Home · Predictions
Predictive desk

Predictions — what the desk says, and what would change the desk's mind

Every prediction on this page is tied to a public list of inputs that would flip the call. We do not write 'safe picks' that hide the inputs, and we do not promise outcomes.

Inputs published openlyPost-match audit alwaysNo certainty language
An analyst checking a weather reading at a cricket ground under cloud cover.
What we publish

The four-step prediction pattern, every match

A

Pre-toss frame

The pitch, the weather, the dew forecast, the venue split for both teams. We publish the frame the morning of the match and label every claim as confirmed (verifiable fact) or projected (analyst inference).

B

Post-toss update

Once the toss is announced and both XIs are confirmed, we update the captain matrix and rewrite the team sheet to match the new conditions. The post-toss read is what we recommend building from.

C

Captain matrix and risk flags

A captain matrix with floor, ceiling, ownership estimate, role volatility and a one-line 'what would change this call' box. Risk flags are highlighted at the top of the matrix so you do not miss them.

D

Post-match audit

Within four hours of the match ending, we publish a short audit: what the captain matrix got right, where it missed, and what the desk will change in the next fixture. Reading the audit is how the framework compounds.

Frequently asked

Questions our readers ask

We publish a pre-toss prediction once the team sheet is reported, label it as a projection, then update after the toss with a confirmed projection. The post-match audit closes the loop and feeds back into the next fixture's prediction.

We write both. Pre-toss predictions are framed as projections with a confidence tag. Post-toss predictions are reframed against the actual XI and pitch.

No. Some matches have inputs so noisy that a confident pick is dishonest. On those matches we publish the framework and a short list of what would change the call.

Evidence model

Six facts we anchor every prediction to

Role share

Expected overs or position

For every player we publish an expected role share — batting position, expected overs, expected impact-player slot. The number is updated after each match.

Venue split

Phase-by-phase at this ground

Venue split covers the last six matches played at this ground by both franchises, broken into powerplay, middle and death. We never use season-aggregate numbers for prediction reads.

Matchup

Last three head-to-heads

Matchup reads cover pace or spin the player faces, the role they usually play against the opposition's likely combination, and the impact on captain matrix confidence.

Workload

Recent overs and rotation

Workload tags show whether the player is being rested, rotated, or held to a fixed share. Three match windows are used (last 1, last 3, last 6).

Confidence

Why we lean this way

Every prediction carries a one-line why this leans this way and a one-line what would flip this. The two lines are a hard requirement on this desk.

Risk flag

What to hedge against

Each prediction carries a risk flag — the inputs that, if confirmed, would override the call. Examples: Chase is 1W-7L here. Bumrah workload caps at 3 overs tonight.

Risk-flag taxonomy

The 6 flags the desk will not let a prediction ship without

1

Weather and dew

Any reading over 70 percent humidity at the ground or a published dew advisory flips the chase-favourable flag and rewrites the captain matrix on spinners.

2

Toss outcome

If the venue chase record is lopsided, the toss literally picks the winner. The desk flags the lopsided venues on the relevant fixtures.

3

Confirmed XI

An unexpected XI choice (an impact-player slot that wasn't telegraphed, an opener change) can rewrite every pre-toss frame. We hold the prediction until XI is in.

4

Batting order

An unusual batting position move — a finisher opening, an anchor demoted — changes the captain matrix entirely. We flag this when broadcast cameras show the order before the toss.

5

Workload cap

A franchise's stated workload cap on a specific player is the single most reliable sign that the player will underperform their season average.

6

Pitch change

If the pitch is different from the previous match's surface — a black-soil change, a relaid square — the read can change completely. We always name the pitch used.

Confidence ladder

How the desk calibrates the confidence on a prediction

High confidence

Three-plus inputs confirmed

A high-confidence prediction sits on at least three confirmed inputs — venue read, role share, recent form — and no unresolved risk flags. We publish the high-confidence read as the desk's lean and recommend a wider captain pool only on contest entries.

Moderate confidence

Two confirmed, one pending

A moderate-confidence prediction sits on two confirmed inputs and one pending — usually the toss or the dew reading. The desk publishes the read with a hold annotation: the matrix is built so the captain pick can be flipped after the toss without re-engineering the team.

Cautious confidence

One confirmed, multiple pending

A cautious-confidence prediction sits on one confirmed input and multiple pending — typically XI is in but venue read is ambiguous. We publish only the framework on cautious-confidence days and a short list of what would change the call. The captain pool is wider than normal but the lean is weaker.

No call

When the desk stays quiet

The desk publishes no call when the inputs are too noisy — typically when there's rain in the forecast, an unconfirmed XI, and a workload question on the headline bowler. We publish a short note explaining why the call is held, and we update as soon as the noise resolves.

Match-type reads

How the desk calibration changes by match type

Day matches

Day matches read drier, with the ball gripping more from over 8 onward. The captain matrix on day matches tends to favour seamers who exploit early movement and batters who can capitalise on the batting-friendly powerplay before the conditions turn. The differential pool is wider than night matches because the deck behaviour is less linear.

Day matches also produce more spin-friendly conditions in the second innings. Watch for the captain's reading on the chase — if the deck reads driest at the start, the side batting first has the advantage.

Night matches

Night matches read flatter, with dew kicking in from over 13 onward at most venues. The captain matrix on night matches tends to favour chasers — the dew makes the second-innings run chase 8-12 percent easier on average. The captain's reading on the toss is the highest-leverage single input.

Night matches also produce more batting-friendly conditions in the second innings. The captain's differential pool widens on the chasing side's middle-order — a Tier-2 captain is often better than a Tier-1 captain in a chasing context.

Afternoon double-headers

Afternoon double-headers pit a day match against an evening match on the same venue. The captain matrix on the first game is typically slower-spin friendly; the second game's matrix is typically chase-friendly. Read both before locking teams.

The differential on double-header days sits on the second game's bowling side — a spinner who can take wickets in the dew-heavy middle overs. The vice-captain role is wider than usual because the role split between the two matches is unusual.

Back-to-back fixtures

Back-to-back fixtures for a franchise typically mean a workload cap on the frontline bowler and a rotation on one of the frontline batters. The captain matrix on a back-to-back fixture is built around the rotated players — the differential pool is wider than usual because the role share is unclear.

Watch the management notes 48 hours out — a workload cap is usually telegraphed there. We log every management note on the desk tracker and update the matrix accordingly.

Play now