How our predictions work

The short version

Every prediction comes from a transparent, rule-based model — no gut feeling, no insider whispers. We combine recent form, head-to-head history, attacking and defensive averages, and home advantage into an expected-performance estimate for each side, then publish the pick with a confidence rating before the match starts.

Football

• Form — each team's last five results, weighted into a strength score.
• Goal data — average goals scored and conceded per game feed an expected-goals estimate for the fixture.
• Home advantage — a fixed boost for the home side, reflecting long-run averages across leagues.
• Head-to-head — recent meetings tilt the edge when one side consistently beats the other.

The model outputs a predicted scoreline, a 1X2 winner pick, both-teams-to-score, and an over/under 2.5 goals call.

Cricket

For T20 matches we blend recent win/loss form, batting output (average innings totals) and bowling economy into a composite strength score per side, adjusted for the venue's typical first-innings total. The model picks a winner and a first-innings runs band.

Confidence bands

Each prediction carries a score from 30–95% mapped to three bands: high (70%+), medium (55–69%) and low (below 55%). Bigger gaps between the teams' numbers mean higher confidence — but no prediction is ever certain.

Honesty policy

Predictions are locked when published and resolved automatically once the result is in. Hits and misses both stay on the accuracy page forever. We never edit a prediction after kick-off, and we never sell "premium" tips.