UFC on ESPN: Blanchfield vs. Barber

1️⃣ Rayanne dos Santos vs Alice Ardelean

• Prediction: Rayanne dos Santos by Decision (or late sub if Ardelean fades).

• Reality: Alice Ardelean by Decision.

What went wrong:

• I overvalued Santos’s pace and presumed she would overwhelm Ardelean with volume and grappling edge.

• I underestimated Ardelean’s physicality, clinch control, and ability to neutralize Santos’s offense.

• Did not fully account for judging bias toward top control and cage pushing — Ardelean played a smart clinch-heavy game.

New lessons:

✅ Fighters with strong cage control can steal decisions even when out-struck in open space — weight this more heavily.

✅ Downgrade fighters like Santos who struggle to impose pace against physically stronger opponents.

✅ Adjust model to better reward clinch-heavy grinding styles in close matchups.

2️⃣ Kurt Holobaugh vs Jordan Leavitt

• Prediction: Holobaugh by TKO (Round 2) or Decision.

• Reality: Leavitt by Submission.

What went wrong:

• I underestimated Leavitt’s ability to aggressively pursue submissions early.

• Overlooked Holobaugh’s defensive grappling liabilities—he was subbed quickly.

• Assumed Holobaugh’s pace would wear Leavitt out first, but Leavitt implemented a fast-start sub attack.

New lessons:

✅ Heavily penalize aging fighters with known grappling defense holes when facing submission hunters.

✅ Be cautious predicting pace-based wins when opponent is a dangerous opportunistic grappler with fast-start potential.

✅ Increase volatility rating when such style clash exists.

3️⃣ Zachary Reese vs Duško Todorović

• Prediction: Duško by TKO (Round 2) – high risk noted.

• Reality: Zachary Reese won.

What went wrong:

• The model properly flagged this as a high-risk volatile fight.

• Todorović’s defensive flaws and chin issues are increasingly unreliable.

• Reese’s size, power, and early aggression were underweighted vs Todorović’s declining durability.

New lessons:

✅ Fighters like Todorović showing clear durability and defensive decline should be downgraded even more heavily in volatile matchups.

✅ Power hitters with size advantages vs declining veterans get an extra volatility bump.

✅ The model needs to better penalize chin decline + poor defensive reactions in matchups with early KO threat.

4️⃣ Billy Ray Goff vs Ramiz Brahimaj

• Prediction: Goff by KO/TKO (Round 2).

• Reality: Brahimaj by Decision.

What went wrong:

• I overvalued Goff’s KO power and assumed Brahimaj’s chin would fold under sustained damage.

• Severely undervalued Brahimaj’s durability improvements and grappling control.

• Did not sufficiently account for Goff’s defensive wrestling gaps or tendency to get stuck on bottom.

• Brahimaj’s smarter, measured gameplan wasn’t fully reflected in the pre-fight read.

New lessons:

✅ Do not over-assume KO finishing paths vs grapplers who can control tempo.

✅ When a fighter has shown improved durability and fight IQ in recent fights (as Brahimaj had), adjust accordingly.

✅ Reward grappling-first fighters who can control rounds and reduce KO variance, especially against wild brawlers.

Summary of global improvements learned:

✅ Increase weighting for clinch control and grinding styles in judging-heavy matchups (Ardelean case).

✅ Penalize aging fighters with grappling defense holes more aggressively when facing sub hunters (Holobaugh case).

✅ Heavily downgrade durability-declining veterans against early KO threats (Duško case).

✅ Reduce KO path reliability when facing improved grapplers with fight IQ and durability upgrades (Brahimaj case).

✅ Add new volatility flag:

• Aging fighters + defensive gaps vs explosive starters = volatility spike.

• Grappler vs wild brawler = volatility if brawler relies only on KO.