Model pick: Karine Silva (51.6%). The model expects the other fighter to win about 48 times in 100 similar fights.
Evidence and reliability
Karine Silva: 8 prior detailed UFC fights
Gabriella Fernandes: 6 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 pickKarine Silva51.6%
Karine SilvaGabriella Fernandes
+Supporting the pick3
Striking prevention against prior oppositionEdge: Karine Silva
Karine Silva 3.27, Gabriella Fernandes -3.68.
3.27-3.68
Striking prevention compared with opponents’ usual output.
Significant strikes absorbed per roundEdge: Karine Silva
Karine Silva 12.72, Gabriella Fernandes 25.47.
12.7225.47
Prior UFC win rateEdge: Karine Silva
Karine Silva 62.5%, Gabriella Fernandes 50.0%.
62.5%50.0%
−Working against the pick3
Takedown defenseEdge: Gabriella Fernandes
Karine Silva 29.4%, Gabriella Fernandes 69.0%.
29.4%69.0%
Striking accuracyEdge: Gabriella Fernandes
Karine Silva 39.7%, Gabriella Fernandes 46.6%.
39.7%46.6%
Significant strikes landed per roundEdge: Gabriella Fernandes
Karine Silva 10.94, Gabriella Fernandes 16.87.
10.9416.87
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 distance59%
KO/TKO17%
Submission24%
Decision59%
Outcome breakdown (who wins × how)
Outcome
Karine Silva
Gabriella Fernandes
KO/TKO
5%
14%
Submission
16%
6%
Decision
30%
28%
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.
Research components calculated 4 Nov 2026, 07:59 UTC. Separate from the locked Winner forecast; no official grading credit.
Stat comparison
Descriptive
Reason through the matchup yourself. The brighter side of each bar marks the fighter with the edge.
Karine SilvaGabriella Fernandes
Biggest statistical gaps
Karine Silva has more takedown production: 0.67 takedowns landed per round to Gabriella Fernandes’s 0.13.
Karine Silva controls more: 1:06 per round to Gabriella Fernandes’s 0:13.
Karine Silva completes more of their takedown attempts: 54.5% to Gabriella Fernandes’s 18.2%.
Career UFC totals from fights before Sat 7 Nov 2026 only (pre-fight cutoff; this bout and anything after it are excluded); latest included fight Sat 18 Apr 2026. These are descriptive numbers, not the model's locked inputs, and they are not opponent-adjusted. Small samples move a lot.
The fighters
Karine Silva
UFC record 5-3 · 8 detailed UFC fights · Submission Hunter · Wrestling Heavy
Ratings
Ratings V2Rated
Striking (offense)64
Power70
Striking defense76
Takedown defense60
WrestlingProvisional measure83
Ratings V2 descriptive 0–100 scores (8 detailed fights; as of Sat 18 Apr 2026). Not win probabilities or rankings. Missing values stay unavailable.
Chin and Cardio · Frozen through October 17, 2026
R1 research scale
ChinEvidence: DEVELOPING83.8
CardioEvidence: DEVELOPING66.7
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 V2 descriptive 0–100 scores (6 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: DEVELOPING81.1
CardioEvidence: DEVELOPING68.3
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.
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.
