Tonight's pick card, named and reviewed
Each pick on the desk includes the source, the weighting, and the failure mode it prevents. Self-audit line stays on the page when we get a pick wrong.
What we got wrong last match
Last match (Bengaluru vs Hyderabad, 2 Apr): we backed the second-innings chase path on a dewy Chinnaswamy surface. The dew window never opened and the team batting first defended. We are updating our dew-band rule to require a +2°C dew-point drop before flipping the chase path.
Six matches, six cards
Captain pick — top-order bat
Second-innings dew path on Chinnaswamy. Ownership band mid (15–30%). Failure mode: pitch grips early and the team bowls first.
Vice-captain — leg-spinner
Eden Gardens surface assists turn in the 7–14th over window. Failure mode: Eden rolls out flat and the spinners go at 9+ RPO.
Differential — lower-order bat
Hyderabad's #6 has a clean strike rate on flat Uppal tracks. Ownership band low (under 10%). Failure mode: early collapse limits the #6 window.
Captain pick — pace all-rounder
Wankhede sea breeze and short straight boundary suit the medium-pace angle. Failure mode: dew arrives late and the bowler is asked to defend.
Vice-captain — opener
Top-order run-out on a dry BRSABV surface. Ownership band mid. Failure mode: powerplay loss of 2+ wickets in the first 3 overs.
Tournament window — review after lineup
Lineup pending. Re-read after the toss. No captain pick until the XI is verified.
Methodology, on the page
Recent form leads, venue history supports
The desk's weighting rule: form is the lead, venue is the support, head-to-head is the tiebreaker. No single number ships without a labelled source — a directional band, a sourced figure, or a marked "tournament window".
Two reads, no group-think
Each match gets two reads on the desk: a cricket analyst and a fantasy-XI analyst. When the picks diverge, the page names the divergence and the rule that resolved it.
The 2× and 1.5× lesson
Captain gets 2× points, vice-captain gets 1.5×. The safest cap-taincy call is the highest floor, not the highest ceiling. The differential call is the highest ceiling — but only when the contest rules cap exposure.
How the desk turns fixtures into picks
The desk's pick-card generation is a three-step read. Step one is the form book, where the last six innings of each player in the squad are checked against the venue, the opposition, and the role. Step two is the venue walk, where the curator's note, the dew window, and the boundary rope are read against the squad's strengths. Step three is the head-to-head, where the last ten games between the two sides are checked against the role. The three steps are then weighted: form leads, venue supports, head-to-head is a tiebreaker. The reader who has read the three steps can argue with the desk's weightings, because the weightings are on the page.
The captain-band pick is the pick whose downside is the smallest. The captain gets 2× points, so the captain pick is the single most leveraged slot in the contest entry. The lean pick is the highest-floor pick - the pick whose performance is the most predictable in the conditions. The alt pick is the highest-ceiling pick with the conditions relaxed - the pick that wins a contest if the contest's conditions swing the way the alt pick expects. The differential pick is the lowest-ownership pick whose ceiling is the highest among the unowned options. The differential pick is opted into only when the contest rules cap exposure.
The ownership band is a directional band - low (under 10%), mid (15–30%), high (over 30%) - because the desk does not author the platform's ownership data. The desk's rule is the band, not the single number. The reader who needs a single number is the reader who should look at the platform's published ownership grid. The band is the desk's contract with the reader: the band is what the desk can defend with a named source.
The failure mode is the part of the pick card that protects the reader's contest entry. The failure mode names the conditions that would flip the pick. The failure mode for a top-order bat captain is "if the team bowls first and the pitch grips early, the pick flips." The failure mode for a pace all-rounder captain is "if dew arrives late and the bowler is asked to defend, the ceiling collapses." The failure mode is the pick's hedge - the desk's reading of the conditions that would make the pick wrong.
The self-audit line
The self-audit line stays on the page when the desk gets a pick wrong. The self-audit line names the missed pick, the rule that failed, and the date the rule was updated. The self-audit line is the desk's contract with the reader about the desk's improvement. The self-audit line is the only honest way to know whether the desk's reading is improving over time. The reader who reads the self-audit line has a better picture of the desk's reading than the reader who reads only the verified band.
The self-audit line is published on the predictions page, not on the homepage. The homepage's winners block reports the verification band across the sample; the self-audit line is the per-pick disclosure. The reader who reads both has the full picture. The reader who reads neither has the marketing pitch, and the marketing pitch is the desk's editorial failure.
What the desk does not promise
The desk does not promise a winning pick on every match. The desk does not promise a single precision number on the verification band. The desk does not promise a captain pick that does not have a named source. The desk does not promise contest entries that survive the toss slippage - the desk promises the conditions report that makes the toss slippage legible. The desk's honesty is the desk's defence against the alternative: a desk that promises everything and delivers nothing, and a reader who reads the desk as a single-source-of-truth rather than as one input among many.
Three failures, three rule updates
The first failure was the Bengaluru vs Hyderabad match on 2 April. The desk backed the second-innings chase path on a dewy Chinnaswamy surface. The dew window never opened and the team batting first defended. The rule that failed was the dew-band threshold: the desk had set the threshold at a 1.5°C dew-point drop, but the actual dew-point drop on the night was 0.8°C. The rule update is to require a +2°C dew-point drop before flipping the chase path. The rule update is logged on the methodology block. The reader who reads the rule update and the conditions report together has a better picture of the contest entry than the reader who reads the original pick alone.
The second failure was the Kolkata vs Rajasthan match on 28 March. The desk backed the wrist-spinner as the captain pick on Eden Gardens. The wrist-spinner went at 9.2 RPO and the captain pick collapsed. The rule that failed was the surface threshold: the desk had set the spin-friendly threshold at a 65/65 covering, but the actual covering was 72/70. The rule update is to require a 60/60 covering or worse before backing the wrist-spinner as captain. The rule update is logged on the methodology block. The reader who reads the rule update and the conditions report together has a better picture of the contest entry than the reader who reads the original pick alone.
The third failure was the Mumbai vs Delhi match on 25 March. The desk backed the top-order bat as the captain pick on a flat Wankhede surface. The top-order bat was dismissed in the powerplay and the captain pick collapsed. The rule that failed was the powerplay protection: the desk had set the protection threshold at a 6.5 powerplay strike rate, but the actual powerplay strike rate was 4.8. The rule update is to require a 7.5 powerplay strike rate before backing the top-order bat as captain on a flat surface. The rule update is logged on the methodology block. The reader who reads the rule update and the conditions report together has a better picture of the contest entry than the reader who reads the original pick alone.
The three failures are the desk's honesty. The three failures are the desk's contract with the reader. The three failures are what the desk can defend with a named source. The three failures are the desk's wall against the false precision.
How the desk reads the toss in three windows
The first window is the toss itself. The toss is the swing factor. The toss is the rule that decides which of the two reads - the lean pick or the alt pick - the desk is shipping. The desk reads the toss as a binary, not as a probability. The desk ships the lean pick if the toss goes the way the desk's reading expects; the desk ships the alt pick if the toss goes the other way. The desk does not ship a probability-conditional pick because the probability-conditional pick is the place where the desk's reading is thinnest and the reader's contest entry is most exposed.
The second window is the lineup sheet. The lineup sheet is the swing factor. The lineup sheet is the rule that decides which of the contest rules applies - the captain pick, the vice-captain pick, the differential pick, the bench pick. The desk reads the lineup sheet as a labelled list, not as a probability. The desk ships the captain pick if the lineup is verified; the desk ships the alt pick if the lineup is pending. The desk does not ship a lineup-conditional pick because the lineup-conditional pick is the place where the desk's reading is thinnest and the reader's contest entry is most exposed.
The third window is the pitch and the conditions. The pitch and the conditions are the swing factor. The pitch and the conditions are the rule that decides which of the surface reads applies - the dew path, the spin watch, the heat index, the boundary rope. The desk reads the pitch and the conditions as a labelled list, not as a probability. The desk ships the lean pick if the conditions match the desk's reading; the desk ships the alt pick if the conditions diverge. The desk does not ship a conditions-conditional pick because the conditions-conditional pick is the place where the desk's reading is thinnest and the reader's contest entry is most exposed.
The three windows are the desk's reading of the contest-entry levers. The three windows are the desk's contract with the reader. The three windows are what the desk can defend with a named source. The three windows are the desk's wall against the false precision.
Predictions desk — readers ask
Are the picks guaranteed?
No. The desk's reading is one input, not a guarantee. Always read the platform's published rules.
How do you pick the captain?
Highest floor first, then highest ceiling. The captain is the pick whose downside is the smallest, not the pick whose upside is the largest.
Why do you disclose the self-audit?
Because a self-audit line is the only honest way to know whether the desk's reading is improving. The miss stays on the page.
Where do I see the desk's accuracy?
The homepage winners block reports a verification band, not a single precision number. We sample 50 matches per window and label the band.