The Scorecard That Says N/A: The Questions Cricket's Data System Never Bothered to Ask

Last week a data pipeline came back empty. Twelve fields — title, source, format, teams, players, time sensitivity, source quality. Against all twelve was the same answer: N/A. Nobody had made a mistake. The scraper had run. The parser had run. The system had simply told the truth: it had nothing.

The emptiness felt familiar. Nine years ago, at a digital startup in Mumbai, I built my first xG model for a league where ball-tracking did not exist as a concept. We had a scorecard, a few match reports, and nothing else. At the time I thought we were collecting data. Later I understood we were recording which questions the system had never bothered to ask. Absence is data too. It just needs the right lens.

In Indian cricket this absence is not in one place. It is a staircase, and the thickness of measurement changes at every step. At the top sits the competition where eight cameras are fixed on every ball. At the bottom sit the matches where the only trustworthy number is the scorecard total. The gap between them is not a technical accident — it is a map of money.

Where the broadcast money is, the measurement is; and where the measurement is, analysis becomes possible.

At the top is the IPL. Hawk-Eye, release point, seam movement, footwork, shot direction, catch probability — this is data that lets you ask questions. When I watch an IPL innings, I have grounds to say which over the bowler changed his line and which over the batter took a risk. That information never emerges from a scorecard. The scorecard says fourteen runs came in the 17th over. Ball-tracking says those fourteen came off three balls on the wrong line, and that the bowler's real weapon was the yorker, which he did not bowl once that over.

One step down sit India's home international series. The tracking here is good, but not complete. Data from a home series is dense; on an away tour the same team suddenly goes thin, because the host board's broadcast structure is different. The direct consequence is that when you compare an Indian batter's home and away record, you are comparing two datasets of different quality — and this is almost never written down.

Then comes the Ranji Trophy. Thirty-eight teams. On many state grounds there is no ball-tracking facility. What survives is the scorecard — runs, wickets, overs, catches, and occasionally a local report or two. I have spent the last few seasons trying to assemble Ranji data, and every time I hit the same wall: you can tell how many overs a spinner bowled in the fourth innings, but not how many balls he turned, how many he kept straight, or how often the batter played the wrong shot.

At the bottom is women's domestic cricket. The Women's Premier League began in March 2026 with five teams, and it changed a great deal in terms of visibility. But the historical record of domestic women's ODI and T20 cricket is so thin that you cannot write the story of a player's five-year development. When a young woman batter has a brilliant season, I have no basis to say whether that is her true level or a good month. This is not a shortage of information. It is a shortage of structure.

This staircase did not build itself. IPL media rights for the 2026-27 cycle sold for roughly ₹48,390 crore. That single number explains why ball-tracking exists in the IPL and not in the Ranji Trophy. Tracking technology is not cheap, and installing it at every ground where not one match of a thirty-eight-team competition is broadcast does not fit any commercial logic.

India's share of the ICC's 2026-2027 revenue model is about 38.5 per cent. That money is the single largest current in the global cricket economy. The question is how it reaches the bottom. My observation is blunt: it reaches the international calendar, the broadcast rights and franchise valuations, but it stops well before the measurement systems on domestic grounds. Ranji does not get the tracking the IPL gets, and that gap belongs to the structure, not to any one season.

Now back to that empty pipeline. What do twelve N/As mean? The most direct answer is that this is a technical failure — a scrape or a parse broke somewhere, and the article that actually existed never reached the system. That matters, because the article may have contained important information on injury, controversy or money that is now invisible.

But the more interesting question is the second one. If that article had been about a Ranji Trophy match, would the pipeline have found enough data to fill itself? My answer is: probably not. And that is the real story. Cricket's biggest data problem is not that a scraper broke. It is that after a full year of cricket, the country's oldest and largest domestic tournament still does not produce the kind of data on which deep analysis can stand.

This is where the actual work begins, and the work is building proxies. When you have no direct measurement, you measure what is available and what is connected to the underlying behaviour. If I want to know how dangerous a domestic spinner is in the fourth innings, I cannot measure turn or drift. But I can measure this: his economy in the fourth innings, his share of his team's overs, and the opposition's win rate in those matches. Together these form a picture that is less precise than direct measurement but does not leave you entirely in the dark.

The condition is non-negotiable: a proxy has to be called a proxy. The moment you present an estimate as a direct measurement, your analysis stops being trustworthy. I have made that mistake myself and watched others make it. Once you have called a spinner's economy an indicator of turn, the reader believes it, and then you start making bigger claims on top of it. It is a debt that compounds.

Metrics do not describe outcomes; they describe choices. I wrote that line for football, and in cricket it is even truer. Strike rate never told me who would win. It told me which team chose to breathe in the middle overs — who deferred risk, who changed the batter, who changed the bowling. When I watch an IPL match, my eye is less on the scoreboard and more on the decisions: who saved his best bowler for which over, who held the yorker back, who set an attacking field to build pressure.

The Impact Player rule, introduced in the 2026 IPL season, is the cleanest example. It does not simply add a player; it rewrites the selection architecture. A team can now buy a sixth bowler, give him a few overs, then bring in a batter for the same slot. When I see a team's Impact Player choice, I treat it as the purchase of a matchup. It is not a statistic; it is a strategic decision — and that is what I want to cover.

The auction price gets stuck exactly here. In the first WPL auction, Royal Challengers Bangalore bought Smriti Mandhana for ₹3.4 crore. That is a big number and it made headlines. It does not tell you how good a batter she is. It tells you how scarce an established, broadcast-ready Indian opening batter was in that market at that moment, and how much a new league was willing to pay for an identity.

A high auction price is a commercial signal, not proof of sporting quality. Whenever an auction report lands, the first thing I check is whether the price is a scarcity premium or a quality premium. Left-arm pace, wicketkeeper-batters, left-arm spin — these three categories almost always command more, regardless of whether the player's international record justifies it. If I do not separate the two, my analysis becomes one more headline.

In the same way, I treat the commentary box as a data source, not a rival. I began cricket commentary for All India Radio from Hyderabad at nineteen. That period taught me that a sentence spoken in a regional language sometimes tells you more than a graph, because it captures the feel of a moment no model can measure. When a local commentator says the pitch is slowing from below, I log it as a hypothesis, and then check whether the scorecard proves it.

The reverse is also true. At the 2026 World Cup I calculated France's PPDA at 14.3, the highest among the contenders, and said they would concede very few chances. They won the tournament and conceded only six goals. At the time a television commentator said on air that I was overcomplicating the game. I remember it because it taught me that a model has to be not only built but spoken — and the speaking is often the harder part.

In 2026 the ISL was played behind closed doors in Goa. I looked at distance covered: home teams' average fell from 108 km to 104 km, and the home win percentage fell from 45 to 32. The stadiums emptied; the distance covered did not change — but the morale behind that distance did. I wrote that home advantage was psychological, not tactical. Several coaches denied it at first. The data held.

Cricket has a direct analogue: home pitches, crowd noise, and the effect of that noise on a bowler's run-up. We tend to put all of this in one bag and call it 'home advantage.' In reality these are three separate variables, and we often lack the data to separate them. When there is no data, the biggest mistake is to reach a strong conclusion because one weak number is available.

Now the contrarian angle, the part I look for in every analysis. This is the most delicate part of the whole argument.

Absence is never the absence of quality. If there is no Ranji Trophy data, that does not mean there are no good players there. It means nobody was sitting there measuring. The distinction is so obvious and so routinely ignored that I think it has become a habit. What we do not measure, we slowly begin to treat as less important.

There is a related danger I call sample-size bullying. When someone argues about a trend in the Women's Premier League, the easiest reply is: 'small sample.' That reply is almost always technically correct, which is exactly why it is comfortable. It saves you from thinking. I have had to accept that some live questions cannot be measured and have to be argued. Say what the sample can and cannot answer, then specify the test that would settle it — but running from the argument is not analysis.

A third danger is specific to my own position. I was born in Australia, live in India, and write about Indian cricket. Distance sometimes helps you see institutional blind spots, but it can easily slide into a register that reads as imported. I have to remember that domestic scouting and regional ecosystems are already doing a great deal of the work. My job is not to diagnose them but to credit them and learn from them.

The fourth danger creeps in most quietly: measuring what the model has, not what cricket is. Working with IPL broadcast data is easy because it is clean. Working with Ranji and women's cricket is hard because there is no data at all. The result is that franchise cricket coverage grows while the rest stays thin. I have made a rule for myself: tag the provenance on every claim. If my evidence comes from a broadcast feed, I say so. And I keep a fixed share of the work for the matches where nobody is tracking.

Calibration is a monk's job. I learned that in my early model-building days, when our xG model consistently undervalued shots taken from distance. We spent three weeks looking only at where the model was wrong. No headlines. No readers. But those three weeks gave us the confidence without which everything else was worthless. Cricket's data culture is missing exactly this discipline, because here most people write the conclusion before they build the model.

Now the question I want to ask last, and the one that is the real axis of this report. If all we have on a player or a match is a scorecard, should we write about it at all?

My answer: yes, but honestly. I will not claim that a scorecard yields deep analysis. I will claim that a scorecard does tell you what a team chose under what conditions. And the history of decisions is still history. Why a Ranji captain gave a spinner the ball in the fourth innings, why a batter attacked in the first innings — these questions can be asked from a scorecard, even if the answers are partial.

What worries me most is that our industry routinely mistakes a shortage of data for a shortage of story. When there are not enough numbers on a domestic player, coverage drops him, and then we say there was nothing to say. That is inverted logic. There was plenty to say; we did not build the system to hear it.

One more thing I want to make clear, about auction and injury news. This is a transfer window and rumours are everywhere. The biggest problem is that the signal-to-noise ratio keeps deteriorating. I have a simple filter: I value the story that shows a contract structure, a release clause, or a wage bill. A story that only says, on the authority of sources, that a player is moving to a team is, to me, a rumour, not news. Agent noise blurs the market's real price, and filtering that noise is the reader's right.

It is also worth remembering that injury news is often the least reliable news in cricket. A team never says how many weeks its fast bowler is out, because disclosing it helps the opposition and can depress an auction price. So when I read an injury story, I accept it only within limits, and I always write down which source it came from.

Now, forward. I will not say cricket's data structure will transform over the next few years, because I see no evidence for it. I will say this: the next time you read a cricket analysis, ask one question. Which match, which format, and which conditions does this claim's data come from? If the writer cannot tell you, what you are reading is not analysis. It is an opinion dressed as a statistic.

And the next time a system returns N/A to you, do not call it a failure immediately. Sometimes it is the most honest answer a system can give. The question is what we build to fill that emptiness — another rumour, or a measurement that tells us what we do not know.

For me the answer is clear. We turn the absence into a proxy, stick a label on the proxy, and then write honestly about what we saw and what we did not. Cricket's biggest data crisis is not that we have too few numbers. It is that we never drew up the list of questions we wanted to ask. The day we draw that list, the pipeline will stop coming back empty.