Data & Methodology

Fight IQ Analytics uses historical UFC bout and fighter data to calculate its analytical ratings and career metrics.

Data Source

Cito API

Analytical Methodology

Fight IQ Analytics proprietary models

Market data

Sportsbook odds supplied through Cito API

Cito is the data provider, not the creator of any Fight IQ Analytics rating. Fight IQ Analytics is not affiliated with or endorsed by the UFC or Cito.

Cito provides the underlying data. Fight IQ Analytics processes, normalizes, and analyzes that data through its own methodologies, including UCFR, Fight IQ, Matchup Factor, and GOAT models. Records are matched by stable provider IDs, duplicates are prevented, and missing values stay missing — they are never estimated or filled in.

Twelve skill scores — Striking, Power, Wrestling, Grappling, Speed, Cardio, Durability, Fight IQ, Finishing, Pressure, Physicality, Experience — are derived from observable, measured bout data on fixed 0–100 scales, so later data never changes earlier scores. Where Cito does not measure something directly (for example Speed), a clearly labelled measurable proxy is used. Cardio requires qualifying round-level data from Round 3 or later; otherwise it shows Insufficient Evidence.

Fight IQ is a Fight IQ Analytics metric designed to evaluate observable tactical performance and decision-making in competition. It is based on measurable fight behavior and historical performance data rather than a claim about a fighter's general intelligence.

Fight IQ ≠ human intelligence.

Where sufficient data exists, observable components can include tactical efficiency, adaptation between rounds, opponent-specific adjustments, position/risk management, shot selection, game-plan execution and foul/penalty avoidance. The current version uses the subset that the available bout statistics can measure (efficiency-based proxies); components without data are not estimated.

A separate signed score (−10 to +10) describing how two styles interact: striking adjusted for durability and Fight IQ, wrestling versus takedown defense, grappling, reach, height, speed, power versus durability, stance, cardio, pressure, recent form, opponent quality and physicality. A higher UCFR never automatically produces a positive Matchup Factor.

Evidence Strength indicates the completeness and reliability of the underlying data supporting a metric. It is kept separate from every rating.

It is NOT a win probability, a prediction, a quality rating, or a measure of human intelligence.

If data is insufficient, the metric shows "Insufficient Evidence" — no number is manufactured.

GOAT scores are calculated from historical UFC bout data and chronological pre-fight UCFR ratings. They measure a career's body of work and are separate from UCFR. Components, each shown independently: Peak, Resume, Championship, Dominance, Longevity, Consistency, Finishing, Fight IQ. Opponents are valued at their pre-fight UCFR — never a later career rating.

Moneylines are converted to vig-free implied probabilities and compared with the model lean on a common scale (FIQA and Market Align, Mild Divergence, FIQA / Market Divergence). Rating differences are published as Model Rating Difference — never as win probability — until a probability model is calibrated on at least 100 historical fights.

FIQA / Market Divergence compares FIQA's normalized pre-fight analytical lean, derived from UCFR differential and Matchup Factor, with the betting market's no-vig assessment. Both sides are normalized to a common directional scale (−100 to +100) before comparison: FIQA Model Lean = UCFR differential × 6 + Matchup Factor × 5; Market Lean = (no-vig probability for Fighter A − 50%) × 200.

The Divergence Score is the distance between the two leans, clamped to 0–100. Higher values indicate greater disagreement; lower values indicate greater agreement. 0–17 FIQA and Market Align · 18–34 Mild Divergence · 35–100 FIQA / Market Divergence. When the two sides favour different fighters it is shown as Opposite-side disagreement. With no valid two-sided odds: No Market Data. Odds older than 24 hours: Market Data Stale.

FIQA / Market Divergence describes difference, not correctness. It is not a win probability, expected-value calculation or betting recommendation. Evidence Strength, Fight IQ and Fight Pulse™ (community picks) are not inputs. Every calculation is stored as a permanent snapshot, and the final pre-fight value is frozen when the bout begins.

Every calculation stores its model version, timestamp and result. Historical ratings are never overwritten — new versions are added alongside earlier ones. Content is labelled Real Data, Model Output, Market Data or AI-Generated.

New model runs are validated across the historical database before publication: checks for duplicates, date and result consistency, title data, division data and look-ahead leakage. Runs become public only after review.