UFC 316: Dvalishvili vs O'Malley 2

1️⃣ Quillan Salkilld vs Yanal Ashmouz

• Our pick: Ashmouz KO/TKO (lean)

• Reality: Salkilld decision

Why it failed:

✅ Underestimated Salkilld’s composure vs chaos

✅ Salkilld handled Ashmouz’s blitz easily → stuck to jab and movement

✅ Ashmouz cardio collapsed completely after R1

✅ Model overweighted early chaos threat, undervalued Salkilld’s discipline

→ Update:

• When predicting for high cardio, composed strikers vs one-round chaos brawlers → lean toward cardio & discipline if fighter has proven to maintain gameplan

• Cardio collapse on tape for Ashmouz was a bigger red flag than weighted

2️⃣ Jeka Saragih vs Yoo Jae-sang

• Our pick: Jeka KO/TKO

• Reality: Yoo KO R1 (28 sec)

Why it failed:

✅ Massive underestimation of Yoo’s sharp counters and fast start

✅ Expected Yoo to play long game → instead countered Jeka’s wild entry perfectly

✅ Model over-trusted Jeka’s chin & brawling → wasn’t ready for sharp technical counters

→ Update:

• When opponent has clear clean counter tools and Jeka’s style is wild entry based → increase volatility rating

• Do not trust Jeka’s chin vs clean, technical counter-strikers → flag this matchup in future

3️⃣ Ariane Lipski vs Wang Cong

• Our pick: Lipski decision

• Reality: Wang Cong decision

Why it failed:

✅ Over-trusted Lipski’s UFC experience

✅ Wang out-volumed, out-paced, and Lipski fought tentative

✅ Model underweighted debuting striker’s output & confidence

→ Update:

• For active, high-volume debutants with clean striking vs historically hesitant veterans → experience factor should not override current form & tempo edge

• Lipski is showing signs of timidity under volume pressure → downgrade this in future

4️⃣ Khaos Williams vs Andreas Gustafsson

• Our pick: Khaos KO/TKO

• Reality: Gustafsson decision

Why it failed:

✅ Underestimated Gustafsson’s relentless pressure & cardio

✅ Expected him to be caught early → but excellent cage pressure, body lock, chained grappling

✅ Khaos low output froze under wrestling threat

→ Update:

• For Khaos vs pressure wrestlers → strongly downgrade if opponent has proven cardio and chain wrestling

• Khaos freezes under consistent clinch work → this must be flagged in future

5️⃣ Sergey Spivak vs Waldo Cortes-Acosta

• Our pick: Spivak TKO/sub

• Reality: Waldo decision

Why it failed:

✅ Spivak failed to implement wrestling gameplan — fought like striker

✅ Model assumed forced grappling — didn’t account for Spivak stalling

✅ Waldo’s improved TDD and anti-grappling footwork were undervalued

→ Update:

• When picking grappling-based fighters → verify recent consistency in actually pursuing takedowns

• Do not overtrust Spivak vs mobile boxers if wrestling has recently declined in urgency

6️⃣ Vicente Luque vs Kevin Holland

• Our pick: Luque KO/TKO

• Reality: Holland sub R2

Why it failed:

✅ Did not expect Holland to proactively grapple

✅ Underweighted Luque’s grappling defense decline after recent damage

✅ Holland very smart gameplan — model assumed striker-vs-striker

→ Update:

• Do not lock in archetype expectation — Holland showed willingness to grapple

• Luque’s grappling defense must now be downgraded — decline in reflexes post-damage

7️⃣ Mario Bautista vs Patchy Mix

• Our pick: Mix by submission

• Reality: Bautista decision

Why it failed:

✅ Overtrusted Mix’s ability to force grappling

✅ Bautista excellent scrambling and pace neutralized Mix

✅ Underweighted Bautista’s cardio & defensive scrambling gains

→ Update:

• Against high-paced scramblers with good cardio → Mix’s takedown forcing ability needs reevaluation

• Do not assume Mix can impose grappling vs upgraded anti-grapplers

8️⃣ Kelvin Gastelum vs Joe Pyfer

• Our pick: Gastelum decision

• Reality: Pyfer decision

Why it failed:

✅ Underestimated Gastelum’s decline in footspeed & reaction

✅ Pyfer won minute-to-minute with activity

✅ Gastelum’s recent damage accumulation underweighted

→ Update:

• Gastelum’s speed & defensive reaction now must be heavily downgraded — he is entering sharp decline phase

• Must adjust model to reflect aging “durable vet” trap