Caps 217 and 218: Two Names, One Hole, and the Question the Scorecard Never Asks

Hook

February 2026. The Indian Super League was being played across three empty stadiums in Goa, and my headphones carried touchline microphones instead of a crowd. I was charting pressing triggers, logging 38 coaching instructions a game, every entry going into the same template: minute, broadcast source, player, action. Two weeks later, writing from that file, I found a delivery logged under the wrong name. The clip showed one bowler; my shorthand had written another. I recomputed the economy for that phase and the number moved.

The clipboard said one thing; the 40 metres said another — and the difference was a single name.

This week Wisden's cricket quiz posed a question: what name do India's men's Test caps 217 and 218 share? At first glance it is a knowledge question — you either remember or you don't. At second glance it is the exact failure that split one bowler into two in my own file. The caps 217 and 218 question is not a question about names. It is a question about identity resolution, and that is precisely where cricket's public data layer is thinnest.

I heard the game before I saw it, and the audio told a different story. This time the story's protagonist was a spelling.

Context

This is not reporting. It is an engagement product, and it should be read as one. The source carries thirteen information points and not a single match result, average, transfer or governance decision. What it does carry is a rule, three worked examples, four explicit exclusions, and a question whose answer the compiler deliberately withholds.

The operative rule is that namesakes must be exact spelling matches. That is why four pairs are struck out in plain terms — Mehidy Hasan against Mahedi Hasan, Claire Taylor against Clare Taylor, Abdul Razzaq against Abdur Razzak, Mohammad Sami against Mohammed Shami. That is not a cricket rule; it is a data-classification rule, and it is the most transferable thing in the entire piece.

The quiz's scope is deliberately messy. It crosses Test, ODI, T20, women's internationals and the County Championship in one bag. Mehidy Hasan and Mahedi Hasan are both Bangladesh spin-bowling all-rounders. Claire Taylor and Clare Taylor were both England women's internationals. Abdul Razzaq bowled right-arm pace for Pakistan; Abdur Razzak bowls left-arm spin for Bangladesh. Mohammad Sami is Pakistan; Mohammed Shami is India.

Five further quiz links sit inside the piece — on the Sharmas in international cricket, on Test-playing Smiths, and three others. That is not incidental. Wisden is running the surname as a franchise, which means the publisher assumes namesake confusion is frequent enough among its readers to sustain a series. For a publisher that is a traffic strategy. For a data writer it is a warning label.

Now the part the quiz does not show. India's Test cap numbers are sequential. Every debutant takes the next number. This is a BCCI and record-keeper convention, not a cricketing law. The implication is direct: only one debut separates cap 217 from cap 218. The pair almost certainly debuted in the same Test, or in consecutive Tests — a high-confidence inference that shrinks the answer space to near-zero verification cost. Open the cap ledger, read two lines, done.

The source does not publish the answer. I will not either, and that restraint is the most honest part of this exercise — because the moment I write a name without checking the ledger, I commit the exact error I am describing.

Core: four exclusions, four different failures

Read the four excluded pairs not as a list but as four distinct failure classes, each with its own data cost.

Class one: same country, same role, same format. Mehidy Hasan (Miraz) and Mahedi Hasan are both Bangladeshi, both spin-bowling all-rounders, both active in roughly the same decade. Team filters fail. Role filters fail. Format filters fail. What survives is spelling and date of birth. This is the class where no convenient filter saves you, which makes it the most dangerous. Merge these two rows and your spin-economy baseline is permanently contaminated, silently.

Class two: same country, same gender, overlapping career windows. Claire Taylor and Clare Taylor, both England women's internationals. Here the team filter and the gender filter are both useless because both players sit inside the same team and the same gender. This is the highest-risk pair in the set for a technical reason, not an emotional one: when every social filter for disambiguation fails, only two fields remain — spelling and date of birth. That is why date of birth belongs in the mandatory key, not the optional columns.

Class three: near-identical name, entirely different function. Abdul Razzaq bowled right-arm pace; Abdur Razzak bowls left-arm spin. This is the most analytically material pair in the quiz, because the error stops being a naming error and becomes a bowling-type error. Economy rate, matchups, phase usage, new-ball versus middle-overs deployment — all of it shifts. Pool these two and you are building the profile of a bowler who is simultaneously fast and slow, which is to say no bowler at all. In my experience this kind of contamination surfaces latest, because the number does not look wrong; it looks slightly odd, and slightly odd numbers get waved through.

Class four: the highest-traffic confusion of its era. Mohammad Sami and Mohammed Shami. Different countries, different decades, similar craft. I am calling this the most frequently mis-merged pair of the 2010s as an inference, not a measurement — there is no public count of how often a name gets mistyped. That absence is the real shape of the problem: errors happen, but errors leave no ledger.

Three worked examples that say more than the rule

First, the two Ollie Robinsons. One born 2026, right-arm pace for Sussex and England, a Test cricketer. One born 2026, wicketkeeper-batter, Durham and Kent, English domestic cricket. The interesting axis is not role but system: both inside the same domestic structure at roughly the same time. When two players share a country, a domestic system and a decade, no convenient database key separates them except date of birth. I will also flag that the source's claim that only one of them has played Test cricket is time-sensitive; I could not verify it today, so it should be read as the source's statement rather than as settled fact.

Second, Mitchell Santner and Daryl Mitchell. Santner made a joke about his New Zealand teammate Daryl Mitchell, invoking the Worcestershire captain of the same name. The joke is the evidence. Comedy only lands when both men occupy the same professional ecosystem at the same time; namesakes separated by decades carry no comic force. Santner's joke is quietly reporting a technical fact: namesake confusion is sharpest among contemporaries, which is exactly what makes it dangerous for data.

Third, Zaheer Abbas and Syed Kirmani. The least discussed and most interesting case. Zaheer Abbas's full name was Syed Zaheer Abbas Kirmani, and he chose to use his middle names. The identity was not changed; it was selected. That is a player-created alias, not a database-created one, and it is why historical records need an alias field rather than a single name string.

What happens to the numbers when the name slips

Here I have to cap my own instinct. I could bury three models and two datasets in one paragraph, but only one consequence changes the conclusion: the moment two players collapse into one row, every rate on that row becomes an average of two careers. An average is not a lie; it simply answers a question nobody asked.

The sharpest version of this appears wherever a name functions as a key rather than a description — auction lists, franchise shortlists, fantasy rosters, endorsement attribution. The Sami-Shami gap is an embarrassment in the commentary box; on a scouting sheet it is a wrong selection. That is inference, not a claim from the article, but it comes from the same place my clip-reference template came from: the habit of trusting a hand-typed name.

One boundary I will not cross. The source contains zero batting averages, strike rates, economy rates or splits for any named player. So I will offer no technique judgment at all. Filling what the source lacks with guesses inferred from a quiz is the same sin this article exists to describe.

Contrarian: the quiz is not the problem, the shorthand is

The easy move is to blame the database. The database is not guilty. Professional archives have stored distinct player identities for decades. The failure happens at the handoff, when data leaves the system and enters a human hand at three in the morning, in a spreadsheet, typed by a tired charter. My own file is a witness to that moment.

A second inversion: the quiz's answer is the least valuable output of this whole exercise. I will admit that my first thought on reading the question was to solve it fast and post it. Being right early is intoxicating, and at 26 that intoxication hits hardest. But solving leaves you with a name and nothing else. Understanding the rule leaves you with a verification method that scales to the next hundred names.

A third inversion, aimed at my own systems optimism: the fix is not grand. No cricket ORCID will descend from the sky. The fix is boring — alias fields, date of birth as a mandatory key, role and bowling type stored as separate structured fields, and every record partitioned by format. Pipeline optimists forget that the weakest joint in any pipeline is its last one, where it meets a person.

And a fourth, the most uncomfortable: by withholding the answer, this quiz is practising good hygiene. Publish the answer and the reader stops verifying and starts remembering. By leaving it out, Wisden forces a trip to the cap ledger — and the trip to the ledger is the cure. A quiz that refuses to give the answer is better than one that gives it and thereby certifies its own error.

Takeaway

I am entering this in my prediction ledger with a date. If identity disambiguation in cricket data does not appear as an explicit player-identity index on at least one major public platform before December 2027, then my claim — that this is now an infrastructure problem rather than a quiz problem — will have been falsified.

The check is cheap. Next time a namesake quiz appears, look for a date-of-birth column. If it is there, the publisher has conceded the problem. If it is not, the publisher is still trusting the reader's memory — and memory is the thing that split one bowler into two in my file.

Every formation hides a question the next match will ask. Every name hides one the next scorecard will ask. The question is not what they were called. The question is how we will know we wrote the right one.