Cricket's Data Chain: Scorebook, Ball-Tracking and Ledger — Who Verifies Cricket's Truth?

Introduction: Two Scorecards at Half Past Midnight

At 12:40 a.m. last Friday I had two scorecards for the same innings open side by side. On one, a fast bowler's death-over economy (overs 17–20) read 9.25. On the other, 8.75. The gap was half a run — exactly one delivery's worth of fate, a ball recorded as a wide in one place and a leg-bye in the other. Both cards carried the same match ID and both claimed the same ball-by-ball data set. Yet there were two truths on two screens.

In 2026, at the ICC Trophy match between Bangladesh and Kenya, I sat in a radio box with a scorebook and a pencil. Today I sit with spreadsheets, GPS logs, ball-tracking confidence intervals and Python scripts. The tools changed; the problem did not. Cricket's biggest data problem is not the wrong number — it is that we have no chain that proves which number is genuine.

This is where blockchain enters the conversation. Not as crypto enthusiasm, but as a straightforward question: if every data point of a match were time-stamped and append-only — where a correction is a new entry, not a hidden overwrite — would I still be sitting at half past midnight comparing two scorecards?

The easy answer is yes, everything would be fixed. That easy answer is wrong. This piece is an attempt to explain why.

---

Context: How Long Cricket's Data Supply Chain Actually Is

As fans, when we say "data" we mean the scorecard. But after a ball is bowled, at least six hands touch it before it reaches a scorecard.

Layer one: the venue scorer, one or two people writing ball by ball by hand or on a tablet. Layer two: ball-tracking systems — Hawk-Eye, first used in cricket on television at Lord's in 2026 for England v Pakistan, primarily as an LBW and review aid. Layer three: broadcast graphics teams, who often run their own live coding and do not always reconcile with the venue scorer. Layer four: the official statistics partner. Layer five: media re-publishers. Layer six: fans, prediction games, fantasy apps and betting markets, all of which build their own arithmetic on top.

Notice that bad data does not flow downward. It flows sideways. The scorer is one source, the graphics team another, a fantasy app a third API. Nobody depends on anybody, nobody verifies anybody. There is no chain — only parallel branches.

In 2026 I learned this lesson in a different sport: the Burnley thread looked like noise until I sorted it by PPDA. A 12.1 PPDA and 38 percent possession read as "passive football" to most observers; sorted properly, it looked like controlled passivity. Cricket's data chain behaves the same way: unreadable as noise, structural once sorted.

But the translation must be cricket-native, or we force football tools onto cricket. The cricket equivalent of PPDA is a deliveries-per-pressure index: how quickly a bowling side manufactures stress, and where that stress originates. Paired with economy it tells a story; economy alone is noise.

---

Core Analysis: Where the Chain Cracks

Baseline First, Claim Later

Before judging any bowler, innings or "trend", four baselines sit on my desk: format, venue and pitch type, era (run-rate adjusted), and phase. Cricket rarely rewards home-away splits in the football sense; phase-based economy deviation does the work instead.

| Phase | Baseline (T20, batting-friendly) | Tolerance | Measured | |---|---|---|---| | Powerplay 1–6 | Economy 7.60–8.10 | ±0.40 | New-ball control % | | Middle 7–15 | Economy 7.10–7.80 | ±0.35 | Dot-ball % | | Death 16–20 | Economy 9.40–10.20 | ±0.70 | Yorker share, slow-ball mix |

Three things matter. First, death-over tolerance is nearly double the powerplay's, because variance is higher — one missed yorker changes a series rate. Analysts who ignore that tolerance and call small swings "form" are simply naming noise. Second, baselines are era-dependent: 2026 death economy and 2026 death economy are different animals. Third, a baseline is itself a claim, and must be auditable — sample size and filters written beneath every table, or the table stops being data and becomes decoration.

Seven Seams Where Discrepancies Enter

Seam one: event definition. Dropped catch or difficult chance? Wide or bye? Human calls, human disagreement. Seam two: attribution. Run-out credit to fielder or keeper? A fielding error to the fielder or to the captain's placement? Seam three: coordinate data. Ball-tracking gives pitch point, impact point, trajectory — but camera angle, frame rate and confidence thresholds shift a ball's pitch point by a metre. Seam four: phase boundaries. When does the death start? 16, or "the last spell of the frontline bowlers"? Three platforms, three answers. Seam five: context layer. Dew, wind, innings break, DLS — stored elsewhere, absent from the scorecard. Seam six: publication and re-publication. An error, once printed, gets quietly updated later while the original leaves no scar. Seam seven: citation decay. A number tweeted without context flows into a fantasy app, then into a preview, and three hands later is accepted as fact with no traceable source.

| Layer | Who records | Common error | Verification today | |---|---|---|---| | Venue scoring | On-site scorer | Definition drift | Match referee check | | Ball-tracking | Tech provider | Thresholds, calibration | No independent audit | | Broadcast coding | Live graphics team | Speed over supervision | Internal review | | Official stats | Data partner | Vendor-dependent interpretation | Contract-limited | | Media | Editorial desk | Citation decay | Source lines, often dropped | | Consumer APIs | Fantasy/prediction | Broken pipelines | Almost none |

Four of six rows on the right read "almost none". That is the real worry: not a lack of technology, a lack of accountability.

The Ten-Match Threshold: Not Sacred, But Necessary

My rule since 2026: no trend claim without at least ten matches. The reason is statistical, not moral — small samples bury signal under variance. But ten is not magic. It is a pre-registered threshold, declared before I look, otherwise I will trim to taste afterwards. Condition-specific variation is legal if announced: spin death-bowling variance is higher, so my minimum there is twelve; pace powerplay control is steadier, so ten suffices.

A worked case: a bowler's death economy reads 6.5 across three matches after 10.8 across the previous seven. Media calls it a death specialist. My log calls the ten-match aggregate 9.8 — inside baseline. Nothing new was created; noise was rearranged.

What a Ledger Gives — And What It Does Not

Blockchain offers three things: append-only records, time-stamping, hash-linking. Each has a cricket use. Append-only means a scorer switching a wide to a leg-bye leaves both entries in history — no scar-free correction, which is the biggest gap in cricket data governance today.

In practice, cricket's ledger experiments have mostly been commercial. In November 2026, the ICC announced a multi-year digital collectibles partnership with FanCraze, turning cricket moments into ownable digital assets. Several franchises and governing bodies have trialled fan tokens, blockchain ticketing and digital memorabilia.

Those matter, but be precise: a ledger of ownership and a ledger of truth are different things. An NFT proves who owns a moment; it does not prove what the ball did. The first is a cryptography problem; the second is an epistemic one.

Which brings back the Modric lesson: Modric ran twelve kilometres, but the map showed where the game turned. Distance was verifiable; interpretation was not. Cricket's ledger will do the same — verify ownership and time. Interpretation stays with coaches, analysts and writers.

Precedent Table: Technology's Promise vs Reality

| Technology | Era | Promise | Outcome | |---|---|---|---| | Ball-tracking (Hawk-Eye debut) | From 2026 | Decision accuracy | Fewer howlers, new umpire's-call debates | | DRS | From 2026 (Colombo, SL v India) | Definitive justice | Clear errors cut, marginal debates up | | Live GPS / player tracking | 2010s | Fitness control | More data, lagging interpretation | | Sensor-ball trials | 2010s | Fully-instrumented ball | Stuck in limited trials | | Digital collectibles / fan tokens | From 2026 | Engagement and ownership | Ownership clear, truth not |

What the table omits matters most: every technology entered the decision layer, never the verification layer. Era-adjust these rows or the table manufactures false equivalence — 2026 Hawk-Eye is not 2026 tracking.

---

Contrarian: An Immutable Ledger Makes Bad Data Permanently Bad

Three traps. Garbage in, permanently garbage out: cryptography proves nobody tampered; it does not prove anyone was right. Who runs the nodes: cricket's power is centralised; a decentralised chain run by the powerful is a centralised ledger with better marketing. Privacy versus permanence: player fitness, injury history and biomechanics deserve a right to be forgotten that an immutable public ledger denies.

And there is an inverted problem: cricket's data crisis may be hoarding, not theft. Rights deals and tracking ownership lock raw delivery-level data behind walls. We get processed summaries, not raw truth. Building tamper-proofing for data nobody releases is solving the wrong problem. Also remember correlation is not causation — transparent books do not automatically mean honest books. Technology is architecture; governance is habit. Habits are harder to change.

---

Takeaway: Three Signals for the Next Cycle

First: whether anyone publishes raw delivery-level data openly, even via API. Second: whether correction logs become visible — where is the old entry, when was it changed, why. Third: whether player-side consent rights appear in data governance frameworks; that is the real benchmark.

One thread ties them: the power to prove and the ownership of data should never be the same hand. Blockchain can technically make that easier, and politically make it harder.

For readers: next time you see an innings statistic, ask how many hands the ball passed through and how many files recorded it. The smaller the number, the greater the trust. And whenever a "new star discovered" headline appears, look for the data's birth date and sample size. Like economy rates, sources and timestamps never lie — they are simply left out.

---

Method Note

Trend claims require a minimum of ten matches (twelve for spin-heavy death-bowling samples), pre-registered. Baselines are built on format + venue type + era adjustment + phase. Injury-affected innings, dead rubbers and DLS-affected innings are filtered separately. Data sources: ball-by-ball scorecards, broadcast logs and my own field logs, reconciled before use. Tables before arguments; ten matches before verdicts.