UFC Fight Night: Nurmagomedov vs. Song  29 Aug 2026 · Oriental Sports Center, Shanghai, China, CN

Sumudaerji defeats Alex Perez by unanimous decision

Loss Alex Perez 26-11-0
Win Sumudaerji 20-7-0
Unanimous decision

Round 3 · 5:00

Perez vs Sumudaerji — Flyweight · UFC Fight Night: Nurmagomedov vs. Song

Predictions summary

What happened

The result, the defining moments, and what the verdict means.

Sumudaerji had the stronger striking at range while Alex Perez found success when he took the fight to the canvas.

In the deciding round, Sumudaerji defended the takedown attempts and scored in the clinch to earn a unanimous decision.

Pre-fight analysis — how we saw it going in

Case for Perez

They met in May and it was waved off at 1:45 of the second round after an accidental groin kick, with nothing resolved. Perez is the more varied finisher — seven knockouts and seven submissions among twenty-six wins, 2.07 takedowns and 0.83 submission attempts per 15 minutes — and the more volatile: he averages 6:47 a fight.

Case for Sumudaerji

Sumudaerji has won three straight, all of them on the cards, and is the tidier striker on paper: 4.40 significant strikes landed a minute against 2.29 absorbed, behind 63% defence. Thirteen of his nineteen wins are knockouts, but the current version banks rounds, and three banked rounds wins this.

What We Were Watching For

  • A rematch of a no contest — their May meeting ended at 1:45 of round two on an accidental groin kick.
  • Both men lose the same way: six of Perez's ten defeats are submissions, and six of Sumudaerji's seven are.
  • Perez's last defeat was Asu Almabayev's flying guillotine, 22 seconds into the third round.
  • Sumudaerji's last three wins are all decisions. Perez has not won one since March 2019.
Records & career stats
Perez vs Sumudaerji
Perez Sumudaerji
  1. OutputSig. strikes landed / min 3.79 4.21
  2. Striking def.Striking defense 58% 64%
  3. DurabilitySig. strikes absorbed / min 3.18 2.10
  4. Takedown def.Takedown defense 76% 70%
  5. FinishingFinish rate 54% 70%
26-10-0
Record
19-7-0
—
Recent formRecent formWin–loss record over the fighter’s last five bouts, newest first, as carried in their UFC/FightMetric career stats.
—
United States
Nationality
China
34
Age
30
5′6″168 cm
Height — Sumudaerji leads
5′8″173 cm
65″166 cm
Reach — Sumudaerji leads
72″183 cm
—
StanceStanceWhich side a fighter leads from — a striking term, not a religious one. Orthodox: right-handed, left foot and hand forward (the common default). Southpaw: the mirror image, right side forward (usually a left-hander). Switch: changes between the two during a fight.
—
54%
Finish rateFinish % — Sumudaerji leadsFinish rateThe share of a fighter’s wins that ended inside the distance — by KO/TKO or submission — rather than going to the judges’ scorecards. Higher means more finishes.
70%
7
Wins by KO/TKOKO wins — Sumudaerji leads
13
7
Wins by submissionSub wins — Perez leads
1
3.79
Sig. strikes landed / minStr / min — Sumudaerji leadsSignificant strikes landed per minuteAverage number of significant strikes a fighter lands for every minute of fight time — a read on striking output and pace. From UFC/FightMetric career stats.
4.21
3.18
Sig. strikes absorbed / minAbs / min — Sumudaerji leadsSignificant strikes absorbed per minuteAverage number of significant strikes a fighter takes for every minute of fight time. Lower is better — it means they get hit less often. From UFC/FightMetric career stats.
2.10
58%
Striking defenseStrike def. — Sumudaerji leadsStriking defenseThe percentage of opponents’ significant strikes that do not land — strikes slipped, blocked or evaded. Higher is better. Not the same as striking accuracy, which counts your own strikes that land. From UFC/FightMetric.
64%
76%
Takedown defenseTD def. — Perez leadsTakedown defenseThe percentage of opponents’ takedown attempts that are stuffed. Higher means harder to put on the mat. From UFC/FightMetric.
70%
07:16
Avg. fight timeAvg. timeAverage fight timeThe mean length of the fighter’s bouts (minutes:seconds) — a rough read on whether they tend to finish early or go the distance. From UFC/FightMetric.
11:05
Team Oyama
Team
Team Alpha Male

Career stats per UFC.com; recent form is each fighter's last five bouts. Record is as of fight night; career stats, recent form and win totals are the latest on file and may include this fight. Bars compare each stat on a fixed scale; the “Edge” marker flags the leading corner where the gap is meaningful. The snapshot plots five career rates — output, striking defense, durability (strikes absorbed, inverted), takedown defense and finish rate — each normalised on its own scale.

Comments

Comments are currently closed while we finish moderation and privacy controls.

Who Called It?

How the analysts' picks held up once the cage door closed.

The MMA Guru
Hear it — 23:22 The MMA Guru on YouTube ↗

Current pick evidence

Picked: Alex Perez

Fight result: Sumudaerji · Decision

Picks Alex Perez in the rematch without committing to a method.

“I'm going with Alex Perez.”

Track record: winner called in 213 of 322 resolved picks. How winner calls are counted

Analyst track record Track record countsWinner counts resolved calls. Method and Round count only after a correct winner call. Underdog counts picks for a strict underdog in frozen pre-fight odds.Methodology

Winner Accuracy 66.1% (213 hit, 109 miss, across 322 resolved calls). Method Accuracy 53.4% (62 hit, 54 miss, across 116 method-scored calls on 322 resolved calls). Round Accuracy 35.9% (14 hit, 25 miss, across 39 round-scored calls on 322 resolved calls). Underdog Accuracy 47.4% (18 hit, 20 miss, across 38 underdog calls on 322 resolved calls).

View full track record (322 calls)
Bedtime MMA
Hear it — 23:29 Bedtime MMA on YouTube ↗

Current pick evidence

Picked: Sumudaerji · Dec

Fight result: Sumudaerji · Decision

Picks Sumudaerji by decision.

“I think Sumudaerji wins by decision.”

Track record: winner called in 132 of 185 resolved picks. How winner calls are counted

Analyst track record Track record countsWinner counts resolved calls. Method and Round count only after a correct winner call. Underdog counts picks for a strict underdog in frozen pre-fight odds.Methodology

Winner Accuracy 71.4% (132 hit, 53 miss, across 185 resolved calls). Method Accuracy 61.9% (60 hit, 37 miss, across 97 method-scored calls on 185 resolved calls). Round Accuracy 40.7% (11 hit, 16 miss, across 27 round-scored calls on 185 resolved calls). Underdog Accuracy 48.7% (19 hit, 20 miss, across 39 underdog calls on 185 resolved calls).

View full track record (185 calls)
TheWeasle
Hear it — 20:02 TheWeasle on YouTube ↗

Current pick evidence

Picked: Sumudaerji · KO/TKO · Round 3

Fight result: Sumudaerji · Decision

Picks Sumudaerji by third-round knockout.

“So, I'm going to go with Sumidarji by a third round TKO.”

Track record: winner called in 115 of 163 resolved picks. How winner calls are counted

Analyst track record Track record countsWinner counts resolved calls. Method and Round count only after a correct winner call. Underdog counts picks for a strict underdog in frozen pre-fight odds.Methodology

Winner Accuracy 70.6% (115 hit, 48 miss, across 163 resolved calls). Method Accuracy 61.8% (47 hit, 29 miss, across 76 method-scored calls on 163 resolved calls). Round Accuracy 36.5% (19 hit, 33 miss, across 52 round-scored calls on 163 resolved calls). Underdog Accuracy 52.9% (9 hit, 8 miss, across 17 underdog calls on 163 resolved calls).

View full track record (163 calls)
FightIQ
Read source

Current pick evidence

Picked: Sumudaerji · Dec

Fight result: Sumudaerji · Decision

FightIQ's public record lists Sumudaerji as its pick over Alex Perez.

Track record: winner called in 159 of 227 resolved picks. How winner calls are counted

Analyst track record Track record countsWinner counts resolved calls. Method and Round count only after a correct winner call. Underdog counts picks for a strict underdog in frozen pre-fight odds.Methodology

Winner Accuracy 70.0% (159 hit, 68 miss, across 227 resolved calls). Method Accuracy 39.9% (59 hit, 89 miss, across 148 method-scored calls on 227 resolved calls). Round Accuracy: No round picks. Underdog Accuracy 55.0% (11 hit, 9 miss, across 20 underdog calls on 227 resolved calls).

View full track record (227 calls)
AJ Devito
Hear it — 55:26 AJ Devito on YouTube ↗

Current pick evidence

Picked: Sumudaerji

Fight result: Sumudaerji · Decision

Picks Sumudaerji but leaves the method open between a knockout and decision.

“My prediction is Sumudaerji by knockout or decision. It's hard to say.”

Track record: winner called in 175 of 264 resolved picks. How winner calls are counted

Analyst track record Track record countsWinner counts resolved calls. Method and Round count only after a correct winner call. Underdog counts picks for a strict underdog in frozen pre-fight odds.Methodology

Winner Accuracy 66.3% (175 hit, 89 miss, across 264 resolved calls). Method Accuracy 58.7% (61 hit, 43 miss, across 104 method-scored calls on 264 resolved calls). Round Accuracy 77.8% (7 hit, 2 miss, across 9 round-scored calls on 264 resolved calls). Underdog Accuracy 36.4% (12 hit, 21 miss, across 33 underdog calls on 264 resolved calls).

View full track record (264 calls)
JakeGerberMMA
Hear it — 48:31 JakeGerberMMA on YouTube ↗

Current pick evidence

Picked: Sumudaerji · Dec

Fight result: Sumudaerji · Decision

Picks Sumudaerji by decision for his speed, footwork, and scrambling against wrestling.

“Sumudaerji by decision.”

Track record: winner called in 101 of 154 resolved picks. How winner calls are counted

Analyst track record Track record countsWinner counts resolved calls. Method and Round count only after a correct winner call. Underdog counts picks for a strict underdog in frozen pre-fight odds.Methodology

Winner Accuracy 65.6% (101 hit, 53 miss, across 154 resolved calls). Method Accuracy 60.0% (39 hit, 26 miss, across 65 method-scored calls on 154 resolved calls). Round Accuracy 35.7% (5 hit, 9 miss, across 14 round-scored calls on 154 resolved calls). Underdog Accuracy 33.3% (4 hit, 8 miss, across 12 underdog calls on 154 resolved calls).

View full track record (154 calls)
DFS By The Numbers
Hear it — 9:41 DFS By The Numbers on YouTube ↗

Current pick evidence

Picked: Sumudaerji · Dec

Fight result: Sumudaerji · Decision

Picks Sumudaerji by decision after flipping the selection.

“I'm flipping my pick to Sumudaerji. If Sumudaerji wins, I think it's by decision and that's my pick.”

Track record: winner called in 112 of 173 resolved picks. How winner calls are counted

Analyst track record Track record countsWinner counts resolved calls. Method and Round count only after a correct winner call. Underdog counts picks for a strict underdog in frozen pre-fight odds.Methodology

Winner Accuracy 64.7% (112 hit, 61 miss, across 173 resolved calls). Method Accuracy 61.0% (61 hit, 39 miss, across 100 method-scored calls on 173 resolved calls). Round Accuracy 38.3% (18 hit, 29 miss, across 47 round-scored calls on 173 resolved calls). Underdog Accuracy 36.4% (8 hit, 14 miss, across 22 underdog calls on 173 resolved calls).

View full track record (173 calls)
We Want Picks
Hear it — 10:08 We Want Picks on YouTube ↗

Current pick evidence

Picked: Sumudaerji

Fight result: Sumudaerji · Decision

Angelo Bodetti of We Want Picks includes Sumudaerji in his parlay.

“I have that parlayed with Sumudaerji.”

Track record: winner called in 93 of 145 resolved picks. How winner calls are counted

Analyst track record Track record countsWinner counts resolved calls. Method and Round count only after a correct winner call. Underdog counts picks for a strict underdog in frozen pre-fight odds.Methodology

Winner Accuracy 64.1% (93 hit, 52 miss, across 145 resolved calls). Method Accuracy 40.0% (2 hit, 3 miss, across 5 method-scored calls on 145 resolved calls). Round Accuracy: No qualifying picks. Underdog Accuracy 34.5% (10 hit, 19 miss, across 29 underdog calls on 145 resolved calls).

View full track record (145 calls)
HopperMMA
Hear it — 16:27 HopperMMA on YouTube ↗

Current pick evidence

Picked: Sumudaerji · Dec

Fight result: Sumudaerji · Decision

Picks Sumudaerji by decision.

“I'm going to go Sumudaerji by decision.”

Track record: winner called in 92 of 148 resolved picks. How winner calls are counted

Analyst track record Track record countsWinner counts resolved calls. Method and Round count only after a correct winner call. Underdog counts picks for a strict underdog in frozen pre-fight odds.Methodology

Winner Accuracy 62.2% (92 hit, 56 miss, across 148 resolved calls). Method Accuracy 60.8% (45 hit, 29 miss, across 74 method-scored calls on 148 resolved calls). Round Accuracy 48.8% (20 hit, 21 miss, across 41 round-scored calls on 148 resolved calls). Underdog Accuracy 26.9% (7 hit, 19 miss, across 26 underdog calls on 148 resolved calls).

View full track record (148 calls)
MMA Prediction Guru (MMAPG)
Hear it — 44:10 MMA Prediction Guru (MMAPG) on YouTube ↗

Current pick evidence

Picked: Sumudaerji

Fight result: Sumudaerji · Decision

Picks Sumudaerji while leaving the exact result between decision and knockout unresolved.

“Going to go Sumudaerji, either dominant decision or KO-TKO victory.”

Track record: winner called in 41 of 69 resolved picks. How winner calls are counted

Analyst track record Track record countsWinner counts resolved calls. Method and Round count only after a correct winner call. Underdog counts picks for a strict underdog in frozen pre-fight odds.Methodology

Winner Accuracy 59.4% (41 hit, 28 miss, across 69 resolved calls). Method Accuracy 57.7% (15 hit, 11 miss, across 26 method-scored calls on 69 resolved calls). Round Accuracy 12.5% (1 hit, 7 miss, across 8 round-scored calls on 69 resolved calls). Underdog Accuracy 30.8% (4 hit, 9 miss, across 13 underdog calls on 69 resolved calls).

View full track record (69 calls)
Andrew Richardson
Read source

Current pick evidence

Picked: Sumudaerji · Dec

Fight result: Sumudaerji · Decision

MMA Mania's UFC Shanghai X-Factor preview has Andrew Richardson picking Sumudaerji by decision.

Track record: winner called in 37 of 66 resolved picks. How winner calls are counted

Analyst track record Track record countsWinner counts resolved calls. Method and Round count only after a correct winner call. Underdog counts picks for a strict underdog in frozen pre-fight odds.Methodology

Winner Accuracy 56.1% (37 hit, 29 miss, across 66 resolved calls). Method Accuracy 73.0% (27 hit, 10 miss, across 37 method-scored calls on 66 resolved calls). Round Accuracy: No qualifying picks. Underdog Accuracy 10.0% (1 hit, 9 miss, across 10 underdog calls on 66 resolved calls).

View full track record (66 calls)
Half The Battle
Hear it — 13:40 Half The Battle on YouTube ↗

Current pick evidence

Picked: Sumudaerji · KO/TKO

Fight result: Sumudaerji · Decision

Picks Sumudaerji to stop Alex Perez, citing their prior meeting.

“Sue is going to come out here and knock out Alex Perez.”

Track record: winner called in 58 of 109 resolved picks. How winner calls are counted

Analyst track record Track record countsWinner counts resolved calls. Method and Round count only after a correct winner call. Underdog counts picks for a strict underdog in frozen pre-fight odds.Methodology

Winner Accuracy 53.2% (58 hit, 51 miss, across 109 resolved calls). Method Accuracy 71.4% (10 hit, 4 miss, across 14 method-scored calls on 109 resolved calls). Round Accuracy 66.7% (2 hit, 1 miss, across 3 round-scored calls on 109 resolved calls). Underdog Accuracy 21.4% (6 hit, 22 miss, across 28 underdog calls on 109 resolved calls).

View full track record (109 calls)

Analyst picks: Bedtime MMA Hit; TheWeasle Hit; FightIQ Hit; AJ Devito Hit; The MMA Guru Miss; JakeGerberMMA Hit; DFS By The Numbers Hit; We Want Picks Hit; HopperMMA Hit; MMA Prediction Guru (MMAPG) Hit; Andrew Richardson Hit; Half The Battle Hit.

11 of 12 nailed it.

Chart: selected analyst picks, not a win probability. Pill badge: Winner hit rate (total scored picks before this event).

The experts above are a lens, not credentials we conferred — analyst, fighter or betting desk.

Results

Each agent’s latest pre-fight call vs result; forecast dates may differ. Market: sportsbook snapshot or historic market reconstruction.

Actual Sumudaerji · Decision · R3

Closest probability ChatGPT and Claude put 64% on Sumudaerji.

ChatGPTForecast: 63 of 108 winner picks correct · 58% hit rate
36% Perez
64% Sumudaerji
PickSumudaerji · DecisionActualSumudaerji · Decision · R3
WinnerHitMethodHitRoundNot applicable — decisions have no finishing roundHistoric market favouriteHistoric market favourite: FavouriteThe selected fighter had the higher market-implied chance of winning.Historic market favourite: favourite call
ClaudeForecast: 61 of 101 winner picks correct · 60% hit rate
36% Perez
64% Sumudaerji
PickSumudaerji · DecisionActualSumudaerji · Decision · R3
WinnerHitMethodHitRoundNot applicable — decisions have no finishing roundHistoric market favouriteHistoric market favourite: FavouriteThe selected fighter had the higher market-implied chance of winning.Historic market favourite: favourite call
GeminiForecast: 68 of 108 winner picks correct · 63% hit rate
45% Perez
55% Sumudaerji
PickSumudaerji · KO/TKO · R2ActualSumudaerji · Decision · R3
WinnerHitMethodMissRoundNot applicable — decisions have no finishing roundHistoric market favouriteHistoric market favourite: FavouriteThe selected fighter had the higher market-implied chance of winning.Historic market favourite: favourite call
AI Predictions — latest pre-fight callsFrozen final publication identity; the calls below retain each agent’s own publication.

AI Predictions

Each agent’s latest pre-fight forecast. Dates are shown per agent.

  • ChatGPT

    63 of 108 winner picks correct · 58% hit rate

    Pick Sumudaerji

    Most likely: Sumudaerji by decision, about 42 in 100

    Win chance: Perez 36%, Sumudaerji 64%.

    gpt-5.6-sol · 23 Aug

    Confidence 6 of 10 · Medium

  • Claude

    61 of 101 winner picks correct · 60% hit rate

    Pick Sumudaerji

    Most likely: Sumudaerji by decision, about 33 in 100

    Win chance: Perez 36%, Sumudaerji 64%.

    claude-opus-5 · 24 Aug

    Confidence 6 of 10 · Medium

  • Gemini

    68 of 108 winner picks correct · 63% hit rate

    Pick Sumudaerji

    Most likely: a tie between Sumudaerji by KO/TKO and Sumudaerji by decision, about 25 in 100 each

    Win chance: Perez 45%, Sumudaerji 55%.

    gemini-3.1-pro-high · 24 Aug

    Confidence 6 of 10 · Medium

3 of 3 models pick Sumudaerji.

Questions

The result 7 questions

Who wins, and how likely?

  • ChatGPT

    Sumudaerji: 64% (about 64 in 100 fights). Perez: 36%.

  • Claude

    Sumudaerji: 64% (about 64 in 100 fights). Perez: 36%.

  • Gemini

    Sumudaerji: 55% (about 55 in 100 fights). Perez: 45%.

What is the single most likely outcome?

  • ChatGPT

    Sumudaerji by decision: 42% (about 42 in 100).

  • Claude

    Sumudaerji by decision: 33% (about 33 in 100).

  • Gemini

    A tie: Sumudaerji by KO/TKO and Sumudaerji by decision, 25% each (about 25 in 100 each).

What are the three most likely outcomes, and why?

  • ChatGPT

    1. Sumudaerji by decision (42%); 2. Sumudaerji by KO/TKO (18%); 3. Perez by submission (15%)

  • Claude

    1. Sumudaerji by decision (33%); 2. Sumudaerji by KO/TKO (28%); 3. Perez by submission (13%)

  • Gemini

    1. Sumudaerji by KO/TKO (25%); 1. Sumudaerji by decision (25%); 3. Perez by decision (20%)

How confident is the model?

  • ChatGPT

    6 of 10, based on Sumudaerji's 64% win chance.

  • Claude

    6 of 10, based on Sumudaerji's 64% win chance.

  • Gemini

    6 of 10, based on Sumudaerji's 55% win chance.

Does it end early or go to the judges?

  • ChatGPT

    Finish (KO/TKO or submission): 47%. Decision: 53%.

  • Claude

    Finish (KO/TKO or submission): 54%. Decision: 46%.

  • Gemini

    Finish (KO/TKO or submission): 55%. Decision: 45%.

How does Perez win?

  • ChatGPT

    Perez wins 36%: KO/TKO 10%, submission 15%, decision 11%.

  • Claude

    Perez wins 36%: KO/TKO 10%, submission 13%, decision 13%.

  • Gemini

    Perez wins 45%: KO/TKO 10%, submission 15%, decision 20%.

How does Sumudaerji win?

  • ChatGPT

    Sumudaerji wins 64%: KO/TKO 18%, submission 4%, decision 42%.

  • Claude

    Sumudaerji wins 64%: KO/TKO 28%, submission 3%, decision 33%.

  • Gemini

    Sumudaerji wins 55%: KO/TKO 25%, submission 5%, decision 25%.

What if 1 question

What is the upset path?

  • Gemini

    Perez's most likely upset: Perez by decision (20%).

Compared with others 1 question

Do the models agree?

  • All models

    The models agree: all 3 pick Sumudaerji.

After the fight 2 questions

What happened vs what the model said?

  • ChatGPT

    Result: Sumudaerji by decision, the model's most likely outcome.

  • Claude

    Result: Sumudaerji by decision, the model's most likely outcome.

  • Gemini

    Result: Sumudaerji by decision, the model's joint number 1 outcome. Its most likely outcomes were Sumudaerji by KO/TKO and Sumudaerji by decision (25% each).

How much chance did it give the actual result?

  • ChatGPT

    42% (about 42 in 100).

  • Claude

    33% (about 33 in 100).

  • Gemini

    25% (about 25 in 100).

More questions (19)

The result 6 questions

KO/TKO vs submission vs decision overall

  • ChatGPT

    KO/TKO 28%, submission 19%, decision 53%.

  • Claude

    KO/TKO 38%, submission 16%, decision 46%.

  • Gemini

    KO/TKO 35%, submission 20%, decision 45%.

Who is more likely to finish the other?

  • ChatGPT

    Perez: 25%, against 22% for Sumudaerji.

  • Claude

    Sumudaerji: 31%, against 23% for Perez.

  • Gemini

    Sumudaerji: 30%, against 25% for Perez.

Can Perez get the finish? Can Sumudaerji?

  • ChatGPT

    Perez: 25% (KO/TKO 10%, submission 15%). Sumudaerji: 22% (KO/TKO 18%, submission 4%).

  • Claude

    Perez: 23% (KO/TKO 10%, submission 13%). Sumudaerji: 31% (KO/TKO 28%, submission 3%).

  • Gemini

    Perez: 25% (KO/TKO 10%, submission 15%). Sumudaerji: 30% (KO/TKO 25%, submission 5%).

Who is more likely to be knocked out, and who to be submitted?

  • ChatGPT

    Perez is knocked out in 18% and submitted in 4%. Sumudaerji is knocked out in 10% and submitted in 15%.

  • Claude

    Perez is knocked out in 28% and submitted in 3%. Sumudaerji is knocked out in 10% and submitted in 13%.

  • Gemini

    Perez is knocked out in 25% and submitted in 5%. Sumudaerji is knocked out in 10% and submitted in 15%.

How lopsided is this fight?

  • ChatGPT

    Lopsided: 64% vs 36%.

  • Claude

    Lopsided: 64% vs 36%.

  • Gemini

    A clear favourite: 55% vs 45%.

How big an upset would it be if the underdog wins?

  • ChatGPT

    Perez wins 36% (about 1 in 3 fights): a big upset.

  • Claude

    Perez wins 36% (about 1 in 3 fights): a big upset.

  • Gemini

    Perez wins 45% (about 1 in 2 fights): a modest upset.

Timing 1 question

Chance it goes to a decision

  • ChatGPT

    53%.

  • Claude

    46%.

  • Gemini

    45%.

What if 1 question

If it goes to the judges, who wins?

  • ChatGPT

    If it goes to the judges: Perez 21%, Sumudaerji 79%.

  • Claude

    If it goes to the judges: Perez 28%, Sumudaerji 72%.

  • Gemini

    If it goes to the judges: Perez 44%, Sumudaerji 56%.

Compared with others 5 questions

Where does this model differ from the others?

  • ChatGPT

    Biggest difference: Sumudaerji by decision, 42% here against 29% on average for the others.

  • Gemini

    Biggest difference: Sumudaerji by decision, 25% here against 38% on average for the others.

How does it compare with analyst picks?

  • ChatGPT

    Of 12 analyst picks, 1 backs Perez and 11 back Sumudaerji. This model picks Sumudaerji (64%).

  • Claude

    Of 12 analyst picks, 1 backs Perez and 11 back Sumudaerji. This model picks Sumudaerji (64%).

  • Gemini

    Of 12 analyst picks, 1 backs Perez and 11 back Sumudaerji. This model picks Sumudaerji (55%).

Which model is most confident on Perez, and which on Sumudaerji?

  • All models

    On Perez: Gemini (45%). On Sumudaerji: ChatGPT and Claude, tied (64%).

Where do the models disagree most?

  • All models

    Sumudaerji by decision: from 25% (Gemini) to 42% (ChatGPT).

Which analyst agrees most with this model?

  • ChatGPT

    Bedtime MMA (Sumudaerji by decision), HopperMMA (Sumudaerji by decision), DFS By The Numbers (Sumudaerji by decision), JakeGerberMMA (Sumudaerji by decision), Andrew Richardson (Sumudaerji by decision) and FightIQ (Sumudaerji by decision): the same winner and method as this model, tied; this model favours Sumudaerji, most likely by decision.

  • Claude

    Bedtime MMA (Sumudaerji by decision), HopperMMA (Sumudaerji by decision), DFS By The Numbers (Sumudaerji by decision), JakeGerberMMA (Sumudaerji by decision), Andrew Richardson (Sumudaerji by decision) and FightIQ (Sumudaerji by decision): the same winner and method as this model, tied; this model favours Sumudaerji, most likely by decision.

  • Gemini

    TheWeasle (Sumudaerji by KO/TKO in round 3), AJ Devito (Sumudaerji), MMA Prediction Guru (MMAPG) (Sumudaerji), Bedtime MMA (Sumudaerji by decision), HopperMMA (Sumudaerji by decision), DFS By The Numbers (Sumudaerji by decision), We Want Picks (Sumudaerji), JakeGerberMMA (Sumudaerji by decision), Half The Battle (Sumudaerji by KO/TKO), Andrew Richardson (Sumudaerji by decision) and FightIQ (Sumudaerji by decision): the same winner as this model, tied; this model favours Sumudaerji.

How sure 4 questions

When was this forecast made, and by which model version?

  • ChatGPT

    On 23 Aug 2026, by gpt-5.6-sol.

  • Claude

    On 24 Aug 2026, by claude-opus-5.

  • Gemini

    On 24 Aug 2026, by gemini-3.1-pro-high.

What can't the model know?

  • FightCard

    Models research public information before the fight, such as records, styles, recent results and coverage. They cannot see inside training camps, injuries nobody has reported, or what happens on the night. A small chance is not no chance: outcomes the model rates at a few percent still happen, and big favourites lose sometimes.

How is "confidence" worked out?

  • FightCard

    Confidence restates the favourite's win chance on a 10-step scale: the win chance in tens, rounded, so 55% is 6 of 10 and 95% or more is 10 of 10. 1 to 3 reads Low, 4 to 6 Medium and 7 to 10 High. A favourite usually has at least half the chance, so most picks show 5 or more. When the model has the fighters exactly even, there is no pick and no confidence level. It is not a separate judgement by the model.

What counts as a KO, a decision or a draw?

  • FightCard

    KO/TKO includes referee, doctor and corner stoppages, and disqualifications. A decision includes a fight stopped early and scored on the judges' cards. A stoppage between rounds counts as a finish in the round that just ended. A draw and a no contest share one outcome.

After the fight 2 questions

Which model was closest?

  • All models

    ChatGPT, which gave the actual result 42%.

Was the actual result in this model's top 3?

  • All models

    Yes.

ChatGPT

Pick

Sumudaerji

Confidence 6 of 10 · Medium

Most likely: Sumudaerji by decision, about 42 in 100

Win chance: Perez 36%, Sumudaerji 64%.

Most likely outcomes

  1. Sumudaerji by decision42%
  2. Sumudaerji by KO/TKO18%
  3. Perez by submission15%
How it ends
  1. KO/TKO: Perez 10%, Sumudaerji 18%.
  2. Submission: Perez 15%, Sumudaerji 4%.
  3. Decision: Perez 11%, Sumudaerji 42%.

AI model's own reasoning — not checked by FightCard

How ChatGPT sees the fight

Sumudaerji is the 64% pick, most likely by decision. His reach, striking efficiency and strong early success in the unfinished first meeting favor him at distance, while Perez's wrestling and Sumudaerji's historical submission vulnerability keep the matchup competitive.

Keys to victory

  • Sumudaerji led the first meeting 26-9 in significant strikes through 6:45 and scored its only knockdown.
  • Sumudaerji has a 6.5-inch reach advantage and stronger career striking accuracy and defensive metrics.
  • Perez's primary upset route is sustained wrestling and submission pressure against an opponent with six submission losses among seven professional defeats.

X-factor

Whether Perez can convert early grappling entries into sustained control; he went 0-for-2 on takedowns in the first meeting but remains the more credible submission threat.

Questions

The result 7 questions

Who wins, and how likely?

  • ChatGPT

    Sumudaerji: 64% (about 64 in 100 fights). Perez: 36%.

Other models' answers

  • Gemini

    Sumudaerji: 55% (about 55 in 100 fights). Perez: 45%.

What is the single most likely outcome?

  • ChatGPT

    Sumudaerji by decision: 42% (about 42 in 100).

Other models' answers

  • Claude

    Sumudaerji by decision: 33% (about 33 in 100).

  • Gemini

    A tie: Sumudaerji by KO/TKO and Sumudaerji by decision, 25% each (about 25 in 100 each).

What are the three most likely outcomes, and why?

  • ChatGPT

    1. Sumudaerji by decision (42%); 2. Sumudaerji by KO/TKO (18%); 3. Perez by submission (15%)

Other models' answers

  • Claude

    1. Sumudaerji by decision (33%); 2. Sumudaerji by KO/TKO (28%); 3. Perez by submission (13%)

  • Gemini

    1. Sumudaerji by KO/TKO (25%); 1. Sumudaerji by decision (25%); 3. Perez by decision (20%)

How confident is the model?

  • ChatGPT

    6 of 10, based on Sumudaerji's 64% win chance.

Other models' answers

  • Gemini

    6 of 10, based on Sumudaerji's 55% win chance.

Does it end early or go to the judges?

  • ChatGPT

    Finish (KO/TKO or submission): 47%. Decision: 53%.

Other models' answers

  • Claude

    Finish (KO/TKO or submission): 54%. Decision: 46%.

  • Gemini

    Finish (KO/TKO or submission): 55%. Decision: 45%.

How does Perez win?

  • ChatGPT

    Perez wins 36%: KO/TKO 10%, submission 15%, decision 11%.

Other models' answers

  • Claude

    Perez wins 36%: KO/TKO 10%, submission 13%, decision 13%.

  • Gemini

    Perez wins 45%: KO/TKO 10%, submission 15%, decision 20%.

How does Sumudaerji win?

  • ChatGPT

    Sumudaerji wins 64%: KO/TKO 18%, submission 4%, decision 42%.

Other models' answers

  • Claude

    Sumudaerji wins 64%: KO/TKO 28%, submission 3%, decision 33%.

  • Gemini

    Sumudaerji wins 55%: KO/TKO 25%, submission 5%, decision 25%.

Compared with others 1 question

Do the models agree?

  • ChatGPT

    The models agree: all 3 pick Sumudaerji.

After the fight 2 questions

What happened vs what the model said?

  • ChatGPT

    Result: Sumudaerji by decision, the model's most likely outcome.

Other models' answers

  • Gemini

    Result: Sumudaerji by decision, the model's joint number 1 outcome. Its most likely outcomes were Sumudaerji by KO/TKO and Sumudaerji by decision (25% each).

How much chance did it give the actual result?

  • ChatGPT

    42% (about 42 in 100).

Other models' answers

  • Claude

    33% (about 33 in 100).

  • Gemini

    25% (about 25 in 100).

More questions (19)

The result 6 questions

KO/TKO vs submission vs decision overall

  • ChatGPT

    KO/TKO 28%, submission 19%, decision 53%.

Other models' answers

  • Claude

    KO/TKO 38%, submission 16%, decision 46%.

  • Gemini

    KO/TKO 35%, submission 20%, decision 45%.

Who is more likely to finish the other?

  • ChatGPT

    Perez: 25%, against 22% for Sumudaerji.

Other models' answers

  • Claude

    Sumudaerji: 31%, against 23% for Perez.

  • Gemini

    Sumudaerji: 30%, against 25% for Perez.

Can Perez get the finish? Can Sumudaerji?

  • ChatGPT

    Perez: 25% (KO/TKO 10%, submission 15%). Sumudaerji: 22% (KO/TKO 18%, submission 4%).

Other models' answers

  • Claude

    Perez: 23% (KO/TKO 10%, submission 13%). Sumudaerji: 31% (KO/TKO 28%, submission 3%).

  • Gemini

    Perez: 25% (KO/TKO 10%, submission 15%). Sumudaerji: 30% (KO/TKO 25%, submission 5%).

Who is more likely to be knocked out, and who to be submitted?

  • ChatGPT

    Perez is knocked out in 18% and submitted in 4%. Sumudaerji is knocked out in 10% and submitted in 15%.

Other models' answers

  • Claude

    Perez is knocked out in 28% and submitted in 3%. Sumudaerji is knocked out in 10% and submitted in 13%.

  • Gemini

    Perez is knocked out in 25% and submitted in 5%. Sumudaerji is knocked out in 10% and submitted in 15%.

How lopsided is this fight?

  • ChatGPT

    Lopsided: 64% vs 36%.

Other models' answers

  • Gemini

    A clear favourite: 55% vs 45%.

How big an upset would it be if the underdog wins?

  • ChatGPT

    Perez wins 36% (about 1 in 3 fights): a big upset.

Other models' answers

  • Gemini

    Perez wins 45% (about 1 in 2 fights): a modest upset.

Timing 1 question

Chance it goes to a decision

  • ChatGPT

    53%.

Other models' answers

  • Claude

    46%.

  • Gemini

    45%.

What if 1 question

If it goes to the judges, who wins?

  • ChatGPT

    If it goes to the judges: Perez 21%, Sumudaerji 79%.

Other models' answers

  • Claude

    If it goes to the judges: Perez 28%, Sumudaerji 72%.

  • Gemini

    If it goes to the judges: Perez 44%, Sumudaerji 56%.

Compared with others 5 questions

Where does this model differ from the others?

  • ChatGPT

    Biggest difference: Sumudaerji by decision, 42% here against 29% on average for the others.

Other models' answers

  • Gemini

    Biggest difference: Sumudaerji by decision, 25% here against 38% on average for the others.

How does it compare with analyst picks?

  • ChatGPT

    Of 12 analyst picks, 1 backs Perez and 11 back Sumudaerji. This model picks Sumudaerji (64%).

Other models' answers

  • Gemini

    Of 12 analyst picks, 1 backs Perez and 11 back Sumudaerji. This model picks Sumudaerji (55%).

Which model is most confident on Perez, and which on Sumudaerji?

  • ChatGPT

    On Perez: Gemini (45%). On Sumudaerji: ChatGPT and Claude, tied (64%).

Where do the models disagree most?

  • ChatGPT

    Sumudaerji by decision: from 25% (Gemini) to 42% (ChatGPT).

Which analyst agrees most with this model?

  • ChatGPT

    Bedtime MMA (Sumudaerji by decision), HopperMMA (Sumudaerji by decision), DFS By The Numbers (Sumudaerji by decision), JakeGerberMMA (Sumudaerji by decision), Andrew Richardson (Sumudaerji by decision) and FightIQ (Sumudaerji by decision): the same winner and method as this model, tied; this model favours Sumudaerji, most likely by decision.

Other models' answers

  • Gemini

    TheWeasle (Sumudaerji by KO/TKO in round 3), AJ Devito (Sumudaerji), MMA Prediction Guru (MMAPG) (Sumudaerji), Bedtime MMA (Sumudaerji by decision), HopperMMA (Sumudaerji by decision), DFS By The Numbers (Sumudaerji by decision), We Want Picks (Sumudaerji), JakeGerberMMA (Sumudaerji by decision), Half The Battle (Sumudaerji by KO/TKO), Andrew Richardson (Sumudaerji by decision) and FightIQ (Sumudaerji by decision): the same winner as this model, tied; this model favours Sumudaerji.

How sure 4 questions

When was this forecast made, and by which model version?

  • ChatGPT

    On 23 Aug 2026, by gpt-5.6-sol.

Other models' answers

  • Claude

    On 24 Aug 2026, by claude-opus-5.

  • Gemini

    On 24 Aug 2026, by gemini-3.1-pro-high.

What can't the model know?

  • FightCard

    Models research public information before the fight, such as records, styles, recent results and coverage. They cannot see inside training camps, injuries nobody has reported, or what happens on the night. A small chance is not no chance: outcomes the model rates at a few percent still happen, and big favourites lose sometimes.

How is "confidence" worked out?

  • FightCard

    Confidence restates the favourite's win chance on a 10-step scale: the win chance in tens, rounded, so 55% is 6 of 10 and 95% or more is 10 of 10. 1 to 3 reads Low, 4 to 6 Medium and 7 to 10 High. A favourite usually has at least half the chance, so most picks show 5 or more. When the model has the fighters exactly even, there is no pick and no confidence level. It is not a separate judgement by the model.

What counts as a KO, a decision or a draw?

  • FightCard

    KO/TKO includes referee, doctor and corner stoppages, and disqualifications. A decision includes a fight stopped early and scored on the judges' cards. A stoppage between rounds counts as a finish in the round that just ended. A draw and a no contest share one outcome.

After the fight 2 questions

Which model was closest?

  • ChatGPT

    ChatGPT, which gave the actual result 42%.

Was the actual result in this model's top 3?

  • ChatGPT

    Yes.

gpt-5.6-sol · 23 Aug63 of 108 winner picks correct · 58% hit rate

Claude

Pick

Sumudaerji

Confidence 6 of 10 · Medium

Most likely: Sumudaerji by decision, about 33 in 100

Win chance: Perez 36%, Sumudaerji 64%.

Most likely outcomes

  1. Sumudaerji by decision33%
  2. Sumudaerji by KO/TKO28%
  3. Perez by submission, tied for number 313%
How it ends
  1. KO/TKO: Perez 10%, Sumudaerji 28%.
  2. Submission: Perez 13%, Sumudaerji 3%.
  3. Decision: Perez 13%, Sumudaerji 33%.

AI model's own reasoning — not checked by FightCard

How Claude sees the fight

Immediate rematch of a May 30, 2026 No Contest (groin kick, R2 1:45): Perez won R1 with pressure and front-headlock threats before Sumudaerji dropped him with a straight left early in R2 (source-1, source-9). The market sits at Sumudaerji -195/-235, ~63-65% no-vig (source-10). I lean with it: his +7-inch reach, 2.29 SApM absorbed and 73% takedown defense across three straight decision wins outweigh Perez's guillotine path — real against a man with six submission losses, but Perez is 34, averages 6:47 of cage time, and is on a 13-week turnaround (source-3, source-5, source-6).

Keys to victory

  • Sumudaerji's 7-inch reach edge and straight left dropped Perez in R2 of the first fight; the market now prices him -195/-235 (source-9, source-10)
  • Perez's path is the guillotine: 6 of Sumudaerji's 7 losses are submissions, but his TDD is 73% and he stuffed the R1 takedown attempt (source-3, source-6)
  • Perez is 34 and averages 6:47 per fight; Sumudaerji averages 10:44 with three straight decisions, so distance math favors blue (source-5, source-6)

X-factor

Foul recurrence: the first bout featured an eye poke and low blows, so another point deduction or no contest is a live outcome (source-1, source-9).

Questions

The result 7 questions

Who wins, and how likely?

  • Claude

    Sumudaerji: 64% (about 64 in 100 fights). Perez: 36%.

Other models' answers

  • Gemini

    Sumudaerji: 55% (about 55 in 100 fights). Perez: 45%.

What is the single most likely outcome?

  • Claude

    Sumudaerji by decision: 33% (about 33 in 100).

Other models' answers

  • ChatGPT

    Sumudaerji by decision: 42% (about 42 in 100).

  • Gemini

    A tie: Sumudaerji by KO/TKO and Sumudaerji by decision, 25% each (about 25 in 100 each).

What are the three most likely outcomes, and why?

  • Claude

    1. Sumudaerji by decision (33%); 2. Sumudaerji by KO/TKO (28%); 3. Perez by submission (13%)

Other models' answers

  • ChatGPT

    1. Sumudaerji by decision (42%); 2. Sumudaerji by KO/TKO (18%); 3. Perez by submission (15%)

  • Gemini

    1. Sumudaerji by KO/TKO (25%); 1. Sumudaerji by decision (25%); 3. Perez by decision (20%)

How confident is the model?

  • Claude

    6 of 10, based on Sumudaerji's 64% win chance.

Other models' answers

  • Gemini

    6 of 10, based on Sumudaerji's 55% win chance.

Does it end early or go to the judges?

  • Claude

    Finish (KO/TKO or submission): 54%. Decision: 46%.

Other models' answers

  • ChatGPT

    Finish (KO/TKO or submission): 47%. Decision: 53%.

  • Gemini

    Finish (KO/TKO or submission): 55%. Decision: 45%.

How does Perez win?

  • Claude

    Perez wins 36%: KO/TKO 10%, submission 13%, decision 13%.

Other models' answers

  • ChatGPT

    Perez wins 36%: KO/TKO 10%, submission 15%, decision 11%.

  • Gemini

    Perez wins 45%: KO/TKO 10%, submission 15%, decision 20%.

How does Sumudaerji win?

  • Claude

    Sumudaerji wins 64%: KO/TKO 28%, submission 3%, decision 33%.

Other models' answers

  • ChatGPT

    Sumudaerji wins 64%: KO/TKO 18%, submission 4%, decision 42%.

  • Gemini

    Sumudaerji wins 55%: KO/TKO 25%, submission 5%, decision 25%.

Compared with others 1 question

Do the models agree?

  • Claude

    The models agree: all 3 pick Sumudaerji.

After the fight 2 questions

What happened vs what the model said?

  • Claude

    Result: Sumudaerji by decision, the model's most likely outcome.

Other models' answers

  • Gemini

    Result: Sumudaerji by decision, the model's joint number 1 outcome. Its most likely outcomes were Sumudaerji by KO/TKO and Sumudaerji by decision (25% each).

How much chance did it give the actual result?

  • Claude

    33% (about 33 in 100).

Other models' answers

  • ChatGPT

    42% (about 42 in 100).

  • Gemini

    25% (about 25 in 100).

More questions (18)

The result 6 questions

KO/TKO vs submission vs decision overall

  • Claude

    KO/TKO 38%, submission 16%, decision 46%.

Other models' answers

  • ChatGPT

    KO/TKO 28%, submission 19%, decision 53%.

  • Gemini

    KO/TKO 35%, submission 20%, decision 45%.

Who is more likely to finish the other?

  • Claude

    Sumudaerji: 31%, against 23% for Perez.

Other models' answers

  • ChatGPT

    Perez: 25%, against 22% for Sumudaerji.

  • Gemini

    Sumudaerji: 30%, against 25% for Perez.

Can Perez get the finish? Can Sumudaerji?

  • Claude

    Perez: 23% (KO/TKO 10%, submission 13%). Sumudaerji: 31% (KO/TKO 28%, submission 3%).

Other models' answers

  • ChatGPT

    Perez: 25% (KO/TKO 10%, submission 15%). Sumudaerji: 22% (KO/TKO 18%, submission 4%).

  • Gemini

    Perez: 25% (KO/TKO 10%, submission 15%). Sumudaerji: 30% (KO/TKO 25%, submission 5%).

Who is more likely to be knocked out, and who to be submitted?

  • Claude

    Perez is knocked out in 28% and submitted in 3%. Sumudaerji is knocked out in 10% and submitted in 13%.

Other models' answers

  • ChatGPT

    Perez is knocked out in 18% and submitted in 4%. Sumudaerji is knocked out in 10% and submitted in 15%.

  • Gemini

    Perez is knocked out in 25% and submitted in 5%. Sumudaerji is knocked out in 10% and submitted in 15%.

How lopsided is this fight?

  • Claude

    Lopsided: 64% vs 36%.

Other models' answers

  • Gemini

    A clear favourite: 55% vs 45%.

How big an upset would it be if the underdog wins?

  • Claude

    Perez wins 36% (about 1 in 3 fights): a big upset.

Other models' answers

  • Gemini

    Perez wins 45% (about 1 in 2 fights): a modest upset.

Timing 1 question

Chance it goes to a decision

  • Claude

    46%.

Other models' answers

  • ChatGPT

    53%.

  • Gemini

    45%.

What if 1 question

If it goes to the judges, who wins?

  • Claude

    If it goes to the judges: Perez 28%, Sumudaerji 72%.

Other models' answers

  • ChatGPT

    If it goes to the judges: Perez 21%, Sumudaerji 79%.

  • Gemini

    If it goes to the judges: Perez 44%, Sumudaerji 56%.

Compared with others 4 questions

How does it compare with analyst picks?

  • Claude

    Of 12 analyst picks, 1 backs Perez and 11 back Sumudaerji. This model picks Sumudaerji (64%).

Other models' answers

  • Gemini

    Of 12 analyst picks, 1 backs Perez and 11 back Sumudaerji. This model picks Sumudaerji (55%).

Which model is most confident on Perez, and which on Sumudaerji?

  • Claude

    On Perez: Gemini (45%). On Sumudaerji: ChatGPT and Claude, tied (64%).

Where do the models disagree most?

  • Claude

    Sumudaerji by decision: from 25% (Gemini) to 42% (ChatGPT).

Which analyst agrees most with this model?

  • Claude

    Bedtime MMA (Sumudaerji by decision), HopperMMA (Sumudaerji by decision), DFS By The Numbers (Sumudaerji by decision), JakeGerberMMA (Sumudaerji by decision), Andrew Richardson (Sumudaerji by decision) and FightIQ (Sumudaerji by decision): the same winner and method as this model, tied; this model favours Sumudaerji, most likely by decision.

Other models' answers

  • Gemini

    TheWeasle (Sumudaerji by KO/TKO in round 3), AJ Devito (Sumudaerji), MMA Prediction Guru (MMAPG) (Sumudaerji), Bedtime MMA (Sumudaerji by decision), HopperMMA (Sumudaerji by decision), DFS By The Numbers (Sumudaerji by decision), We Want Picks (Sumudaerji), JakeGerberMMA (Sumudaerji by decision), Half The Battle (Sumudaerji by KO/TKO), Andrew Richardson (Sumudaerji by decision) and FightIQ (Sumudaerji by decision): the same winner as this model, tied; this model favours Sumudaerji.

How sure 4 questions

When was this forecast made, and by which model version?

  • Claude

    On 24 Aug 2026, by claude-opus-5.

Other models' answers

  • ChatGPT

    On 23 Aug 2026, by gpt-5.6-sol.

  • Gemini

    On 24 Aug 2026, by gemini-3.1-pro-high.

What can't the model know?

  • FightCard

    Models research public information before the fight, such as records, styles, recent results and coverage. They cannot see inside training camps, injuries nobody has reported, or what happens on the night. A small chance is not no chance: outcomes the model rates at a few percent still happen, and big favourites lose sometimes.

How is "confidence" worked out?

  • FightCard

    Confidence restates the favourite's win chance on a 10-step scale: the win chance in tens, rounded, so 55% is 6 of 10 and 95% or more is 10 of 10. 1 to 3 reads Low, 4 to 6 Medium and 7 to 10 High. A favourite usually has at least half the chance, so most picks show 5 or more. When the model has the fighters exactly even, there is no pick and no confidence level. It is not a separate judgement by the model.

What counts as a KO, a decision or a draw?

  • FightCard

    KO/TKO includes referee, doctor and corner stoppages, and disqualifications. A decision includes a fight stopped early and scored on the judges' cards. A stoppage between rounds counts as a finish in the round that just ended. A draw and a no contest share one outcome.

After the fight 2 questions

Which model was closest?

  • Claude

    ChatGPT, which gave the actual result 42%.

Was the actual result in this model's top 3?

  • Claude

    Yes.

claude-opus-5 · 24 Aug61 of 101 winner picks correct · 60% hit rate

Gemini

Pick

Sumudaerji

Confidence 6 of 10 · Medium

Most likely: a tie between Sumudaerji by KO/TKO and Sumudaerji by decision, about 25 in 100 each

The model's predicted finish round: 2.

Win chance: Perez 45%, Sumudaerji 55%.

Most likely outcomes

  1. Sumudaerji by KO/TKO, tied for number 125%
  2. Sumudaerji by decision, tied for number 125%
  3. Perez by decision20%
How it ends
  1. KO/TKO: Perez 10%, Sumudaerji 25%.
  2. Submission: Perez 15%, Sumudaerji 5%.
  3. Decision: Perez 20%, Sumudaerji 25%.

AI model's own reasoning — not checked by FightCard

How Gemini sees the fight

Sumudaerji possesses a significant 7-inch reach and 2-inch height advantage that will complicate distance management for Perez. Although Perez has the wrestling edge and a high striking volume, bridging the distance against a long, southpaw Sanda striker will be challenging. Sumudaerji's youth and outside precision striking give him a clear path to victory, provided he can avoid Perez's takedown attempts.

Keys to victory

  • Sumudaerji's massive 7-inch reach and 2-inch height advantage.
  • Perez's strong grappling base and top control advantage on the mat.
  • Sumudaerji's southpaw stance and outside distance management.

X-factor

Perez's ability to safely bridge the distance and enter the pocket without absorbing significant damage from Sumudaerji's long-range attacks.

Questions

The result 7 questions

Who wins, and how likely?

  • Gemini

    Sumudaerji: 55% (about 55 in 100 fights). Perez: 45%.

Other models' answers

  • ChatGPT

    Sumudaerji: 64% (about 64 in 100 fights). Perez: 36%.

  • Claude

    Sumudaerji: 64% (about 64 in 100 fights). Perez: 36%.

What is the single most likely outcome?

  • Gemini

    A tie: Sumudaerji by KO/TKO and Sumudaerji by decision, 25% each (about 25 in 100 each).

Other models' answers

  • ChatGPT

    Sumudaerji by decision: 42% (about 42 in 100).

  • Claude

    Sumudaerji by decision: 33% (about 33 in 100).

What are the three most likely outcomes, and why?

  • Gemini

    1. Sumudaerji by KO/TKO (25%); 1. Sumudaerji by decision (25%); 3. Perez by decision (20%)

Other models' answers

  • ChatGPT

    1. Sumudaerji by decision (42%); 2. Sumudaerji by KO/TKO (18%); 3. Perez by submission (15%)

  • Claude

    1. Sumudaerji by decision (33%); 2. Sumudaerji by KO/TKO (28%); 3. Perez by submission (13%)

How confident is the model?

  • Gemini

    6 of 10, based on Sumudaerji's 55% win chance.

Other models' answers

  • ChatGPT

    6 of 10, based on Sumudaerji's 64% win chance.

  • Claude

    6 of 10, based on Sumudaerji's 64% win chance.

Does it end early or go to the judges?

  • Gemini

    Finish (KO/TKO or submission): 55%. Decision: 45%.

Other models' answers

  • ChatGPT

    Finish (KO/TKO or submission): 47%. Decision: 53%.

  • Claude

    Finish (KO/TKO or submission): 54%. Decision: 46%.

How does Perez win?

  • Gemini

    Perez wins 45%: KO/TKO 10%, submission 15%, decision 20%.

Other models' answers

  • ChatGPT

    Perez wins 36%: KO/TKO 10%, submission 15%, decision 11%.

  • Claude

    Perez wins 36%: KO/TKO 10%, submission 13%, decision 13%.

How does Sumudaerji win?

  • Gemini

    Sumudaerji wins 55%: KO/TKO 25%, submission 5%, decision 25%.

Other models' answers

  • ChatGPT

    Sumudaerji wins 64%: KO/TKO 18%, submission 4%, decision 42%.

  • Claude

    Sumudaerji wins 64%: KO/TKO 28%, submission 3%, decision 33%.

What if 1 question

What is the upset path?

  • Gemini

    Perez's most likely upset: Perez by decision (20%).

Compared with others 1 question

Do the models agree?

  • Gemini

    The models agree: all 3 pick Sumudaerji.

After the fight 2 questions

What happened vs what the model said?

  • Gemini

    Result: Sumudaerji by decision, the model's joint number 1 outcome. Its most likely outcomes were Sumudaerji by KO/TKO and Sumudaerji by decision (25% each).

Other models' answers

  • ChatGPT

    Result: Sumudaerji by decision, the model's most likely outcome.

  • Claude

    Result: Sumudaerji by decision, the model's most likely outcome.

How much chance did it give the actual result?

  • Gemini

    25% (about 25 in 100).

Other models' answers

  • ChatGPT

    42% (about 42 in 100).

  • Claude

    33% (about 33 in 100).

More questions (19)

The result 6 questions

KO/TKO vs submission vs decision overall

  • Gemini

    KO/TKO 35%, submission 20%, decision 45%.

Other models' answers

  • ChatGPT

    KO/TKO 28%, submission 19%, decision 53%.

  • Claude

    KO/TKO 38%, submission 16%, decision 46%.

Who is more likely to finish the other?

  • Gemini

    Sumudaerji: 30%, against 25% for Perez.

Other models' answers

  • ChatGPT

    Perez: 25%, against 22% for Sumudaerji.

  • Claude

    Sumudaerji: 31%, against 23% for Perez.

Can Perez get the finish? Can Sumudaerji?

  • Gemini

    Perez: 25% (KO/TKO 10%, submission 15%). Sumudaerji: 30% (KO/TKO 25%, submission 5%).

Other models' answers

  • ChatGPT

    Perez: 25% (KO/TKO 10%, submission 15%). Sumudaerji: 22% (KO/TKO 18%, submission 4%).

  • Claude

    Perez: 23% (KO/TKO 10%, submission 13%). Sumudaerji: 31% (KO/TKO 28%, submission 3%).

Who is more likely to be knocked out, and who to be submitted?

  • Gemini

    Perez is knocked out in 25% and submitted in 5%. Sumudaerji is knocked out in 10% and submitted in 15%.

Other models' answers

  • ChatGPT

    Perez is knocked out in 18% and submitted in 4%. Sumudaerji is knocked out in 10% and submitted in 15%.

  • Claude

    Perez is knocked out in 28% and submitted in 3%. Sumudaerji is knocked out in 10% and submitted in 13%.

How lopsided is this fight?

  • Gemini

    A clear favourite: 55% vs 45%.

Other models' answers

  • ChatGPT

    Lopsided: 64% vs 36%.

  • Claude

    Lopsided: 64% vs 36%.

How big an upset would it be if the underdog wins?

  • Gemini

    Perez wins 45% (about 1 in 2 fights): a modest upset.

Other models' answers

  • ChatGPT

    Perez wins 36% (about 1 in 3 fights): a big upset.

  • Claude

    Perez wins 36% (about 1 in 3 fights): a big upset.

Timing 1 question

Chance it goes to a decision

  • Gemini

    45%.

Other models' answers

  • ChatGPT

    53%.

  • Claude

    46%.

What if 1 question

If it goes to the judges, who wins?

  • Gemini

    If it goes to the judges: Perez 44%, Sumudaerji 56%.

Other models' answers

  • ChatGPT

    If it goes to the judges: Perez 21%, Sumudaerji 79%.

  • Claude

    If it goes to the judges: Perez 28%, Sumudaerji 72%.

Compared with others 5 questions

Where does this model differ from the others?

  • Gemini

    Biggest difference: Sumudaerji by decision, 25% here against 38% on average for the others.

Other models' answers

  • ChatGPT

    Biggest difference: Sumudaerji by decision, 42% here against 29% on average for the others.

How does it compare with analyst picks?

  • Gemini

    Of 12 analyst picks, 1 backs Perez and 11 back Sumudaerji. This model picks Sumudaerji (55%).

Other models' answers

  • ChatGPT

    Of 12 analyst picks, 1 backs Perez and 11 back Sumudaerji. This model picks Sumudaerji (64%).

  • Claude

    Of 12 analyst picks, 1 backs Perez and 11 back Sumudaerji. This model picks Sumudaerji (64%).

Which model is most confident on Perez, and which on Sumudaerji?

  • Gemini

    On Perez: Gemini (45%). On Sumudaerji: ChatGPT and Claude, tied (64%).

Where do the models disagree most?

  • Gemini

    Sumudaerji by decision: from 25% (Gemini) to 42% (ChatGPT).

Which analyst agrees most with this model?

  • Gemini

    TheWeasle (Sumudaerji by KO/TKO in round 3), AJ Devito (Sumudaerji), MMA Prediction Guru (MMAPG) (Sumudaerji), Bedtime MMA (Sumudaerji by decision), HopperMMA (Sumudaerji by decision), DFS By The Numbers (Sumudaerji by decision), We Want Picks (Sumudaerji), JakeGerberMMA (Sumudaerji by decision), Half The Battle (Sumudaerji by KO/TKO), Andrew Richardson (Sumudaerji by decision) and FightIQ (Sumudaerji by decision): the same winner as this model, tied; this model favours Sumudaerji.

Other models' answers

  • ChatGPT

    Bedtime MMA (Sumudaerji by decision), HopperMMA (Sumudaerji by decision), DFS By The Numbers (Sumudaerji by decision), JakeGerberMMA (Sumudaerji by decision), Andrew Richardson (Sumudaerji by decision) and FightIQ (Sumudaerji by decision): the same winner and method as this model, tied; this model favours Sumudaerji, most likely by decision.

  • Claude

    Bedtime MMA (Sumudaerji by decision), HopperMMA (Sumudaerji by decision), DFS By The Numbers (Sumudaerji by decision), JakeGerberMMA (Sumudaerji by decision), Andrew Richardson (Sumudaerji by decision) and FightIQ (Sumudaerji by decision): the same winner and method as this model, tied; this model favours Sumudaerji, most likely by decision.

How sure 4 questions

When was this forecast made, and by which model version?

  • Gemini

    On 24 Aug 2026, by gemini-3.1-pro-high.

Other models' answers

  • ChatGPT

    On 23 Aug 2026, by gpt-5.6-sol.

  • Claude

    On 24 Aug 2026, by claude-opus-5.

What can't the model know?

  • FightCard

    Models research public information before the fight, such as records, styles, recent results and coverage. They cannot see inside training camps, injuries nobody has reported, or what happens on the night. A small chance is not no chance: outcomes the model rates at a few percent still happen, and big favourites lose sometimes.

How is "confidence" worked out?

  • FightCard

    Confidence restates the favourite's win chance on a 10-step scale: the win chance in tens, rounded, so 55% is 6 of 10 and 95% or more is 10 of 10. 1 to 3 reads Low, 4 to 6 Medium and 7 to 10 High. A favourite usually has at least half the chance, so most picks show 5 or more. When the model has the fighters exactly even, there is no pick and no confidence level. It is not a separate judgement by the model.

What counts as a KO, a decision or a draw?

  • FightCard

    KO/TKO includes referee, doctor and corner stoppages, and disqualifications. A decision includes a fight stopped early and scored on the judges' cards. A stoppage between rounds counts as a finish in the round that just ended. A draw and a no contest share one outcome.

After the fight 2 questions

Which model was closest?

  • Gemini

    ChatGPT, which gave the actual result 42%.

Was the actual result in this model's top 3?

  • Gemini

    Yes.

gemini-3.1-pro-high · 24 Aug68 of 108 winner picks correct · 63% hit rate

AI-generated · Not betting advice · 21+. Gambling problem? Call 1-800-MY-RESET

Market Odds

No current consensus line is saved on this page.

Not betting advice · 21+. Gambling problem? Call 1-800-MY-RESET

Independent fan-made fight recap. Not affiliated with, endorsed by, or sponsored by the UFC or any other MMA promotion. Analysis is editorial opinion, not betting advice. Fighter names and records are used for commentary. Image credits · Methodology · Privacy · Terms.