UFC 23: Ultimate Japan 2: Predictions & Analysis
UFC 23: Ultimate Japan 2 lands on Friday, November 19, 1999 in Chiba, Japan with 7 bouts on the card. The card is headlined by a championship fight. Below is our fight-by-fight breakdown, combining Elo ratings, rolling statistical trends, style matchup data, and betting market context into a pick for every bout.
Quick Picks
| Matchup | Pick | Confidence | Prob |
|---|---|---|---|
| Kevin Randleman vs Pete WilliamsHeavyweight | Kevin Randleman | Lean | 61% |
| Pedro Rizzo vs Tsuyoshi KohsakaHeavyweight | Pedro Rizzo | Toss-up | 54% |
| Kenichi Yamamoto vs Katsuhisa FujiiOpen Weight | Kenichi Yamamoto | Lean | 55% |
| Joe Slick vs Jason DeLuciaMiddleweight | Joe Slick | Lean | 56% |
| Eugene Jackson vs Keiichiro YamamiyaMiddleweight | Eugene Jackson | Confident | 69% |
| Kenichi Yamamoto vs Daiju TakaseOpen Weight | Kenichi Yamamoto | Strong | 83% |
| Katsuhisa Fujii vs Masutatsu YanoOpen Weight | Katsuhisa Fujii | Strong | 79% |
Fight-by-Fight Breakdown
Kevin Randleman vs Pete Williams
The Heavyweight championship matchup features Kevin Randleman (3-3) taking on Pete Williams (3-4).
Randleman is rated at 1189 — 336 points above Williams's 853. Gaps this large usually mean one fighter has been consistently beating better opponents.
Both fighters land in our "Wrestler" archetype — fighters who win by dictating where the fight takes place, grinding out control time and wearing opponents down. When mirror matchups like this happen, the edge usually goes to whoever can impose their preferred pace and range.
A few statistical edges stand out. Williams throws significantly more leather — a 0.7 sig. strike per minute gap. Randleman is far more active with takedowns, averaging 6.5 more per 15 minutes. Williams has tighter striking defense, making opponents miss more often.
The Pick: Kevin Randleman over Pete Williams. The model gives Randleman a slight nod at 61% — this could easily go either way.
Pedro Rizzo vs Tsuyoshi Kohsaka
The Heavyweight matchup features Pedro Rizzo (8-5) taking on Tsuyoshi Kohsaka (3-2).
Rizzo is rated at 1268 — 254 points above Kohsaka's 1014. Gaps this large usually mean one fighter has been consistently beating better opponents.
Stylistically this is Rizzo's striker game against Kohsaka's submission artist approach. Rizzo brings a versatile approach, while Kohsaka is most dangerous on the ground, constantly threatening chokes and joint locks. Historically these archetypes are dead-even when they collide.
A few statistical edges stand out. Kohsaka throws significantly more leather — a 0.4 sig. strike per minute gap. Kohsaka is far more active with takedowns, averaging 2.7 more per 15 minutes. Kohsaka has tighter striking defense, making opponents miss more often.
The Pick: Pedro Rizzo over Tsuyoshi Kohsaka. This is essentially a pick'em. The model nudges toward Rizzo at 54%, but there's almost nothing separating these two.
Kenichi Yamamoto vs Katsuhisa Fujii
The Open Weight matchup features Kenichi Yamamoto (2-0) taking on Katsuhisa Fujii (0-0).
Yamamoto is rated at 1092 — 202 points above Fujii's 890. Gaps this large usually mean one fighter has been consistently beating better opponents.
A few statistical edges stand out. Fujii throws significantly more leather — a 0.0 sig. strike per minute gap. Fujii is far more active with takedowns, averaging 0.0 more per 15 minutes. Fujii has tighter striking defense, making opponents miss more often.
The Pick: Kenichi Yamamoto over Katsuhisa Fujii. The model gives Yamamoto a slight nod at 55% — this could easily go either way.
Joe Slick vs Jason DeLucia
The Middleweight matchup features Joe Slick (1-0) taking on Jason DeLucia (1-1).
Slick carries a modest Elo edge (974 to 920), the kind of gap that reflects a slightly better run of form rather than a talent chasm.
A few statistical edges stand out. DeLucia throws significantly more leather — a 0.4 sig. strike per minute gap. DeLucia is far more active with takedowns, averaging 1.9 more per 15 minutes. DeLucia has tighter striking defense, making opponents miss more often.
The Pick: Joe Slick over Jason DeLucia. The model gives Slick a slight nod at 56% — this could easily go either way.
Eugene Jackson vs Keiichiro Yamamiya
The Middleweight matchup features Eugene Jackson (3-3) taking on Keiichiro Yamamiya (0-0).
The Elo gap here is razor-thin — Jackson at 882, Yamamiya at 904. On paper, this is as close to a coin flip as it gets.
A few statistical edges stand out. Jackson throws significantly more leather — a 3.3 sig. strike per minute gap. Yamamiya is far more active with takedowns, averaging 0.0 more per 15 minutes. Yamamiya has tighter striking defense, making opponents miss more often.
The Pick: Eugene Jackson over Keiichiro Yamamiya. We're leaning Jackson here at 69%, a solid but not overwhelming edge.
Kenichi Yamamoto vs Daiju Takase
The Open Weight matchup features Kenichi Yamamoto (2-0) taking on Daiju Takase (0-2).
Yamamoto is rated at 1092 — 298 points above Takase's 794. Gaps this large usually mean one fighter has been consistently beating better opponents.
A few statistical edges stand out. Yamamoto throws significantly more leather — a 0.1 sig. strike per minute gap. Takase is far more active with takedowns, averaging 0.0 more per 15 minutes. Yamamoto has tighter striking defense, making opponents miss more often.
The Pick: Kenichi Yamamoto over Daiju Takase. The model is firm on this one: Yamamoto at 83%.
Katsuhisa Fujii vs Masutatsu Yano
The Open Weight matchup features Katsuhisa Fujii (0-0) taking on Masutatsu Yano (0-0).
Fujii carries a modest Elo edge (890 to 833), the kind of gap that reflects a slightly better run of form rather than a talent chasm.
A few statistical edges stand out. Fujii throws significantly more leather — a 2.9 sig. strike per minute gap. Fujii is far more active with takedowns, averaging 3.2 more per 15 minutes. Yano has tighter striking defense, making opponents miss more often.
The Pick: Katsuhisa Fujii over Masutatsu Yano. The model is firm on this one: Fujii at 79%.
Methodology
Predictions are generated by our ensemble model combining LightGBM (65%) and CatBoost (35%), trained on every UFC fight since 1994. The model uses 23 features including Elo ratings, rolling 5-fight statistical averages, style matchup history, physical attributes, and market odds when available.
On our held-out test set (402 fights from January-September 2023), the model achieves 63.4% accuracy with a log-loss of 0.626. High-confidence picks (>75% probability) hit at 82.7%. For full model transparency, visit our Model page.