his last appearance lasted barely a minute.

That makes it difficult to know exactly how he has developed or how he responds if Hiestand can force him into a sustained high-tempo fight.

The Nakamura path

This is where the matchup becomes interesting.

Nakamura may simply be the better wrestler.

If he can stuff Hiestand’s entries, reverse positions or punish the openings created when Hiestand commits forward, the entire dynamic changes.

And there is another possibility.

Hiestand’s commitment to pressure could leave him exposed.

The model expects Nakamura to win but prefers a decision. Fred sees a more damaging route.

If Hiestand repeatedly forces his way into range and Nakamura starts timing him coming forward, there could be opportunities to hurt him rather than simply out-wrestle him.

Fight Lab verdict

The model has Nakamura at 57.8%, with decision as the most likely method.

Fred agrees with Nakamura but sees the finish differently.

Hiestand’s pressure should create plenty of exchanges, but Nakamura looks better equipped to win the wrestling battle and may also have the sharper opportunities when Hiestand starts forcing the action.

If Nakamura can punish those entries rather than simply defending them, the fight could finish before the scorecards.

Model: Rinya Nakamura, 57.8%, with decision the preferred method

Fred’s Pick: Rinya Nakamura by KO/TKO

John Castaneda vs Alatengheili

Finally, a fight where the model is willing to plant its feet.

John Castaneda comes out at 77.3%, the second-highest win probability on the card.

This is one of the cleaner matchups to read.

Castaneda should have the advantage in overall activity and has enough variety to make Alatengheili solve more than one problem.

He can strike in combination, mix his targets and bring wrestling into the equation when needed.

Alatengheili is experienced and durable enough that simply walking through him is unlikely.

But across 15 minutes, Castaneda should have more ways to build winning rounds.

The Castaneda path

Volume without getting reckless.

Castaneda does not need to chase a finish here.

Keep Alatengheili reacting, vary the offence and avoid giving him long periods where the pace slows enough to set clean counters.

The longer the fight stays busy, the more Castaneda should be able to separate himself.

The Alatengheili path

Alatengheili needs to make the fight smaller.

Fewer exchanges.

More patience.

More opportunities to land something meaningful rather than trying to match Castaneda’s output.

If Castaneda is allowed to dictate the pace for three rounds, this becomes difficult very quickly.

Fight Lab verdict

This is one of the strongest agreements between Fred and the model.

Castaneda’s ability to produce more offence, mix his attacks and win minutes gives him a clear path across three rounds.

Alatengheili should be difficult enough to put away, which points towards the judges.

Model: John Castaneda, 77.3%

Fred’s Pick: John Castaneda by decision

Vanessa Demopoulos vs Yazmin Jauregui

The biggest model number of the week belongs to Yazmin Jauregui.

88.4%.

That is a big number in MMA.

Big enough that it deserves challenging rather than simply accepting.

Fights are too chaotic and Jauregui is returning after a lengthy period away herself, so treating an 88% prediction as anything approaching certainty would be foolish.

But the stylistic case is easy to see.

Jauregui should have significant advantages in the striking exchanges.

She is faster, sharper in combination and far more dangerous when allowed to settle into her offence.

Demopoulos needs to make sure that does not happen.

The Demopoulos path

Chaos.

That is the fight.

Standing in front of Jauregui and kickboxing for three rounds is unlikely to end well.

Demopoulos needs grappling exchanges, awkward scrambles and opportunities to turn the fight into something far less technically clean.

The more uncomfortable the fight becomes, the more realistic her chances become.

The Jauregui path

The biggest danger may be trying to prove how large the striking advantage is.

She does not need to.

Stay patient.

Manage distance.

Punish entries.

Force Demopoulos to continually reset.

If Jauregui keeps the fight structured, she should be able to win long stretches of it without taking unnecessary risks.

Interestingly, the model does not see the most likely victory as another knockout.

It prefers the decision.

Fred agrees.

Fight Lab verdict

Jauregui is the clearest winner pick on the card.

The 88.4% model number might look extreme, particularly after a long period out, but the basic matchup still heavily favours her.

Demopoulos has a route if she can turn this into a messy grappling fight.

If she cannot, Jauregui should control the striking exchanges for most of the 15 minutes.

Model: Yazmin Jauregui, 88.4%

Fred’s Pick: Yazmin Jauregui by decision

Final Word

There is a temptation when analysing every UFC event to pretend each card contains the same amount of information.

It doesn’t.

This one certainly doesn’t.

There are several fights on Saturday where the available UFC sample is thin, the fighters are still developing or there simply is not enough separation to justify a strong opinion.

The model makes that pretty obvious.

Nine low-separation fights from a 12-fight card is not a green light to start digging harder until an edge magically appears.

Sometimes uncertainty is the answer.

The five fights covered here are the ones that offer something more substantial.

Barcelos over Rosas Jr. is barely a lean and should be treated that way. Perez over Dumont looks similarly competitive, with a narrow decision the most plausible outcome.

Nakamura is the more interesting disagreement between the model and Fred, with both taking the Japanese prospect but Fred expecting his power to matter more than the method model does.

Castaneda and Jauregui are different.

Those are the two places where the numbers finally produce meaningful separation, and Fred agrees with both.

Not every UFC card needs bold predictions everywhere.

Sometimes the edge is knowing where the information stops.