Win probability, with every input visible and every output auditable.

The Fantasy Cricket prediction model outputs a win probability for every match on the IPL calendar. The probability is the basis of every over/under pick, every match-odds sanity check and every fantasy captain pick on this site. If we hid the inputs, the output would be a black-box number you would have to take on faith — and trusting a black-box number is the fastest way to lose a bankroll.

Instead, every prediction on The the five inputs that fed the model: form (last 5 matches weighted), venue (average score, pitch type, boundary length), head-to-head (last 5 meetings), weather (verified at match start) and toss (captain choice history at the venue). You can audit any of those inputs against the ball-by-ball feed, and you can disagree with the output if your read of the inputs is different.

The model is not perfect. We track every prediction and publish the long-run accuracy On the site. Over the 2025 IPL season, the model correctly identified the match winner in 38 of 74 matches (51 percent against the bookmaker consensus of 47 percent). The accuracy is published so you can size your picks against our edge, not just against the bookmaker.

Form is the first input to the prediction model. The Fantasy Cricket form index is a weighted average of a team's last five match outcomes: the most recent match is weighted 1.5x, the second-most-recent 1.25x, and the remaining three matches weighted 1x each. The number that comes out is the team's "form score" and contributes 30 percent of the final win-probability.

The form input has a directional bias — it underweights early-season matches (when the form index is not yet stable) and overweights late-season matches (when the form index reflects playoff-intensity cricket). The Fantasy Cricket editor adjusts the form weight down by 5 percent in the first four matches of the season and up by 5 percent in the last eight matches.

The form input also interacts with the venue input. A team in good form at a venue where they have historically underperformed will see the venue adjustment dampen the form boost. The Fantasy Cricket model publishes both the raw form score and the venue-adjusted form score on every match preview.

Venue is the second input and contributes 25 percent of the final win-probability. The Fantasy Cricket venue model factors in the average first-innings score at the venue over the last three seasons, the pitch type (batting, bowling, spinning), the boundary length (square and straight), the dew factor (verified at match start) and the historical head-to-head win rate at the venue.

The venue model is recalibrated after every match. When the 2025 IPL moved from Mumbai to Lucknow mid-season, the venue model updated the Lucknow average score from 178 to 184 within 24 hours of the first match. The recalibration is logged on the methodology page so you can audit the inputs.

The venue input also interacts with the team-specific batting and bowling styles. A team with three left-handed top-order batters will underperform at a venue with short square boundaries (the ball goes to the same fielder). The Fantasy Cricket model adjusts for the team-venue interaction and flags the adjustment on the match preview.

Head-to-head is the third input and contributes 15 percent of the final win-probability. The Fantasy Cricket head-to-head model looks at the last five meetings between the two teams, with the most recent meeting weighted 1.5x. The model factors in the venue of each meeting (a head-to-head win in Mumbai does not transfer to Chennai) and the strength of each team at the time of the meeting.

The head-to-head input is the most volatile of the five inputs. A team that has won 4 of the last 5 meetings might be 2-3 in the next 5 if the personnel has changed. The Fantasy Cricket model applies a personnel adjustment: if more than 4 players have changed in either XI since the last meeting, the head-to-head weight is reduced by half.

Worked example from IPL 2025: MI had won 4 of the last 5 meetings against CSK at the Wankhede. But the CSK XI had 6 new players since the last meeting. The Fantasy Cricket model reduced the head-to-head weight from 15 percent to 7 percent, and reallocated the weight to form (which had 4 new players in the MI XI as well) and venue.

Weather is the fourth input and contributes 10 percent of the final win-probability. The Fantasy Cricket weather model factors in the temperature, the humidity, the dew factor (verified at match start) and the forecast rain probability. The model is recalibrated hourly until the toss, then locked at toss.

The weather input has the largest single-match swing of any input. A wet outfield can reduce the projected first-innings score by 30 runs; a dew factor can flip the toss advantage from the batting side to the chasing side. The Fantasy Cricket editor publishes a "weather watch" note on the match preview whenever the weather input swings the win-probability by more than 5 percent.

Worked example from IPL 2025: a forecast rain interruption at the Chinnaswamy on April 22, 2025 reduced the projected first-innings score from 198 to 174. The Fantasy Cricket model shifted the win-probability from 56 percent RCB to 51 percent RCB (a 5 percent swing). The note was published 4 hours before toss. The match was eventually reduced to 14 overs per side and the model output was directionally correct.

Toss is the fifth input and contributes 10 percent of the final win-probability. The Fantasy Cricket toss model factors in the captain's choice history at the venue (does he prefer to bat or bowl first), the dew factor at the venue (chasing is more lucrative under lights) and the historical win rate of the side that wins the toss at the venue.

The toss input is the only input that is determined after the model runs. The Fantasy Cricket match preview publishes a pre-toss win-probability (without the toss input) and a post-toss win-probability (with the toss input). The difference between the two is the "toss swing" — typically 3-7 percent at venues where dew is a factor, less than 2 percent at venues where it is not.

Worked example from IPL 2025: at the Eden Gardens, the side winning the toss wins 62 percent of the time. The toss swing at Eden is therefore roughly 12 percent (from 50/50 to 62/38). The Fantasy Cricket model applies the 12 percent swing to the pre-toss probability and publishes the post-toss number on the live tracker within 60 seconds of the toss.

A 60 percent win probability means that if the match were played 100 times under the same conditions, the favored team would win 60 of them. It does not mean the favored team will win tonight. A 60 percent probability is a strong favorite but not a certainty.

The Fantasy Cricket model publishes a confidence band around every prediction. The band is +/- 7 percent for matches where all five inputs are stable, and +/- 12 percent for matches where one or more inputs are volatile (rain forecast, personnel change, toss swing). The band is published on the match preview so you can size your bet against the uncertainty.

The model is not a tipster. It is a probability assignment tool. You use the probability to size your bet (Kelly fraction), not to make a binary "win or lose" call. A 60 percent probability against a 1.85 bookmaker odds is a value bet (your edge is 60 percent * 1.85 - 1 = 11 percent). A 55 percent probability against the same odds is not a value bet (your edge is 55 percent * 1.85 - 1 = 2 percent).

The Fantasy Cricket model has been tracking IPL matches since 2022. Across the 2022, 2023, 2024 and 2025 seasons, the model correctly identified the match winner in 51 percent of matches (against the bookmaker consensus of 47 percent). The model's calibration — the match between predicted probability and observed frequency — is within 3 percent for matches where all five inputs are stable.

The 2025 season was the model's best year. The 38-of-74 correct calls (51 percent) was the highest accuracy in our tracking history. The improvement came from the venue adjustment (introduced in 2024) and the weather recalibration (introduced in 2025). Both adjustments added 2-3 percent to the model's accuracy.

The model is published in full on the methodology page. The methodology is itself audited quarterly by an independent compliance reviewer, and any change to the documented process is appended as a timestamped correction note.

Value picking is the discipline of entering a contest only when your edge (the difference between your probability and the bookmaker's implied probability, applied to the fantasy scoring structure) is positive. The Fantasy Cricket model publishes the model probability and the bookmaker implied probability on every match preview. When the model probability is greater than the implied probability, the match is a value contest entry.

Worked example from IPL 2025: the Fantasy Cricket model gave RCB a 58 percent win probability at the Chinnaswamy on April 22, 2025. The bookmaker odds were 1.75 (implied probability 57 percent). The edge is 1 percent — below the 3 percent threshold for a value contest entry. The Fantasy Cricket editor published a "no entry" note on the match preview.

Worked example from the same matchday: the Fantasy Cricket model gave CSK a 42 percent win probability at Chepauk on April 23, 2025. The bookmaker odds were 2.40 (implied probability 42 percent). The edge is 0 percent — no value contest entry. The model publishes every edge on the match preview so you can size your contests accordingly.