Model pick: JeongYeong Lee (62.1%). The model expects the other fighter to win about 38 times in 100 similar fights.
Evidence and reliability
Julian Erosa: 18 prior detailed UFC fights
JeongYeong Lee: 4 prior detailed UFC fights
Official forecasts require at least 3 prior detailed UFC fights for both fighters. The estimate is a probability from one model, not a guarantee; small differences between fighters carry real uncertainty.
Why the model leans this way
Locked before the fight
Model pickJeongYeong Lee62.1%
Julian ErosaJeongYeong Lee
+Supporting the pick3
AgeEdge: JeongYeong Lee
Julian Erosa 37.3 yrs, JeongYeong Lee 31.0 yrs.
37.3 yrs31.0 yrs
Striking prevention against prior oppositionEdge: JeongYeong Lee
Julian Erosa -1.53, JeongYeong Lee 2.72.
-1.532.72
Striking prevention compared with opponents’ usual output.
Prior UFC lossesEdge: JeongYeong Lee
Julian Erosa 9, JeongYeong Lee 2.
92
−Working against the pick3
Striking output against prior oppositionEdge: Julian Erosa
Julian Erosa 8.31, JeongYeong Lee -6.14.
8.31-6.14
Striking output compared with opponents’ prior defense.
Significant strikes landed per roundEdge: Julian Erosa
Julian Erosa 24.43, JeongYeong Lee 14.60.
24.4314.60
Takedowns landed per roundEdge: Julian Erosa
Julian Erosa 0.49, JeongYeong Lee 0.20.
0.490.20
The model's own locked explanation fields, captured before the fight: what moved the model, not causes. "Against prior opposition" measures compare a fighter with what their opponents usually allow. Opponent-adjusted career evidence; no injury, camp, fight-date strength index or direct 730-day form input. Not prospectively proven.
Method and distance
Separate component · small live sample
Goes the distance58%
KO/TKO29%
Submission13%
Decision58%
Outcome breakdown (who wins × how)
Outcome
Julian Erosa
JeongYeong Lee
KO/TKO
7%
22%
Submission
9%
4%
Decision
22%
36%
Separate method and distance components with a small live sample. They do not change the winner estimate. The outcome breakdown is an approximation built from the winner estimate and the fight-level finish estimate.
Stat comparison
Descriptive
Reason through the matchup yourself. The brighter side of each bar marks the fighter with the edge.
Julian ErosaJeongYeong Lee
Biggest statistical gaps
Julian Erosa brings more wrestling volume: 1.09 takedown attempts per round to JeongYeong Lee’s 0.30.
Julian Erosa has more takedown production: 0.49 takedowns landed per round to JeongYeong Lee’s 0.20.
Julian Erosa scores more knockdowns: 0.03 per round to JeongYeong Lee’s 0.00.
Striking
Significant strikes landedper round
Julian Erosa 24.4, JeongYeong Lee 14.6. Edge: Julian Erosa.
24.414.6
Significant strikes absorbedper round · lower is better
Julian Erosa 25.1, JeongYeong Lee 17.5. Edge: JeongYeong Lee.
25.117.5
Striking differentiallanded minus absorbed, per round
Julian Erosa -0.7, JeongYeong Lee -2.9. Edge: Julian Erosa.
-0.7-2.9
Striking accuracyshare of significant strikes landed
Julian Erosa 47.9%, JeongYeong Lee 48.8%. Edge: JeongYeong Lee.
47.9%48.8%
Striking defenseshare of opponents’ significant strikes avoided
Julian Erosa 49.2%, JeongYeong Lee 51.2%. Edge: JeongYeong Lee.
49.2%51.2%
Knockdownsper round
Julian Erosa 0.03, JeongYeong Lee 0.00. Edge: Julian Erosa.
0.030.00
Wrestling and grappling
Takedown attemptsper round · wrestling volume
Julian Erosa 1.09, JeongYeong Lee 0.30. Edge: Julian Erosa.
1.090.30
Takedowns landedper round · takedown production
Julian Erosa 0.49, JeongYeong Lee 0.20. Edge: Julian Erosa.
0.490.20
Takedown accuracyshare of own attempts completed
Julian Erosa 44.7%, JeongYeong Lee 66.7%. Edge: JeongYeong Lee.
44.7%66.7%
Takedown defenseshare of opponents’ attempts stopped
Julian Erosa 61.5%, JeongYeong Lee 73.8%. Edge: JeongYeong Lee.
61.5%73.8%
Control timeminutes:seconds per round
Julian Erosa 0:37, JeongYeong Lee 0:32. Edge: Julian Erosa.
0:370:32
Submission attemptsper round
Julian Erosa 0.20, JeongYeong Lee 0.00. Edge: Julian Erosa.
0.200.00
Fight profile
Finish rateshare of UFC wins by KO/TKO or submission
Julian Erosa 66.7%, JeongYeong Lee 0.0%. Edge: Julian Erosa.
66.7%0.0%
Wins by KO/TKOUFC wins
Julian Erosa 2, JeongYeong Lee 0. Style difference, no edge.
20
Wins by submissionUFC wins
Julian Erosa 4, JeongYeong Lee 0. Style difference, no edge.
40
Wins by decisionUFC wins
Julian Erosa 3, JeongYeong Lee 2. Style difference, no edge.
32
Average fight timeelapsed time per UFC bout (minutes:seconds)
Julian Erosa 8:05, JeongYeong Lee 10:22. Style difference, no edge.
8:0510:22
Detailed UFC fightsmore fights = more evidence behind these numbers
Julian Erosa 18, JeongYeong Lee 4. Edge: Julian Erosa.
184
Career UFC totals from fights before Sat 31 Oct 2026 only (pre-fight cutoff; this bout and anything after it are excluded); latest included fight Sat 28 Mar 2026. These are descriptive numbers, not the model's locked inputs, and they are not opponent-adjusted. Small samples move a lot.
Ratings V2 descriptive 0–100 scores (18 detailed fights; as of Sat 28 Mar 2026). Not win probabilities or rankings. Missing values stay unavailable.
Chin and Cardio · Frozen through October 17, 2026
R1 research scale
ChinEvidence: ESTABLISHED66.6
CardioEvidence: DEVELOPING80.5
Source: R1 public_candidate_display · status: Research candidate · as of Sat 17 Oct 2026. Separate frozen research scale; fights after October 17, 2026 are not reflected.
About these ratings
Ratings are domain-specific descriptive scales. Equal point differences across different domains do not imply equal absolute skill differences.
Striking (offense) · Ratings V2
Ratings V2 striking-offense score built from recorded significant strikes landed per round, significant-strike accuracy and significant-strike differential.
Power · Ratings V2
Recorded knockdown and KO/TKO-win threat under the Ratings V2 contract (knockdowns per round, KO/TKO win rate); not literal measured punching force.
Striking defense · Ratings V2
Recorded strike avoidance (striking defense) and bounded absorption (strikes absorbed per round) under the Ratings V2 contract.
Takedown defense · Ratings V2
Recorded defense against incoming takedown attempts (takedown defense rate) under the Ratings V2 contract.
Wrestling · Ratings V2
Ratings V2 wrestling-offense score: recorded takedowns landed per round and takedown accuracy. Not ground control or complete grappling.
Chin · R1 research scale
Exposure-adjusted recorded knockdown resistance. This is not a direct biological measurement of chin, recovery, or injury resistance.
Cardio · R1 research scale
Recorded later-round striking activity and effectiveness retention. This is not a direct physiological fitness measurement.
Ratings V2 scores follow each fighter's latest recorded state; R1 research-scale scores are frozen through 17 Oct 2026. Recovery is separate metadata, not a rating. No combined rating exists.
Ratings V2Provisional — fewer than 5 detailed UFC fights
Striking (offense)74
Power65
Striking defense69
Takedown defense80
WrestlingProvisional measure76
Ratings V2 descriptive 0–100 scores (4 detailed fights; as of Sat 10 May 2025). Not win probabilities or rankings. Missing values stay unavailable.
Chin and Cardio · Frozen through October 17, 2026
R1 research scale
ChinEvidence: LIMITED81.3
CardioEvidence: LIMITED65.9
Source: R1 public_candidate_display · status: Research candidate · as of Sat 17 Oct 2026. Separate frozen research scale; fights after October 17, 2026 are not reflected.
About these ratings
Ratings are domain-specific descriptive scales. Equal point differences across different domains do not imply equal absolute skill differences.
Striking (offense) · Ratings V2
Ratings V2 striking-offense score built from recorded significant strikes landed per round, significant-strike accuracy and significant-strike differential.
Power · Ratings V2
Recorded knockdown and KO/TKO-win threat under the Ratings V2 contract (knockdowns per round, KO/TKO win rate); not literal measured punching force.
Striking defense · Ratings V2
Recorded strike avoidance (striking defense) and bounded absorption (strikes absorbed per round) under the Ratings V2 contract.
Takedown defense · Ratings V2
Recorded defense against incoming takedown attempts (takedown defense rate) under the Ratings V2 contract.
Wrestling · Ratings V2
Ratings V2 wrestling-offense score: recorded takedowns landed per round and takedown accuracy. Not ground control or complete grappling.
Chin · R1 research scale
Exposure-adjusted recorded knockdown resistance. This is not a direct biological measurement of chin, recovery, or injury resistance.
Cardio · R1 research scale
Recorded later-round striking activity and effectiveness retention. This is not a direct physiological fitness measurement.
Ratings V2 scores follow each fighter's latest recorded state; R1 research-scale scores are frozen through 17 Oct 2026. Recovery is separate metadata, not a rating. No combined rating exists.
A written review of what happened versus the locked forecast has not been published for this bout yet.
Simulate in the Lab
The Lab's simulators currently run on a fixed library of historical example fights, so this exact matchup is not available there yet. They are research tools and can disagree with the official forecast.
