# UFC predictions — 12 September 2026

Prepared 8 September 2026. The current official Noche UFC card contains thirteen matchups. Likely methods are separate analyst judgements conditional on the picked fighter winning; they have not been validated by a finish-method model.

**The percentages are raw model estimates, not established confidence levels.** The original chronological winner test achieved 65.1%; the revised model has not had an independent future test. Its predecessor was overconfident, and the revised model remains vulnerable to sparse records.

The reasons below describe supporting and opposing historical evidence. They are not mathematical feature attributions or proven causes. Current news context is described separately from what was present in the frozen model inputs.

**Each percentage refers only to the winner, not to winning by the suggested method.** No method probabilities or exact rounds are estimated.

| Matchup | Pick and raw estimate | Why, and the main reservation |
|---|---|---|
| Jean Silva vs Jose Miguel Delgado | **Jean Silva — 51%** | Silva has the stronger opponent history, including his win over Arnold Allen. Delgado has the better recorded striking margin and longer reach. This is effectively even. |
| Brandon Moreno vs Joseph Morales | **Brandon Moreno — 63%** | Moreno has stronger opposition experience, higher striking output and better recorded takedown defence. His two consecutive losses and Morales's two submission wins temper the pick. |
| Tommy McMillen vs Marwan Rahiki | **Tommy McMillen — 69%** | McMillen's two UFC wins show an exceptional striking-volume advantage; he landed 252 significant strikes against Montes. Both fighters have only two UFC bouts, so this estimate is fragile. |
| Manon Fiorot vs Alexa Grasso | **Manon Fiorot — 61%** | Fiorot has the stronger recorded striking differential and takedown defence, with an 8–1 UFC record. Grasso's recent stoppage of Maycee Barber is clear counterevidence. |
| Waldo Cortes Acosta vs Curtis Blaydes | **Waldo Cortes Acosta — 58%** | Cortes-Acosta has three stoppage wins in his last five and the better recent knockdown rate. Blaydes's wrestling and control history give him a clear route to overturn the pick. |
| David Martinez vs Dan Ige | **David Martinez — 72%** | Martinez is 3–0 in the UFC, including wins over Font and Vera; Ige is 1–4 in his last five. Ige's first bantamweight appearance makes the division change an additional uncertainty. |
| Tim Elliott vs Edgar Chairez | **Edgar Chairez — 63%** | Chairez has three straight wins and a substantial age advantage. Elliott's takedowns and control remain a serious counter: the model's lean does not establish that Chairez can stop his wrestling. |
| Ignacio Bahamondes vs Muslim Salikhov | **Ignacio Bahamondes — 68%** | Bahamondes has the youth, reach and historical striking-output advantages. He has lost his last two, and his move from lightweight to welterweight makes past rates imperfect comparisons. |
| Yousri Belgaroui vs Djorden Santos | **Yousri Belgaroui — 82%** | Belgaroui outlanded his first two UFC opponents 196–40 in significant strikes, winning both by stoppage. That supports the direction; two bouts and his conceded takedowns do not justify treating 82% as dependable. |
| Drakkar Klose vs Tommy Gantt | **Tommy Gantt — 73%** | Gantt's debut showed six takedowns, substantial control and a TKO. However, it was his only UFC bout. Klose is 10–3 in the archive, and Elo favours Klose: this is a particularly questionable confidence estimate. |
| Rafa Garcia vs Rongzhu | **Rafa Garcia — 61% (provisional)** | Garcia's recent wins and opponent-adjusted record support a modest lean. Rongzhu also has three straight wins. Adding his omitted May Road result reduces Garcia's estimate from 64.0% to 60.5%; full statistics were not added. |
| Sean King vs Jessie Rosas | **Sean King — 51%** | Neither fighter has a recorded UFC bout in the model. There is no demonstrated performance advantage behind this small lean; treat the fight as a toss-up. |
| JJ Aldrich vs Regina Tarin | **Regina Tarin — 57%** | Tarin showed high output in her debut win, but it is one fifteen-minute sample. Aldrich has won four of her last five; both Elo and the older model favour Aldrich. This is a disputed, low-evidence pick. |

## Likely means of victory

These are routes I would lean towards **if the picked fighter wins**. Historical counts are descriptive, not method probabilities; opponent quality and small samples matter. Decision means a judges' decision without predicting whether it will be unanimous or split. KO/TKO includes a referee or doctor stoppage from strikes.

| Pick | Likely method | Why this route; main uncertainty |
|---|---|---|
| **Jean Silva** | **KO/TKO** | Four of Silva's six captured UFC wins were stoppages by strikes, including one doctor stoppage; five were finishes overall. That makes a striking stoppage the method lean if he wins. Delgado has never been stopped in his five captured UFC bouts; his only defeat and his last three fights went to decisions. Silva also won his latest bout by decision, and the winner forecast itself is close. |
| **Brandon Moreno** | **Decision** | Five of Moreno's eleven captured UFC wins were decisions, including his two latest wins over Albazi and Erceg. A points win is the cautious lean. Moreno has six UFC finish wins too, so decision is only the largest single category. Morales has a prior KO loss and won all three of his captured UFC victories by submission. |
| **Tommy McMillen** | **KO/TKO** | Both of McMillen's UFC wins ended by strikes, against Zecchini and Montes. His observed UFC winning methods support a TKO lean. This is only two UFC bouts. Rahiki is also 2–0 with two stoppages and no captured UFC loss, so there is no demonstrated UFC stoppage vulnerability to rely on. |
| **Manon Fiorot** | **Decision** | Five of Fiorot's eight UFC wins were decisions. Four of Grasso's five UFC defeats also went to the judges. Fiorot stopped Jasudavicius in her latest bout, and Grasso returned with a knockout of Barber. Their recent finishes keep a striking finish possible. |
| **Waldo Cortes Acosta** | **KO/TKO** | Cortes Acosta has five KO/TKO wins in ten UFC victories. Five of Blaydes' six UFC defeats were striking stoppages, including a doctor stoppage; Waldo's latest three wins were knockouts. Half of Waldo's UFC wins were decisions, and Blaydes' latest defeat went the distance. Blaydes also offers a wrestling route that makes the winner call uncertain. |
| **David Martinez** | **Decision** | Martinez won his last two UFC bouts by decision; nine of Ige's ten UFC defeats went to the judges. Ige was knocked out in his latest bout, and Martinez has a UFC knockout win. This is Ige's first announced bantamweight bout, so prior featherweight durability may not transfer directly. |
| **Edgar Chairez** | **Submission** | Three of Chairez' four UFC wins came by submission, while Elliott has five submission losses in twelve UFC defeats. Elliott's other seven defeats went to decisions, and he has three UFC submission wins himself. Chairez also needed a split decision against Bunes. |
| **Ignacio Bahamondes** | **KO/TKO** | Three of Bahamondes' six UFC wins were KO/TKOs; Salikhov has three KO/TKO losses among five UFC defeats, including his latest bout. Bahamondes also has two submission wins, and his last two bouts were decision losses. Moving from lightweight to welterweight changes the matchup context; prior finishing power is not proven at this new weight. |
| **Yousri Belgaroui** | **KO/TKO** | Both Belgaroui UFC wins ended by strikes, against Bekoev and Abdul-Malik. Only two UFC bouts are available. Santos has not been knocked out in his three captured UFC bouts; his defeats were a decision and a submission. |
| **Tommy Gantt** | **KO/TKO** | Gantt's only UFC win was a TKO, and two of Klose's three UFC defeats were KO/TKOs. Gantt's current official biography also reports six consecutive finishes, comprising four TKOs and two guillotines. One UFC appearance is a very small sample. Klose has gone to decisions in nine of thirteen UFC bouts, and Gantt's official biography lists five professional submission wins, so submission or decision remain credible alternatives. |
| **Rafa Garcia** | **Decision** | Five of Garcia's seven UFC wins were decisions, including two of his latest three wins. His captured UFC history supports a points win. Garcia also stopped Gordon recently. Rongzhu's additional May 2026 TKO win and three older Road wins sit outside the frozen six-bout history; the disclosed one-result update remains incomplete. |
| **Sean King** | **Unclear** | There is not enough reliable matchup evidence to choose a method. Both fighters are making their UFC debuts. Regional sources suggest a possible striking-finish route for King but disagree on his complete professional record. Rosas' official profile lists two KO wins and four submission wins, so the absence of UFC history does not imply a lack of finishing ability. |
| **Regina Tarin** | **Decision** | Tarin won her only UFC bout by decision. Aldrich has gone to decisions in fourteen of seventeen UFC bouts, including four of her six losses. Tarin has only one UFC appearance. Her current official biography lists four professional KO wins and two submission wins, so her wider record contains substantial finishing threat. |

Method evidence is drawn from the bout histories in the linked evidence file. Any supplementary public source is recorded against its matchup there. These additions do not change the winner probabilities or the original sealed forecasts.


**My assessment of the evidence:** Martinez, Fiorot and Moreno have more defensible supporting histories than the small-sample Gantt forecast. Silva–Delgado and King–Rosas are essentially even. Gantt, Tarin and the exact 82% assigned to Belgaroui deserve particular scepticism.

## The Garcia–Rongzhu correction

UFC's current Rongzhu profile confirms a first-round TKO over Victor Martinez at 1:10 on 28 May 2026, at Road To UFC Season 5 Opening Rounds. That result is absent from the original archive. A separate calculation inserts the verified outcome, opponent and duration chronologically, keeping its unavailable paired statistics missing. With the exact same fitted V3 model, Garcia changes from 63.9710% to 60.5118%. The other twelve estimates are unchanged within numerical precision. This is a limited sensitivity calculation, not a complete supplementary-statistics update.

A subsequent finite check of all 26 current fighters' official pages found this was the only confirmed missing recent UFC/Road result among their dated 2025–26 narratives. Rongzhu's profile also lists three older Road tournament wins absent from the archive. Those older wins have not been inserted into this one-result calculation, so 60.5% remains a partial-history estimate. No complete professional-career coverage is claimed.

The original 7 September forecasts, original data, fitted models and future trial record remain preserved. This separately dated user briefing does not silently replace the original sealed test issue or the workbench's original data.

## Sources and evidence

- [Current official card](https://www.ufc.com/event/ufc-fight-night-september-12-2026), verified live 8 September.
- [UFC fight-by-fight preview](https://www.ufc.com/news/fight-fight-preview-noche-ufc-phoenix-silva-delgado-2026): Ige's bantamweight move, Bahamondes's welterweight move and current form.
- [UFC on Delgado's late replacement](https://www.ufc.com/news/jose-miguel-delgado-from-yuma-with-love-noche-ufc): three weeks to prepare for five rounds; this contextual fact did not alter the sealed forecast.
- [UFC record book](https://statleaders.ufc.com/fight) and [McMillen feature](https://www.ufc.com/news/theres-no-party-like-a-tommy-mcmillen-party-noche-ufc): 252 significant strikes against Montes.
- [Rongzhu's official profile](https://www.ufc.com/athlete/zhu-rong?page=1): omitted Road result and current three-fight winning streak.
- [Machine-readable evidence](September-12-Forecast-Evidence.json): saved probabilities, exact archived profiles, bout histories and the separate result-only calculation.

The older Ramiro Jimenez–Rodrigo Vera pairing is absent from the current official card and is excluded from these thirteen current picks. Its original alternative-source forecast remains preserved.
