How we calculate win chances

Some upcoming match pages show each side’s estimated chance of winning and a short form guide. The numbers are calculated automatically from completed results in crix11’s own data, with no AI. They are estimates, not a guarantee. When a match has a preview article, the page shows the figures the preview was written from, with the date of the latest result they use.

Form guide

The form guide lists each team’s last five completed matches in our data from the 12 months before the match, newest first, in any format: won (W), lost (L), tied (T), drawn (D) or no result (NR), with the opponent, margin and date. A tie decided by a Super Over counts as a win for the side that won the Super Over, as on our team pages. Men’s and women’s teams are always separate.

Win chances

• Rating. Every team starts at 1500. After each completed result, the winner gains points and the loser drops by the same amount (an Elo-style rating, 24 points per result at most). Beating a higher-rated side moves the ratings more than beating a lower-rated one. A tie counts as half a win; a tie decided by a Super Over counts as a win for the Super Over winner. No-result matches are not rated. Margins of victory are not used.
• Same format only. T20 internationals and T20 leagues are rated together, ODIs and other 50-over matches are rated together. Results in one format do not affect the other.
• Last two years. Only results from the two years before the match count.
• Breaks between seasons. When a team has not played in that format for more than 90 days, its rating is pulled a third of the way back to 1500 before its next match, because squads change between seasons.
• Men and women separate. A men’s side and a women’s side are never rated against each other.
• Enough data. Both teams need at least five rated results in that format in our data, and the two teams must be linked through results (they have played each other, or played teams that played each other). Otherwise we show the form guide only.
• Pulled toward 50/50. The rating gap gives a raw chance, which is then pulled toward 50/50 when data is thin: with n rated results for the team with fewer, the gap from 50% is multiplied by n ÷ (n + 8). No side is shown below 15% or above 85%.
• Whole numbers. The two chances are rounded to whole percentages that add up to 100.
• No Tests. Test and other multi-day matches can end in a draw, so no win chance is shown for them.

How we test it

Before any numbers are shown, we replay the method on every completed match in our data: each match is estimated only from the results before it, exactly as it would have appeared, and compared with what happened. We score this with the Brier score (the average squared gap between the estimate and the result; lower is better). Calling every match 50/50 scores 0.25. We show win chances for a format only after at least 30 replayed matches and only while the method scores better than calling every match 50/50.

Results from 18 Apr 2026 onward, updated with each data refresh. Formats not listed have no replayed matches yet.

Limits

The estimate only knows results that are in our data: the series we follow, from our data feed. Most of the series we follow involve India, so many opponents are rated mainly from their matches against India. It does not know the playing XI, injuries, pitch, weather, toss or home advantage. A team with few matches in our data may be rated poorly. Cricket results often go against the numbers.

crix11 does not offer, promote or link to betting, gambling or real-money fantasy games, and these estimates are not betting advice.