Cuotario methodology: data, samples and football predictions

1. Fixtures, results and statistics

Cuotario organizes available records by competition, team and match. Fixture and score coverage can differ from coverage for corners, shots, fouls or cards. Missing observations are not zeros; every statistic can have a different sample. The provider may correct times, lineups and other records.

2. Official leagues and cups

Enabled team markets compare team production with the opponent’s conceded output using official club history from leagues, cups and international competitions. Teams are identified by their IDs. Each competition retains its current season or phase within a maximum 180-day window before the cutoff day.

Friendlies, national-team fixtures, partial statistics and matches identified as involving extra time or penalties are excluded. Duplicate fixtures count once. The analyzed match and later information must not enter a reconstructed pre-match prediction.

Expansion is enabled by competition and market after retrospective comparison. Other segments retain their existing basis. In the Champions League, enabled corners, shots on target, fouls, cards and offsides use the expansion; total shots retain the earlier basis. This expansion does not change the 1X2 model or individual player markets.

3. Observed and effective samples

Time weighting halves every 90 days. A match in another competition receives a factor of 0.70. Where sufficient information exists, similarity in pre-match favoritism and the reference lineup also affects weights. Venue, scenarios and referee context are retained.

A raw match count is not the effective sample after weighting. The detail identifies both and shows the contributing competitions. At least three effective observations are retained on each side of the comparison before suggesting a selection; that minimum is not lowered to show more markets.

Public descriptive averages are unweighted summaries, so they can differ from model projections.

4. Averages, projections and probabilities

An average summarizes observations. A projection estimates a quantity for a match. A probability estimates whether a particular condition will hold. A projected quantity is not a guaranteed exact result, and a high probability still allows a loss.

A small sample can vary greatly. Different metrics or markets may have different coverage and must not be treated as interchangeable.

5. Lineups and missing players

Distinguish a confirmed eleven from a projected one and check the information timestamp. For enabled markets, lineup similarity influences earlier-match weights when coverage permits. The descriptive volume associated with missing players is not automatically subtracted from the projection. Player markets retain their separate model.

6. Odds and evaluation

The model’s minimum odds are a comparison threshold, not an operator’s offered price or a guarantee of profit. Compare equivalent settlement rules and current prices. Historical accuracy alone does not establish profitability.

Evaluate selections recorded before kickoff against later outcomes. A retrospective reconstruction of the current model is different from a record of what was actually published at the time. Do not infer live performance from a small or selectively chosen example set.

7. Reading the public pages

English dates and times on the today page use UTC. The interactive app labels local kickoff times. Team profiles retain the competition in their address and identify historical periods when appropriate. Coverage and update times vary by market.

Explore the football-statistics guides and the today page to compare these parts separately. No prediction guarantees a result or a financial return.