UFC 333, scheduled for 2026-10-24 at Etihad Arena in Abu Dhabi, presents a fourteen-fight card whose defining feature is not any single matchup but the narrowness of the model’s separation across most of it. The elo-v1.1 forecast, built from UFC results between 2019-01-01 and the present, assigns a mean favourite probability of 56.4% across the card and a rating band spanning 1488.9 to 1702.8. Ten of the fourteen fights carry low confidence, four carry medium, and none carries high. The card therefore tests something specific: whether a rating system that has been deliberately conservative in its confidence labels can still discriminate between fighters when the underlying numbers sit close together, and whether the two genuinely wide gaps on the card — Merab Dvalishvili vs. Petr Yan at 13.4 points and Azamat Murzakanov vs. Dominick Reyes at 11.4 points — behave like the mismatches their ratings suggest.
The main event adds a second test. Alexander Volkanovski, rated 1620.5 with a 9-3 record in the window, is forecast as the underdog at 39.6% against Movsar Evloev, rated 1693.6 and 10-0. That is a 10.4-point edge for the undefeated challenger, and it is the model’s fourth-widest margin on the card rather than its widest. The featherweight title picture thus sits inside a card where the model’s largest disagreement with the eye is not the headliner at all. Across the full slate the model expects 6.2 finishes and assigns a 0.03% probability that every favourite wins — a figure that frames the entire breakdown: this is not a card the model treats as a formality.
The card in one page
Medium confidence (4 fights)
- Bantamweight: Merab Dvalishvili (63.4%, rating 1702.8, 13-1) vs. Petr Yan (36.5%, rating 1607.0, 9-4) — edge 13.4 pts; finish probability 35.8%.
- Light Heavyweight: Azamat Murzakanov (61.4%, rating 1606.5, 7-1) vs. Dominick Reyes (38.6%, rating 1525.8, 6-5) — edge 11.4 pts; finish probability 50.7%.
- Light Heavyweight: Nikita Krylov (38.9%, rating 1510.5, 6-6) vs. Abdul-Rakhman Yakhyaev (61.1%, rating 1589.2, 4-0) — edge 11.1 pts; finish probability 55.8%.
- Featherweight (main event): Alexander Volkanovski (39.6%, rating 1620.5, 9-3) vs. Movsar Evloev (60.4%, rating 1693.6, 10-0) — edge 10.4 pts; finish probability 30.0%.
Low confidence (10 fights)
- Lightweight: Grant Dawson (58.4%, rating 1664.5, 12-2) vs. Nurullo Aliev (41.6%, rating 1605.5, 5-0) — edge 8.4 pts; finish probability 42.5%.
- Featherweight: Daniel Santos (58.1%, rating 1545.4, 4-2) vs. Bolaji Oki (41.9%, rating 1488.9, 3-3) — edge 8.1 pts; finish probability 44.6%.
- Middleweight: Abubakar Vagaev (42.9%, rating 1522.9, 1-0) vs. Jacob Malkoun (57.1%, rating 1572.3, 6-3) — edge 7.1 pts; finish probability 30.0%.
- Heavyweight: Alexander Volkov (56.7%, rating 1634.5, 10-4) vs. Rizvan Kuniev (43.3%, rating 1587.5, 4-1) — edge 6.7 pts; finish probability 35.8%.
- Women’s Strawweight: Nicolle Caliari (44.5%, rating 1499.9, 2-2) vs. Iasmin Lucindo (55.5%, rating 1558.6, 5-2) — edge 5.5 pts; finish probability 42.5%.
- Flyweight: Ramazan Temirov (53.0%, rating 1566.4, 3-0) vs. Lone’er Kavanagh (46.9%, rating 1545.2, 4-2) — edge 3.0 pts; finish probability 55.0%.
- Bantamweight: Colby Thicknesse (52.1%, rating 1514.6, 2-1) vs. Asaf Chopurov (47.9%, rating 1500.0, 0-0) — edge 2.1 pts; finish probability 30.0%.
- Middleweight: Abus Magomedov (48.2%, rating 1562.3, 5-3) vs. Cam Rowston (51.8%, rating 1575.1, 4-1) — edge 1.8 pts; finish probability 51.4%.
- Featherweight: Aaron Pico (49.3%, rating 1509.2, 1-1) vs. Losene Keita (50.7%, rating 1514.0, 1-1) — edge 0.7 pts; finish probability 46.2%.
- Lightweight: Noah Gugnon (50.3%, rating 1527.6, 1-0) vs. Kody Steele (49.6%, rating 1525.2, 2-1) — edge 0.3 pts; finish probability 65.0%.
High confidence: none.
The shape is a tight card with a finish-heavy tail. Six of the fourteen fights carry a finish probability at or above 44.6%, and the single highest finish figure on the card — 65.0% for Gugnon vs. Steele — belongs to the closest fight on the card by rating, a 0.3-point edge. The card is not mismatch-heavy: only two fights clear an 11-point edge, and the median edge sits well inside single digits. It is also not a chalk card in any meaningful sense, given the 0.03% probability attached to all favourites winning.
Main event: Alexander Volkanovski vs. Movsar Evloev
The number
Evloev is forecast at 60.4% with a rating of 1693.6 and a 10-0 record in the window. Volkanovski is forecast at 39.6% with a rating of 1620.5 and a 9-3 record. The edge is 10.4 points, and the model labels the fight medium confidence. The finish probability is 30.0%, split as Evloev KO/TKO 9.1% and submission 9.1%, and Volkanovski KO/TKO 5.9% and submission 5.9%. That is the second-lowest finish probability among the four medium-confidence fights and the lowest of any fight on the card above low confidence.
The case for Evloev
The case rests on the rating gap and the undefeated record. Evloev’s 1693.6 is the second-highest rating on the entire card, behind only Dvalishvili’s 1702.8, and his 10-0 mark in the rating window means the model has no UFC loss to discount against him. Volkanovski’s 9-3 record in the same window is strong but carries three losses, and the Elo construction treats those as real information rather than noise. The 10.4-point edge is the model’s way of saying the gap is meaningful without being decisive — it is a medium-confidence read, not a statement that the fight is one-sided. The finish split reinforces the point: the model gives Evloev a combined 18.2% chance of a stoppage, which is modest for a fighter with his rating advantage and reflects that Volkanovski’s own durability is priced into the rating rather than assumed away.
The case for Volkanovski
Volkanovski’s 39.6% is not a token number. A rating of 1620.5 places him comfortably inside the card’s band, and the model’s medium confidence label means the 10.4-point edge is treated as real but not overwhelming. The 30.0% finish probability cuts both ways: it is the lowest finish figure on the card above low confidence, which means the model expects this fight to go to the scorecards more often than not, and a decision-heavy fight is one where a 10.4-point rating gap translates into a smaller practical advantage than the raw percentage suggests. Volkanovski’s combined 11.8% stoppage share is nearly two-thirds of Evloev’s 18.2%, so the paths to a Volkanovski win are not confined to the judges.
What would change the read
The read would shift if the rating inputs changed — a different rating window, or a documented inactivity adjustment applied to either fighter, would move the 1620.5 and 1693.6 figures and therefore the 39.6/60.4 split. The medium confidence label itself is the signal: it means the model sees the gap as real but not robust enough to promote to high confidence, and any change to the underlying results would be more likely to move this fight’s number than the low-confidence fights where the edges are already small. The 30.0% finish probability would also move if the method distribution changed, since it is built from each fighter’s finish record rather than from any assessment of how the styles interact.
Co-main: Merab Dvalishvili vs. Petr Yan
The model’s widest edge on the card belongs here: Dvalishvili at 63.4% with a rating of 1702.8 and a 13-1 record, against Yan at 36.5% with a rating of 1607.0 and a 9-4 record. The rating gap is 95.8 points, the largest on the card, and the edge is 13.4 points. The finish probability is 35.8%, split as Dvalishvili KO/TKO 10.2% and submission 12.5%, Yan KO/TKO 5.9% and submission 7.2%. Dvalishvili’s combined 22.7% stoppage share is more than double Yan’s 13.1%.
The fight earns medium confidence rather than high for a structural reason: the model’s confidence labels are driven by the size and stability of the rating gap, and while 13.4 points is the widest on this card, it is not wide enough in absolute terms to clear the threshold the model uses for high confidence. The 35.8% finish probability is moderate — the model expects a decision more often than a stoppage — and a decision-heavy fight compresses the practical effect of a rating gap. Dvalishvili’s 1702.8 is the highest rating on the card, which is why the edge is the widest, but the model does not treat the highest rating on a tight card as equivalent to a high-confidence mismatch.
The rest of the main card
Azamat Murzakanov vs. Dominick Reyes (Light Heavyweight). Murzakanov is forecast at 61.4% with a rating of 1606.5 and a 7-1 record; Reyes at 38.6% with a rating of 1525.8 and a 6-5 record. The edge is 11.4 points, the second-widest on the card, and the finish probability is 50.7% — the highest among the medium-confidence fights. The split is Murzakanov KO/TKO 15.6% and submission 15.6%, Reyes KO/TKO 9.8% and submission 9.8%. The 80.8-point rating gap and the elevated finish probability together make this the model’s clearest candidate for a stoppage among the medium-confidence bouts, though the 50.7% figure still leaves nearly half the probability mass on a decision or other outcome.
Nikita Krylov vs. Abdul-Rakhman Yakhyaev (Light Heavyweight). Yakhyaev is forecast at 61.1% with a rating of 1589.2 and a 4-0 record; Krylov at 38.9% with a rating of 1510.5 and a 6-6 record. The edge is 11.1 points and the finish probability is 55.8%, the second-highest on the card. The split is Yakhyaev KO/TKO 17.1% and submission 17.1%, Krylov KO/TKO 10.8% and submission 10.8%. Yakhyaev’s undefeated record in the window is doing the rating work here, and the 78.7-point gap is the third-widest on the card. The high finish probability reflects the method distribution rather than any assessment of how the two styles match up.
Grant Dawson vs. Nurullo Aliev (Lightweight). Dawson is forecast at 58.4% with a rating of 1664.5 and a 12-2 record; Aliev at 41.6% with a rating of 1605.5 and a 5-0 record. The edge is 8.4 points and the finish probability is 42.5%, split as Dawson KO/TKO 11.2% and submission 13.7%, Aliev KO/TKO 8.0% and submission 9.7%. Dawson’s 1664.5 is the third-highest rating on the card, but the 59-point gap to Aliev is not enough to lift the fight out of low confidence. Aliev’s undefeated record means the model has no loss to discount against him, which is part of why the edge stays inside single digits.
Daniel Santos vs. Bolaji Oki (Featherweight). Santos is forecast at 58.1% with a rating of 1545.4 and a 4-2 record; Oki at 41.9% with a rating of 1488.9 and a 3-3 record. The edge is 8.1 points and the finish probability is 44.6%, split as Santos KO/TKO 12.9% and submission 12.9%, Oki KO/TKO 9.3% and submission 9.3%. Oki’s 1488.9 is the lowest rating on the card, which is why the 56.5-point gap produces only an 8.1-point edge — the model’s rating band is anchored at the bottom by this fight.
Alexander Volkov vs. Rizvan Kuniev (Heavyweight). Volkov is forecast at 56.7% with a rating of 1634.5 and a 10-4 record; Kuniev at 43.3% with a rating of 1587.5 and a 4-1 record. The edge is 6.7 points and the finish probability is 35.8%, split as Volkov KO/TKO 10.2% and submission 10.2%, Kuniev KO/TKO 7.8% and submission 7.8%. The 47-point rating gap is real but modest, and the 35.8% finish probability is the joint-lowest among the low-confidence fights with a clear favourite.
Nicolle Caliari vs. Iasmin Lucindo (Women’s Strawweight). Lucindo is forecast at 55.5% with a rating of 1558.6 and a 5-2 record; Caliari at 44.5% with a rating of 1499.9 and a 2-2 record. The edge is 5.5 points and the finish probability is 42.5%, split as Lucindo KO/TKO 10.6% and submission 13.0%, Caliari KO/TKO 8.5% and submission 10.4%. The 58.7-point rating gap is larger than the 5.5-point edge might suggest, which is a reminder that the edge is a probability difference and not a direct rescaling of the rating gap.
Ramazan Temirov vs. Lone’er Kavanagh (Flyweight). Temirov is forecast at 53.0% with a rating of 1566.4 and a 3-0 record; Kavanagh at 46.9% with a rating of 1545.2 and a 4-2 record. The edge is 3.0 points and the finish probability is 55.0%, the third-highest on the card. The split is Temirov KO/TKO 14.6% and submission 14.6%, Kavanagh KO/TKO 12.9% and submission 12.9%. A 21.2-point rating gap producing a 3.0-point edge is the model’s clearest example of how small rating differences translate into near-even probabilities.
Colby Thicknesse vs. Asaf Chopurov (Bantamweight). Thicknesse is forecast at 52.1% with a rating of 1514.6 and a 2-1 record; Chopurov at 47.9% with a rating of 1500.0 and an 0-0 record. The edge is 2.1 points and the finish probability is 30.0%, split as Thicknesse KO/TKO 7.8% and submission 7.8%, Chopurov KO/TKO 7.2% and submission 7.2%. Chopurov’s 0-0 record means the model has no UFC results to rate him on, and the 1500.0 figure is the model’s neutral starting point rather than a measured rating — a limitation that the low confidence label reflects.
Abus Magomedov vs. Cam Rowston (Middleweight). Rowston is forecast at 51.8% with a rating of 1575.1 and a 4-1 record; Magomedov at 48.2% with a rating of 1562.3 and a 5-3 record. The edge is 1.8 points and the finish probability is 51.4%, split as Rowston KO/TKO 12.0% and submission 14.7%, Magomedov KO/TKO 11.2% and submission 13.6%. The 12.8-point rating gap is the second-smallest on the card, and the 51.4% finish probability sits above the card’s 6.2-finish expectation per fight.
Aaron Pico vs. Losene Keita (Featherweight). Keita is forecast at 50.7% with a rating of 1514.0 and a 1-1 record; Pico at 49.3% with a rating of 1509.2 and a 1-1 record. The edge is 0.7 points and the finish probability is 46.2%, split as Keita KO/TKO 11.7% and submission 11.7%, Pico KO/TKO 11.4% and submission 11.4%. A 4.8-point rating gap and a 0.7-point edge make this the second-closest fight on the card by probability.
Noah Gugnon vs. Kody Steele (Lightweight). Gugnon is forecast at 50.3% with a rating of 1527.6 and a 1-0 record; Steele at 49.6% with a rating of 1525.2 and a 2-1 record. The edge is 0.3 points — the closest fight on the card — and the finish probability is 65.0%, the highest on the entire card. The split is Gugnon KO/TKO 14.7% and submission 18.0%, Steele KO/TKO 14.5% and submission 17.8%. The model treats this as a near coin-flip with a strong likelihood of a stoppage, which is the most internally tense read on the card.
Get the free foundational lesson
The MCMA System starter lesson, plus new MCMA articles by email. No spam, unsubscribe any time.
Card-level reads
The distribution of probabilities implies a card where the model expects upsets in volume. With a mean favourite probability of 56.4% and a 0.03% probability that every favourite wins, the model is not describing a slate of formality. The 6.2 expected finishes across fourteen fights is a moderate figure — below one finish every two fights — and it is concentrated in the low-confidence bouts: Gugnon vs. Steele at 65.0%, Krylov vs. Yakhyaev at 55.8%, Temirov vs. Kavanagh at 55.0%, Magomedov vs. Rowston at 51.4%, and Murzakanov vs. Reyes at 50.7% are the five highest finish probabilities on the card, and four of those five are low-confidence fights.
The chalk risk sits in two places. The first is the medium-confidence tier, where the model’s four widest edges all live: Dvalishvili at 63.4%, Murzakanov at 61.4%, Yakhyaev at 61.1%, and Evloev at 60.4%. Those are the four fights where a favourite loss would be most costly to the model’s calibration, and they are also the four fights where the model has the most to lose from a single result. The second is the cluster of near-coin-flips — Pico vs. Keita at 0.7 points, Magomedov vs. Rowston at 1.8 points, Thicknesse vs. Chopurov at 2.1 points, Temirov vs. Kavanagh at 3.0 points, and Gugnon vs. Steele at 0.3 points — where the model’s own numbers say the outcome is close to a toss-up and where a run of favourites winning would be a genuine surprise rather than a confirmation.
The rating band of 1488.9 to 1702.8 spans 213.9 points, but the probability spread is much narrower than that range would suggest: the widest edge on the card is 13.4 points and the closest is 0.3 points. That compression is the card’s signature. It is a slate where the model’s ratings differ meaningfully but its probabilities do not, and where the confidence labels — ten low, four medium, zero high — are the most important output in the payload.
Methodology and limits
The elo-v1.1 ratings are built only from UFC results in the window 2019-01-01 to the present, sourced from the ESPN official scoreboard. The model applies a documented inactivity adjustment, which means a fighter’s rating reflects time away from competition as well as results. Method distributions — the KO/TKO and submission splits reported for each fight — are derived from each fighter’s finish record rather than from any assessment of how two styles interact. The model does not incorporate striking volume, takedown statistics, or control statistics; those inputs are not in the model, and any reasoning that depends on them is outside what these numbers can support. The model has not watched tape, does not weigh in, and does not know about injuries, camps, or news. The confidence labels — ten low, four medium, zero high — are properties of the rating gaps and the underlying data, not of any external assessment.
How to read this before Saturday
First, treat the confidence labels as the headline. Ten low-confidence fights and four medium-confidence fights mean the model is telling the reader that most of this card is close, and the percentages should be read as narrow separations rather than predictions of dominance.
Second, note that the widest edge on the card, Dvalishvili vs. Yan at 13.4 points, is still a medium-confidence read. A 13.4-point edge is meaningful but not decisive, and the 35.8% finish probability means the model expects a decision more often than a stoppage in that fight.
Third, the finish probabilities and the win probabilities are separate outputs. Gugnon vs. Steele is the closest fight on the card at 0.3 points and also the most likely to end inside the distance at 65.0%; a high finish probability does not imply a lopsided fight.
Fourth, the 0.03% probability that every favourite wins is the card-level summary: the model does not expect a chalk card, and the reader should expect the results to diverge from the favourites somewhere on the slate.
Forecasts are model probabilities, not guarantees and not betting advice.
Sources
Card artwork: UFC.
