Ledger vs Scoreboard: Auditing Death Overs, Ball-Tracking Chains and the Ten-Match Threshold

The Night Two Screens Told Two Truths

In a recent BPL season I sat at my table in Rangpur with two screens lit: one showing the franchise league's official broadcast graphics, the other a third-party data dashboard. Same bowler, same spell, four death overs. One screen said economy 7.75. The other said 8.17. A gap of 0.42 runs per over looks small until you multiply it across four overs — roughly 1.7 runs, often the difference between defending and losing in the last over.

I wrote both numbers in my notebook and added a question: which one is right? Then a second, more important question: where did the number come from, who entered it, when, and did anyone change it after the match ended?

Since that night I have been circling one problem. Cricket now records the geography of every delivery — line, length, speed, batter position, fielder angle. But how trustworthy is the record itself? And if the record is in question, how much weight does every analysis built on it — including my own — actually carry?

Context: Cricket Data Is Not Built in One Layer

Cricket data is produced in at least three layers, and the three layers have different owners.

The first layer is the scorer. A human sits at the ground and logs each ball: runs, wickets, extras, which batter, which bowler. This is the official scorecard and the legal document. In franchise leagues the scorer is employed by the league, and entries can be amended later — rain, Duckworth-Lewis recalculations, or simple error correction.

The second layer is ball-tracking. Hawk-Eye or an equivalent system derives trajectory, pitch location, bounce, swing and seam movement from camera frames. The system is installed with the broadcaster's money, and the data lands on the broadcaster's servers.

The third layer is enrichment: wagon wheels, pitch maps, control percentage, dot-ball geography, expected-runs models. This layer is almost entirely proprietary, and definitions differ between companies.

Here is my real worry. Layer one and layer two disagreeing is normal, because they measure different things. Layer three disagreeing with layer three is not normal. That night's 7.75 versus 8.17 gap was born exactly there. One provider counted wides as dot balls in the over; the other did not. One treated leg-byes as dots; the other separated them as batter's deliveries.

Both are "true." But both cannot be true at once unless you first state which question you are answering.

My method comes from that problem, and I publish it so readers can audit me. I make no tactical claim without ten matches of data. I fix the format first — T20, ODI or Test. Then the venue: Mirpur's slow surface is not Chattogram's batting-friendly deck. Then the era: a run rate of 8.1 in 2026 is not 8.1 in 2026. Then the phase: powerplay, middle, death. Then the quality of the opposition.

The Burnley thread looked like noise until I sorted it by PPDA. Cricket follows the same rule. A run rate is noise until you sort it by phase and venue. Sorted, it becomes information.

Core Analysis: Baselines Before Verdicts

Baseline Table 1: BPL run rate by phase (my compiled sheet)

| Season | Powerplay (1-6) | Middle (7-16) | Death (17-20) | |---|---|---|---| | 2026 | 6.9 | 7.1 | 8.1 | | 2026 | 7.1 | 7.3 | 8.4 | | 2026 | 7.4 | 7.6 | 8.9 | | 2026 | 7.6 | 7.8 | 9.2 | | 2026 | 7.5 | 7.9 | 9.0 | | 2026 | 7.9 | 8.2 | 9.6 | | 2026 | 7.8 | 8.1 | 9.4 | | 2026 | 8.2 | 8.5 | 10.1 | | 2026 | 8.3 | 8.6 | 10.3 | | 2026 | 8.5 | 8.8 | 10.6 |

I built this table from my own tracking sheet, not from official league records, because official records do not split by phase — and I do not judge a bowler without a phase split. Each match in my sheet carries at least six fields per ball: over number, ball number, bowler, batter, runs, and outcome type. My season sample runs 35 to 46 matches, roughly 8,500 to 11,000 legal deliveries.

The table states a structural truth: death-over run rate rose about 2.5 runs between 2026 and 2026, while powerplay run rate rose only about 1.6. The whole game got faster, but the last four overs got faster fastest. That is not only better bats — it is changed bowling strategy. In 2026 the death-over weapons were the yorker and the slower ball. By 2026 they include the wide yorker, the hard-length cutter, the slow bouncer, and a "not-a-bad-ball" philosophy: prevent boundaries rather than hunt dots.

That is where I part with conventional analysis. We judge death bowlers by economy. But economy is a ratio — runs divided by overs. And overs are no longer filled with dot balls; they are filled with attempts to avoid boundaries.

Table 2: Sources of death-over dot balls (my sheet, 2026-2026, 2,314 dots)

| Dot-ball type | Share | |---|---| | Wide yorker / toe-crushing length | 27% | | Slower ball (knuckle/off-cutter) | 21% | | Hard length, stump to top of stump | 18% | | Slow bouncer | 14% | | Batter error (mis-hit, late swing) | 11% | | Fielding-driven dot (dive, direct hit) | 9% |

This is my most valuable table, because it shows a dot ball is not a dot ball. One dot comes from bowler control; another comes from batter error. The first is repeatable; the second is often luck. Without a ten-match threshold I cannot separate them — because over three matches, luck landing twice in a row looks like skill.

The Ten-Match Audit: One Death Bowler's Case

Take a death bowler's first ten matches of a season (I withhold the name; the point is method, not profile).

Table 3: Bowler-K, death overs, rolling ten matches

| Matches | Overs | Runs | Wkts | Economy | Dot% | |---|---|---|---|---|---| | 1-2 | 6.0 | 41 | 2 | 6.83 | 42% | | 3-4 | 7.0 | 68 | 1 | 9.71 | 29% | | 5-6 | 6.2 | 52 | 3 | 8.21 | 35% | | 7-8 | 7.0 | 74 | 0 | 10.57 | 24% | | 9-10 | 6.1 | 49 | 2 | 7.94 | 39% |

Ten-match economy: 8.65. But the average is meaningless here, because matches 3-4 and 7-8 do not resemble matches 1-2 and 9-10. My question: is the swing about opposition quality, venue, or match state?

Table 4: Stability check — Bowler-K, death economy

| Split | Sample | Economy | Dot% | |---|---|---|---| | Mirpur | 5 matches | 7.40 | 41% | | Chattogram | 3 matches | 10.20 | 26% | | Sylhet | 2 matches | 9.10 | 31% | | vs top-four batting lineups | 4 matches | 10.85 | 23% | | vs the rest | 6 matches | 7.20 | 42% | | Defending | 6 matches | 7.90 | 38% | | Chasing | 4 matches | 9.80 | 27% |

Now the story is clear. Change the venue and 2.8 runs move. Change the opposition and 3.65 runs move. Change the match state and 1.9 runs move. This bowler's death skill is real inside one environment and questionable outside it. Had I written "reliable death bowler" off the 8.65 average, I would have been wrong. Had I written "disaster" off matches 7-8, I would also have been wrong.

I do not write without ten matches. That is not laziness; it is method. And even at ten, I split by venue, opposition and match state, because otherwise the average lies to me.

Precedent Table: Change the Era, Change the Meaning

Table 5: Era-adjusted death-bowling precedents

| Period | League | Baseline death economy | "Good" | "Exceptional" | |---|---|---|---|---| | 2026-2026 | IPL | 8.0 | 7.0 | 6.0 | | 2026-2026 | IPL | 8.8 | 7.8 | 6.8 | | 2026-2026 | IPL | 9.4 | 8.4 | 7.4 | | 2026-2026 | IPL | 10.2 | 9.0 | 8.0 | | 2026-2026 | BPL | 8.4 | 7.4 | 6.4 | | 2026-2026 | BPL | 10.3 | 9.1 | 8.1 |

Mustafizur Rahman took 17 wickets in IPL 2026 and was named Emerging Player — a documented fact. But to understand that season properly you compare it with the 2026 IPL death baseline of 8.8, not the 2026 baseline of 10.2.

This is where I disagree with how transfer markets price death bowlers. Franchises set value from one season's economy. One season is 14 to 18 matches, of which death overs are only 30 to 40. Blend a boundary-heavy venue with a slow one across 40 overs and the resulting number cannot carry a multi-crore bid. Transfer-market news means a timestamp, or it is just a story. Transfer valuation means era adjustment, or it is just the market.

The Ledger Layer: What Blockchain Fixes and What It Does Not

If every ball-by-ball entry were written to a tamper-evident ledger — each entry hash-chained to the previous one, each timestamped — three things change.

First, correction and concealment separate. An amendment stays possible, but the old entry remains beneath the new one. The record shows who changed what, and when. Had that night's 0.42-run gap lived on a chain, I would know exactly which dashboard applied which definition, and when.

Second, smart contracts automate contractual conditions. Franchise deals include death-bowling bonuses — a set number of wickets, an economy below a threshold. Today that arithmetic happens by hand, at season's end, amid argument. A verifiable ledger records the trigger the moment it fires, and both parties read the same number.

Third, fan-level ownership. In 2026 FanCraze partnered with the ICC on cricket digital collectibles, and in 2026-22 Rario worked with Cricket Australia. The core idea is a verifiable ownership record — which fan holds which moment — written to a ledger rather than a central server. For a franchise league the extension is natural: a digital copy of an iconic death over, bundled with that over's verified ball-by-ball data.

And here is my central caution.

Contrarian: A Verified Number Is Not a Correct Number

A blockchain proves the entry was not altered. It does not prove the entry was true.

That sounds philosophical, but it is deeply practical. Return to my two screens. Suppose both dashboards write to a hash chain. I can now learn who wrote 7.75 and who wrote 8.17, and when. I cannot learn which definition answers the cricket question correctly. Whether a wide counts as a dot is not a data problem; it is a definition problem. And definition problems are not solved by chains, they are solved by governance.

Second: a ledger can only record events someone entered. If nobody logs the fielder's diving stop as a dot, the chain will not conjure it back. The ledger protects the integrity of the record, not the completeness of the event. In cricket, what is not entered does not exist — an old data disease that chains do not cure.

Third, and my largest fear: a hash-sealed number looks more credible than an unverified one, even when it is wrong. If a selector or franchise manager stops questioning a number because it carries a "verified" tag, we have audit theatre — the appearance of transparency without the substance of verification.

Fourth, the process itself creates a market. Whoever runs the ledger decides which balls get entered, which definitions govern, who can read. That is not decentralisation; it is a transfer of power from broadcaster to company. Cricket has seen this shift before — when ball-tracking arrived in the 2000s we asked who would install the cameras. The answer was whoever paid.

Modric ran twelve kilometres, but the map showed where the game turned. A hash-sealed 8.17 likewise shows the number was not altered — it does not show which question that 8.17 answers.

Method Note: My Four Verification Steps

Step one: identify format, venue, season and phase. Without the format, economy comparisons are meaningless — a T20 8.0 and an ODI 5.2 are never the same thing.

Step two: secure a minimum ten-match sample. Below ten I write "observation," never "verdict." I set the threshold in advance, not after seeing the data, or sample selection becomes biased.

Step three: stability checks. Split by venue, opposition and match state. If any split falls below a sample of five, I label it "insufficient sample."

Step four: era adjustment. Place the number in a precedent table and see how far it sits above or below its own era's baseline.

Outside these four steps I make no tactical claim. It is slow, but it is citable. And this stubbornness costs me entertaining stories — in exchange it saves me from a great many errors.

What I Will Watch Next

Three things. First, a published data dictionary at league level — an official list of definitions covering wides, leg-byes and dot balls. Without it, every data argument stays unresolved. Second, when the league death-over baseline crosses 11. My sheet points that way, and the day it happens, "8.5 economy" stops defining a good bowler. Third, whether a verifiable ledger arrives in cricket, and whether it actually solves the definition problem. Because cricket's real crisis was never a shortage of data — it is agreeing on which number we all believe.