The Empty Spreadsheet: Why 'N/A' Is Cricket Analytics' Most Dangerous Answer

The Empty Spreadsheet: Why 'N/A' Is Cricket Analytics' Most Dangerous Answer

Hook: The Report That Contained Nothing

Last night a scouting-and-analysis report sat on my desk. Eight sections — format and match analysis, player technique and data, team landscape and ranking, league and commercial ecosystem, rules and governance, risk matrix, public narrative and expectation, and the cricket industry transmission map. Every cell of every table returned the same line: 'N/A — insufficient information, cannot assess.'

No innings score. No powerplay average. No pitch report. No player's name. Only an empty frame, and row after row of blank cells inside it.

From the outside this looks like failure. I read it differently. An empty dataset is the most honest mirror cricket analysis has. It proves what a real analyst does when the information isn't there — he does not invent it. He writes 'N/A.'

The trouble is that in the real world, nobody writes 'N/A.' They write 'the momentum shifted.' They write 'he's a big-match player.' The blank cells get filled with imagination. This piece is about that filling-in process — and why it is the most dangerous habit in cricket.

Context: Cricket's Two Parallel Worlds

The volume of data I hold now was unthinkable a decade ago. Hawk-Eye and ball-tracking systems measure the speed, line, length, seam movement and bounce of every delivery. The contact point, bat angle and trajectory of every shot are logged. In T20, an 'expected runs' (xR) model now runs for every batter — a concept parallel to football's xG, though the two games are not built the same way.

Yet beside this bright picture sits another world, far larger: the world of data poverty. Associate-nation matches, women's domestic cricket, and pre-2026 domestic records — here there is no ball-tracking, sometimes only a handwritten scorebook.

South Asian cricket analysis grew up inside that poverty. This is not a shortage of talent; it is a shortage of structure. Where ball-by-ball data does not exist, the analyst works with inference and observation. That compulsion bred a certain caution in this region's analysis — and a certain laxity too.

The reason is simple. When data exists, the number delivers the verdict. When data does not exist, who delivers it? The answer to that question hides the core of today's argument.

One more layer has been added by the fantasy and betting market. That market wants fast answers, however weak they are. So 'certain predictions' built on small samples and inference spread across thousands of screens. Market demand does not improve data quality; it only improves the performance of confidence.

Core Analysis: An Audit Trail From Zero to Verdict

I do not treat my work as a match report. I treat it as a model audit. A hypothesis, a dataset, an anomaly, a recalibration, then a verdict — in that order. The verdict arrives last, and it is unhedged.

The Discipline of Sample Size

Cricket's biggest statistical trap is the small sample. A T20 innings is 120 balls. A batter faces roughly 15 to 25 balls per innings. So 'form' across three matches means 60 to 75 balls in total. In that sample, a commentator will declare a batter 'back in form' after three innings at a 200 strike rate. But much of the gap created across 75 balls is variance, not skill.

For bowlers the arithmetic is crueler. A T20 bowler delivers 24 balls per match. Four matches means 96 balls. Across those 96 balls, an economy rate leaping from 6 to 10 is normal, with no tactical change at all.

I follow this rule: before publishing any claim, its sample size, time window, format and venue adjustment must be written down. Without them, I do not print the number.

Here is an example. Say a bowler has kept a 5.5 economy across his last four matches. The headline writes itself: 'back in brilliant rhythm.' But the underlying data says two of those four matches were on spin-friendly pitches, and the opposition was a lower-table side. The sample is 96 balls, the opposition's batting depth is shallow. Here the environment contributes more than the bowler's skill. Without a context-integrity note, the number is meaningless.

Era Adjustment: When 85 Strike Rate Is Heroic, When It Is Ordinary

Against nostalgia I have one weapon — rate adjustment. In 1990s ODI cricket, a strike rate of 70 was normal and 85 was elite. Today 85 is below par. The same number, a different era, an inverted meaning.

Test batting averages deceive in exactly the same way. An average of 50 built on uncovered pitches, weak fielding and slow over-rates is not the same as today's 50. So before building any 'greatest of all time' list, without era adjustment it is romance, not analysis.

Mixing Formats: One Metric, Two Meanings

Another silent trap is mixing formats. Placing a Test average beside a T20 strike rate to prove someone is an 'all-rounder' produces confusion. In Tests a batter's goal is different — occupying time, protecting wickets, surviving the new ball. In T20 the goal is the opposite. The same player's two numbers tell two different games' stories; blend them and the story becomes false.

The Mathematical Death of the Word 'Momentum'

In commentary, 'momentum' is the most-heard word, yet it has no definition. What is momentum, really? Two overs of heavier scoring? Then it is a number — strike rate or run rate. Where something can be explained by a specific mechanism, the word 'momentum' is redundant.

Instead of momentum, I measure specific things: the run rate of the last five overs, the longest streak of consecutive dot balls, the fall in scoring rate in the two overs after a wicket. Read together, these three numbers show whether the game truly turned — or only the commentary did.

Death-Over Entropy: When Pressure Can Be Measured

Pressure is not a feeling; it is a system. How predictable a team's scoring plan is in the death overs can be measured with entropy. If, from overs 16 to 20, a side plays only two shots — the big hit and the yorker-ball-out — its plan is predictable, meaning entropy is low. The opposing captain sets the field easily.

I have seen this pattern repeatedly in Bangladesh's chasing record. Of the chases lost on home soil between 2026 and 2026, a large share saw the run rate collapse in the last four overs — because the plan had gone monochrome. Pressure peaks exactly when the variety of the plan is lowest.

The 2026 Ghost Matches: A Natural Experiment

After the pandemic in 2026, cricket returned to empty stadiums. In football, against the previous 306 Bundesliga matches in which home teams won 43.2 percent, the 83 matches played without crowds saw that figure fall to 33.7 percent. Goals fell from 3.1 to 2.7. Cricket's sample is smaller, but the question is identical: how much of home advantage is crowd, and how much is travel fatigue?

This window matters because it is a controlled natural experiment — same teams, same grounds, only the crowd changed. But caution is due. The year 2026 cannot become the explanation for everything. I must decide in advance which effect I will call '2026-specific,' and which data would falsify that explanation.

Pre-Registration: Decide First, Do Not Hunt Later

This is my most controversial habit. Before starting an analysis I write down which results I will treat as support and which as refutation. Because if you hunt afterwards, you always find what you want. There are so many angles and sub-groups inside the data that finding evidence for your side is easy.

This is the data-mining trap. Run twenty separate tests on twenty variables and, by coincidence alone, one 'significant' result will appear — if you choose to pick it. Cricket's sub-groups are countless: left-hand versus right-hand, powerplay versus middle, home versus away. Without pre-registration, analysis becomes advocacy — decision first, evidence later.

The Bounded Role of the Eye Test

This is where I make my biggest mistake — or used to. A love of metrics taught me to neglect observation. But in a world of poverty, the eye is sometimes the only witness. So I do not remove the eye test; I define its role. The eye is a hypothesis generator, not a verdict judge. If the eye and the model disagree, I do not suppress the verdict — I publish the disagreement itself.

Venue, Toss and Dew: Environmental Variables

In Tests, home teams historically win more — a blend of pitch, weather and familiarity. In T20, dew changes the second innings' equation and ruins spinners' grip. The toss matters too. Unless these variables are separated out, false credit or false blame lands on a bowler's shoulders.

The Bangladesh Context: Discipline Inside Poverty

Sitting in Rangpur, I built my first expected-value model — I was eighteen. It taught me that where there is no structure, the analyst must build the structure himself. In Bangladesh's domestic cricket, that compulsion is daily. Many Dhaka Premier League matches have no detailed ball-by-ball data; there is only the result and a handwritten score.

Yet inside this poverty lie extraordinary achievements. In international cricket, the dual milestone of over 14,000 runs and over 700 wickets belongs to one man alone — Shakib Al Hasan. Mustafizur Rahman's reliance on the cutter, or Tamim Iqbal's longevity — these too are not mere luck; they are the product of structural continuity. But even in using such numbers, unless sample and format are separated, the analysis goes astray.

Contrarian Angle: Complete But Biased Data Is More Dangerous Than Empty Data

Here is the most unwelcome truth. We think empty data is the big problem. The big problem is complete but biased data — because it answers wrongly, with confidence.

Picture a model trained only on home-match data. Every cell is filled, every metric is present, the report looks immaculate. But the model does not know how far that player's strike rate falls outside home conditions. A blank cell warns you; a full cell puts you to sleep. So the real audit question is not 'is there data,' it is 'is there the right data.'

This is why I am careful with metric imports. Moving from football's xG to cricket's xR, I state in advance which logic transfers and which does not. In football a shot's value depends on location and the assist; in cricket a ball's value depends on line, length, pitch and match situation. A one-to-one translation is impossible. Where they do not match, I do not force the translation.

One more uncomfortable point. Complete data carries another trap — overfitting. When a model explains the past perfectly, it goes blind to the future. In cricket, those who claim 'perfect prediction' must remember every failed forecast as much as every successful one.

Takeaway: What to Watch in the Next Signal

Ball-tracking is now spreading into associate and women's cricket. Over the next two years 'N/A' will shrink — that is nearly certain. The question is whether our discipline grows with the data. Or whether we grow more confident and make bigger mistakes.

I say this: a model is a monastery — you enter with noise and you leave with discipline. Before the next match, ask yourself one question — do you truly hold data, or only a story? If the answer is the second, the bravest act is to write one word: 'N/A.'