Caps 217 and 218: Cricket's Real Crisis Is Data Built on Names

Cricket's most expensive data error is not the one that is wrong

Cricket's most expensive data error is not the one that is wrong. It is the one that is perfectly correct and belongs to somebody else.

A Wisden quiz asks a simple question: what name do India's men's Test caps 217 and 218 share? The answer is withheld. What matters far more than the question is the rule the quiz is built on — exact matches only. One letter off and the pair is disqualified. That is why the quiz explicitly excludes four pairs: Mehidy Hasan against Mahedi Hasan, Claire Taylor against Clare Taylor, Abdul Razzaq against Abdur Razzak, and Mohammad Sami against Mohammed Shami.

Those four exclusions are not a quizmaster's random choices. They are four pairs real people have really merged.

In a Chennai startup I learned that empty tables make loud enemies. Through IPL 2026 I logged all sixty matches ball by ball — 14,400 deliveries. The database started empty; the first real finding arrived as a fight. After the final, where Mumbai Indians beat Rising Pune Supergiant by one run in Hyderabad on 21 May 2026, I published the argument that the margin was set forty overs earlier. My bigger problem was not that anyone disputed the model. It was that a query had merged two different players into a single string and I did not notice for three days.

I repeat it because it questions the foundation of the whole building: cricket's records rest on names, and names are not identifiers.

Context: a quiz, a numbering system, and a rule

This Wisden product carries thirteen information points and not one of them is a match result, a transfer, or a governance decision. The honest professional response is not to manufacture depth that is not there. It is to mark where information is absent and extract the substance that genuinely exists — player-identity disambiguation and record-keeping integrity.

The mechanism first. The BCCI assigns a sequential Test cap number. Cap one is Lala Amarnath, 2026. Every new Test debutant takes the next number. It is never reused, never retired, and never applied to any other format.

A cap number is what a name could never be: a permanent, unique, sequence-based identity.

The structural inference is strong and it is the most reliable thing in this entire source. Caps 217 and 218, being adjacent, almost certainly debuted in the same Test or in consecutive Tests. Sequential debut numbering permits nothing else. That narrows the answer space to two names in one match.

And here I will say what a data analyst says and a broadcast panellist does not: I am not going to guess. Live xG taught me to watch the second screen and the room. A model is only as honest as the questions it refuses. I do not hold a verified answer, and guessing without verification is precisely the error this article is written about — the name looks right, so you move on. When the answer is published, the real question that day will be whether the record that produced it is clean.

Three worked examples appear in the quiz, each from a different world. Two Ollie Robinsons, one of whom played Test cricket for England. New Zealand's Daryl Mitchell, subject of Mitchell Santner's joke about phoning the Worcestershire captain of the same name. And Zaheer Abbas against Syed Kirmani, the most intricate case in the set.

Then the rule. Exact match, or nothing. That rule is not a sporting rule. It is a data-classification rule, and it is the most transferable element in the whole quiz — despite sitting inside an engagement product.

Name classification is a real problem in cricket, and it is not new. English county scorebooks have carried it since the 1950s; the subcontinent has carried it since Mohammad, Muhammad and Mohammed began entering the same register. What has changed is the speed. Once a wrong name lived for a day in a single newspaper column. Now it reaches three thousand fantasy teams in fifteen minutes.

Sixty-four matches, one feed, and a growing list of doubts — that was my condition when I was only watching. When I am also holding the data, one item joins the list: does this row belong to the man whose name is on it?

Core: six pairs, one economy rate, and the question nobody asks

Razzaq and Razzak: when bowling type is a field, not an adjective

Abdul Razzaq bowled right-arm medium-fast for Pakistan and batted lower order. Abdur Razzak bowled left-arm orthodox for Bangladesh. One vowel of difference in the name. A completely different job.

This is where analysis begins, because the error here is not about a name. It is about a bowling type, and bowling type is a structured field in cricket, not a commentator's adjective. Pool these two and you do not spoil a batting record; you spoil the matchup baseline underneath it.

Consider what actually moves. The economy rate of a left-arm spinner against right-handed openers is a different universe from that of a right-arm medium-pacer. Across two hundred matches, if forty belong to the wrong bowling type, your baseline slides quietly. Nobody gets suspicious because the name looks right.

Cricket's most dangerous data error is the one a name certifies as valid.

That lesson cost me money in Chennai. When I built the death-overs exposure metric in 2026, my first question was not whether the metric was right. It was whether every ball sat under the correct bowler. Hand-checking showed the errors were not scattered. They clustered exactly where names looked alike — and those were exactly the places nobody audits.

Claire and Clare: the pair no filter can break

Of the four excluded pairs, the most consequential is not obvious. The hardest one is. Claire Taylor and Clare Taylor.

Both England women's internationals. Same country. Same gender. Overlapping decades. One letter of difference in spelling, and a difference that is invisible in speech.

Clare Taylor (born 2026) played for England from 2026 to 2026 as a seam bowler and also represented England at football. Claire Taylor (born 2026) played from 2026 to 2026 as a batter and became the first woman named among Wisden's Five Cricketers of the Year in 2026.

Notice which filters fail. Country? Both England. Gender? Both women. Format? Both played ODIs. Era? Both active in the 2000s. Role? You can separate a bowler from a batter, but only as reliably as a commentary-box adjective, which is to say not reliably at all.

What remains is date of birth and exact spelling. When two identities are demographically indistinguishable, the only dependable basis for separation is a date of birth — a field never spoken aloud on air.

This is why the problem is larger in women's cricket than men's, and why so few people say so. In men's databases, same-name players from one country are usually separated by decades. In women's cricket the population is small, careers are shorter, and collisions within one generation are likelier. Coverage is thinner. When coverage is thin, nobody catches the error.

I spent one season in a thirty-seat Chepauk press box and I was the only woman in it. The biggest thing I learned there was not about cricket. It was that whatever everybody in the room knows is exactly what nobody checks.

Sami and Shami: the highest-traffic confusion

Mohammad Sami, Pakistan, right-arm fast, born 2026, career roughly 2026 to 2026. Mohammed Shami, India, right-arm fast, born 2026, Test debut 2026.

This is the most searched pair on the list, and the reason is obvious. The error does not begin with spelling. It begins with a name that sounds almost identical attached to a job that is almost identical. Both fast. Both take the new ball. Both bat with some comfort low down. Search 'Mohammad Sami fast bowler' and half the results belong to the other man.

Now watch what merging them does. Sami's international career opens in 2026. Shami's opens in 2026 and runs to the present. Pool the rows and you get a career stretching 2026 to 2026 — twenty-three years for a fast bowler. That is biologically impossible. And the error survives anyway, because nobody audits a long career when the name looks fine.

One further property makes this pair worse than it looks: the confusion peaks precisely while the two countries are playing each other. Maximum attention, minimum verification.

Mehidy and Mahedi: when country and role both fail

Mehidy Hasan Miraz, Bangladesh off-spinner, born 2026, Test debut 2026. Mahedi Hasan, Bangladesh off-break bowler and all-rounder, born 2026, international cricket from 2026.

Same country. Same bowling type. Same approximate era. Same format overlap. One vowel between them.

Here even the country filter and the bowling-type filter fail. A query for 'Bangladesh off-spinner' returns two answers, and knowing the Test bowling average settles nothing.

This pair is where translation matters. My previous work was in football analytics, on live xG. Football tools look seductive in cricket. But football's logic forgives homogeneous errors, because thousands of passes and dozens of competitors push a number back toward truth. Cricket does not. In a low-scoring sport, forty deliveries from the wrong bowling type do not roll back off your phase model. They leave a mark, and sometimes they flip the direction.

Cricket models run on small samples and large assumptions; an identity error there is not a detail, it is a foundational mistake.

Two Ollies, two Daryls: overlap is what makes the joke work

Ollie Robinson (born 2026), Sussex right-arm fast, Test debut at Lord's against New Zealand in June 2026. Ollie Robinson (born 2026), Durham and later Kent wicketkeeper-batter, domestic cricket.

The quiz says only one of them has played Test cricket. That may be true and still be a claim with an expiry date. 'Only one has played a Test' can be true today, false in two years, and true again in four, when the first retires and the second debuts.

The most dangerous row in any database is the one marked 'still true'.

The Daryl Mitchells are safer only because their formats differ. Daryl Mitchell (born 2026), Worcestershire captain, County Championship, 2026 to 2026. Daryl Mitchell (born 2026), New Zealand middle order, from 2026.

Santner's joke about phoning the Worcestershire captain lands because the two men were near-contemporaries in the same professional ecosystem. Separate them by four decades and there is no joke.

A structural rule follows, and it is the most useful in this whole subject: the dangerous name collisions are not the ones separated by time. They are the ones that overlap. A distant player is separated by a date. A contemporary is not, because there is no clean dividing line to find.

Zaheer Abbas and Kirmani: when the player changes his own identifier

The most instructive pair in the set.

Pakistan's Zaheer Abbas was born Syed Zaheer Abbas Kirmani. He preferred to use his middle names. India's wicketkeeper of the same era was Syed Kirmani. The result: two international cricketers, two countries, and a name that sounds like two halves of one man.

This case differs from every other because the error was not made by a database. It was made by the player, and it was entirely his right. That is why historical records need an alias field rather than a single name string: some identities are created by the player, some by the scorebook, and some by the media.

I once reconciled a domestic tournament's match sheets against the original scorebooks. Four of nineteen names appeared in different spellings. Two were simple misspellings. Two were entirely different names that had mutated from one source to the next. That day I told my team that the bigger our database grew, the faster our history was changing — and that the speed was under nobody's control.

The cap number: what the name could not do

India's Test cap number is an identity system that has worked for nearly a century. Cap one, Lala Amarnath, 2026. Next number for every new Test player. Unique, permanent, sequential, irreversible. Cap 217 will be the same man for as long as the record exists.

Compare the system we actually use daily. A name is not unique, not permanent, and certainly not sequential. It can change. Its spelling can change. It can be translated. It can belong to two people at once, and that fact lives in no register.

So why not use the system that works? Because it covers one country, one gender, one format. It tells you a man played a Test for India and nothing else. It does not tell you his IPL franchise, his domestic side, or who his female counterpart is. The moment you leave one format, you are back on names — which is why every cross-format fact in cricket travels by name.

Cricket has had the right identity model since the 1930s, in one small corner, while the rest of the game runs on names.

Contrarian: where I disagree with myself

State the base rate first. Most of the time, in most databases, most names are handled correctly. The people who build cricket data are not fools. They know Sami and Shami are different men. This is a minority problem — but minority problems are what determine a public database's credibility, the way two delayed trains determine a railway's reputation.

The first objection to my own argument is that I am inflating an entertainment quiz into an infrastructure crisis. Half true. Serious analysis never catches this problem, because serious analysis never works at row level. When I analyse death overs I work on aggregates — fifty bowlers averaged together. One wrong row drowns. A quiz, by contrast, must decide whether this name belongs to this man. It performs the strictest audit in the sport. Cricket's tightest identity check happens in trivia and its loosest in analysis.

The second objection is more uncomfortable: the problem is not in the database. It is in the delivery layer. Nobody makes a commentary box bad. Its conditions do. Live, no feed, one ball to react, a name to say, usually read off a screen glanced at three seconds ago. I said from a distance that they make errors. Sitting in the room, I understood they had one ball and I had six months.

If you want to help that chair, you have to help it at the instant of delivery. A screen layer showing birth date, team, bowling hand and cap number alongside a name is technically possible today and cheaper than an extra camera. The day it exists, identity errors fall by half — and that reduction comes from an interface, not an analysis.

The third point is the largest. Cricket's correction culture was built for numbers, not identities. A score printed as ninety instead of a hundred produces a correction the next day. A name swapped for another name usually produces nothing, because the error has no clear owner. A wrong score feels like a lost match. A wrong name feels like a misplaced person, and no data system has priced that.

That is why the problem persists. Its cost is invisible. No sponsor funds it. No franchise hires for it. And that is why, eight years later, my first database still shows the gap it had on day one — the names looked right, so nobody checked them.

Takeaway: the quiz's answer is not the next step

I will not give you the names at caps 217 and 218. I do not hold a verified answer, and my exposure there is greater than the value of being first. The numbering structure gives one dependable fact: both almost certainly debuted together or in consecutive Tests. Everything beyond that is inference, and inference is not what this piece is about. It is about how many people will ask, on the day the answer is published, whether the record that produced it is clean.

I stopped counting trophies and started counting habits. The first number that came back: the rate of identity discrepancies in my own database was three times what I had assumed. That is not a survey. It is the unsentimental result of auditing my own system — and when you find that in your own system, you stop asking other people's systems angry questions and start asking them precise ones.

The next time you open a scorecard and pause on a name, look for one thing: is the birth date, the team and the format printed beside it. If even one is missing, what you are reading is not data. It is a claim.

And the next time someone tells you cricket's statistics have improved, ask by what measure. Because in this country the statistics improved. The data did not. The distance between those two things is exactly the distance between the two Ollie Robinsons — one of whom played Test cricket, while the other still stands in the same scorecard wearing somebody else's name.