CricRatings Methodology

How CricRatings scores a performance

CricRatings converts each innings (and match) into a normalized performance score. The idea is simple: cricket is not played on a spreadsheet, so the model evaluates output and difficulty. Since we use scorecards (not ball-by-ball), the system is designed to extract context from what scorecards reliably contain across eras.

1) The base layer: production still rules

The base score starts with the obvious: batting runs, bowling wickets, and supporting scorecard signals (e.g., strike rate / economy, workload, contribution to team totals, and role-appropriate expectations where available). If you don’t score runs or take wickets, you don’t magically become Player of the Match here. (Nice try.)

2) The context layer: why that production mattered

The model then adjusts performance using multiple context lenses, including:

• Peer relativity within the match: how your output compares to everyone else in the same innings/match (a “same-day difficulty” filter).
• Opponent strength: runs against elite attacks and wickets against elite batting line-ups carry higher value than easy-mode cricket.
• Match situation / leverage: contributions when the game is alive matter more than padding after the result is effectively sealed.
• Home vs away: away performance is normalized differently because conditions and familiarity aren’t equal.
• Era benchmarking: numbers are normalized against the scoring/wicket environment of their time, so 1930 isn’t judged with 2025 goggles.
• Big-match importance: World Cups, ICC events, knockouts, finals, and decisive games have higher leverage.

3) The “outstanding performance” triggers (what makes an innings special)

CricRatings especially rewards performances that cricket people intuitively call “proper match-shaping.” Three recurring criteria show up again and again in top-rated innings/spells:

• Fourth-innings runs (and first-innings wickets): fourth-innings chases amplify pressure; first-innings wicket bursts often define the match’s entire shape. In Tests, these are classic “this decided the game” moments.
• Big-match performance: World Cups, ICC tournaments, finals/knockouts, and high-stakes deciders are weighted higher. Same output, higher consequence.
• Innings situation value: context like 5/3, 20/2, 80/5, or a chase wobble turns “runs” into “rescue runs.”

4) Normalization & aggregation: turning scores into ratings

Once innings scores are computed, they are normalized so they can be compared across eras, formats, conditions, and opponent profiles. Those normalized outputs are then aggregated differently depending on the module: recency-weighted for “Latest Ratings”, career accumulation + consistency for “All-Time Ratings”, and time-sliced for annual/historical views.

What we intentionally don’t use

CricRatings does not use ball-by-ball data. That means we do not directly model micro-events like dot-ball pressure, phase-by-phase scoring, or wagon-wheel intent. Instead, we infer match difficulty and leverage using robust scorecard-level proxies and relativity inside the match/era.