For entertainment only — MMA analysis only, not financial advice.

Pat Sabatini vs William Gomis, UFC 328: Chimaev vs. Strickland prediction and deep analysis

FIGHT COMPLETEResult: Pat SabatiniEngine pick: William Gomis ✗ MISSUpcoming predictions →
// UFC 328: Chimaev vs. Strickland

Pat Sabatini/William Gomis

featherweight
ENGINE FORECAST
// PROJECTED WINNER
William Gomis.
PROJECTED METHOD IF Gomis WINS · LOCKED
MOST LIKELY EXACT OUTCOMELOCKED
FIGHT GOES TO DECISIONLOCKEDEither fighter · total probability
CONFIDENCELOCKED
EVIDENCE COVERAGELOCKEDEligible assigned committee mass
PRIMARY EDGELOCKED
LARGEST RISKLOCKED
// WIN PROBABILITY
SabatiniGomis
45%55%
WAR ROOM COMMITTEE · UNIFIED CONSENSUS
Blueprint MMAblueprintmma.com
// 01 · Winner Committee

Winner Committee

Thirty-three specialists answer distinct winner-only questions. Eligible evidence is renormalized only inside its original lane.

Seven-lane committee detail is included with Sharp.
Unlock full analysis ->
// 02 · Fight-State Projection

Fight-State Projection

These are distributions of eligible committee signal mass—not invented percentages of fight time.

Fight-State Projection is included with Sharp.
Unlock full analysis ->
// 03 · Conditional Method

Conditional Method Model

Each row is conditional on that fighter already winning; every fighter row sums to 100%.

Conditional Method Model is included with Sharp.
Unlock full analysis ->
// 04 · Conditional Timing

Conditional Round Model

Round-event mass is derived from survival-conditioned hazards. The final scheduled round is absorbing.

Conditional Round Model is included with Sharp.
Unlock full analysis ->
// 05 · Probability Reconciliation

Probability Reconciliation

The exact-outcome grid preserves both winner marginals and sums to 100%. Decision branches have no round.

Probability Reconciliation is included with Sharp.
Unlock full analysis ->
// 06 · Output Selection

Output Selection

The winner and every outcome label come from one immutable reconciled distribution. Oversight may widen uncertainty, never move the point forecast.

Output Selection is included with Sharp.
Unlock full analysis ->

Key Takeaways

  • The War Room Committee picks William Gomis (55% win probability).
  • William Gomis path. "Counter Volume To The Cards": Manage range with the rear straight in open stance, deny clean level changes
  • Pat Sabatini path. "Submission Off The Chain": Establish takedown threat early — body lock entries off Gomis's southpaw straight
  • Risk: Gomis's takedown defense is undertested
  • Risk: Sabatini chin against a sharp counter southpaw

Frequently Asked Questions

Who won Pat Sabatini vs William Gomis?

Pat Sabatini won. Blueprint MMA's pre-fight pick was William Gomis.

Who wins Pat Sabatini vs William Gomis?

Blueprint MMA's War Room Committee picks William Gomis with a 55% win probability.

What could change the outcome of Pat Sabatini vs William Gomis?

Gomis's takedown defense is undertested: The TDD column reads as unavailable in the stats block, and Sabatini's takedown-volume edge plus submission rate is the exact attack profile that exposes a striker who hasn't had to defend it before. Sabatini chin against a sharp counter southpaw: His only clean KO loss came against Diego Lopes when he tried to close distance recklessly — Gomis's left hand in open stance is structurally the same problem if Sabatini's entries get sloppy. Late-round volume swing: Gomis's pace profile shows round-three output peaking sharply, which means a close fight on the cards bends his way the longer it stays standing.

How does Blueprint MMA make this prediction?

Blueprint MMA validates and freezes the fight data, then runs 68 Kimi K2.6 agent executions: 33 Winner, 10 Method, 20 Timing, and 5 non-voting Oversight. Kimi K2.6 is the only inference model; LangGraph only orchestrates stage order and state. Winner agents fix the point probability first, Method and Timing agents fill conditional branches without reopening it, and Oversight audits without voting. The public probability is held to a sane 5-95% range and logged before release. Published winner probabilities are AI-only. Market prices may be shown for benchmarking, CLV, and workflow context, but they do not select, rank, or alter the published forecast. If the direct pipeline cannot return a valid read, the forecast is held. Methodology is documented at https://blueprintmma.com/manifesto and accuracy is tracked publicly at https://blueprintmma.com/track-record.