Research methodology

Evidence. Uncertainty. Accountability.

Averix predictions are selected by hand, published before the event and scored against the official result. They are analysis, not guarantees.

How predictions are made

Every Averix prediction is selected and entered manually by our analyst. Picks are not generated by an AI model, machine-learning system or automated probability model. Fight cards are imported from the official UFC schedule; the prediction for each bout is a human judgment.

What the analysis considers

Each matchup is assessed on factors including:

  • Fighting style and how the styles match up
  • Striking
  • Wrestling and grappling
  • Recent performance
  • Quality of opposition
  • Physical attributes
  • Cardio
  • Historical fight data
  • Market and odds context, where available

The weight given to each factor depends on the matchup. No formula converts these factors into a pick.

What each prediction includes

  • Winner: the fighter the analyst expects to win.
  • Confidence: a qualitative Low, Medium or High assessment of the pick. It is not a percentage or a probability.
  • Method: the predicted finish: KO/TKO, submission or decision.
  • Round: the predicted round, within the bout’s scheduled rounds.
  • Analysis: the written reasoning behind the pick.

Method and round are analyst predictions, not certainties.

Market context

Where a current Polymarket market is confidently matched to a bout, its market-implied percentages appear beside the matchup for reference. Those figures belong to the market: Averix publishes no probability of its own, and market prices never select a pick.

Publication and record-keeping

Picks remain private drafts until the card is reviewed and published, which must happen before the event. Each publication is timestamped and kept as a numbered version that cannot be edited. A revision is published as a new version, and earlier versions are retained.

How results are scored

Each published pick is scored against the official result. Each fight counts once, using the first version published before the event, so a later revision cannot replace an earlier loss. Draws and no contests are excluded from accuracy, and picks without an official result stay pending. Scoring is kept internally for grading and review; Averix does not publish a historical record, accuracy figure or win rate.

Limits

MMA outcomes are uncertain, and small samples can mislead. Averix provides sports analysis and predictions, not guaranteed outcomes. No prediction guarantees accuracy, returns or betting success, and past performance is not a guarantee of future results.

  1. 01Import the official card
  2. 02Assess styles and matchup
  3. 03Review form and opposition
  4. 04Check market context
  5. 05Pick winner, method and round
  6. 06Rate confidence
  7. 07Write the analysis
  8. 08Publish before the event
  9. 09Record the official result
View the current predictions