Missing Data Is Data: The Silent Nulls Inside Cricket Scouting

Missing Data Is Data: The Silent Nulls Inside Cricket Scouting

Two in the morning on a balcony in Sylhet, and I have a scouting spreadsheet open in front of me. Fourteen columns. Thirteen are full — average, strike rate, powerplay economy, death-over splits, a ten-match trend line. The fourteenth is blank: away economy rate. In the cell beside it the model has stamped a verdict in green: 'No concerns. Risk: low.' My hands started shaking, because the model had read that empty cell as a zero. An away economy of 0.00 means a bowler is superhuman — or it means he has never bowled a ball away from home. Two explanations, worlds apart, and on a spreadsheet they look identical. After fifteen years in Sylhet's regional press I left match reports behind in 2026, at forty-nine, and moved into spatial analysis. The first lesson from that move sits at the centre of this piece: a blank cell is never a zero; a blank cell is unknown — and treating the unknown as zero is the most expensive mistake in the transfer market.

Every club dossier is really a two-stage pipeline. Stage one is where scouts and data feeds pull out the events — who bowled how many balls, in which over, against which field setting, at what match state. Stage two is where an analyst extracts meaning — which bowler cracks under pressure, which batter finds the ball in the half-space, which team's line drops back at which minute. The relationship is simple: stage two can only build as much meaning as stage one supplies events. When stage one comes back empty, stage two cannot produce anything but a beautifully arranged framework — it can only make emptiness look professional.

The problem is cultural, not technical. When a column is blank, the easiest move is to fill it with a zero, because a zero looks clean, tidy, and never argues back with software. Yet zero and missing are entirely different objects. A bowler who has never bowled in England has an unknown England economy; it is not nil. Someone who faces three balls and strikes at two hundred in the death overs has not displayed skill; he has displayed a small sample. When a model delivers the same confident verdict in both cases, the human making the decision is being misled.

We are inside a transfer window now, and this is the season when blankness is most expensive. Headlines are full of fees and shortlists, but the release-clause structure and the wage bill are the real story. When a club commits three years on six months of data, what it is really buying is not performance but a confidence, and half the columns behind that confidence are empty. A club that flags its blank columns can price the deal; a club that reads them as zero discovers on the training ground that the bowler it signed loads onto his back foot before he brings the new ball down — a fact no dossier contained.

Agents trade precisely in that gap. Where data is missing, narrative is inevitable, and narrative always costs more than numbers. An agent's job is never to create information; it is to package the empty cell as 'hidden talent'. This silent cost never appears in a transfer fee line item, yet it leaves its fingerprint on every digit of that fee.

Core analysis: zero and missing are not the same thing

On the half-space: the half-space is where the game hides its intentions. By the same logic, a dataset's blank cell is where a model hides its ignorance. In both cases the real information lives at the boundary, not in the middle. If a batter's map shows him as outstanding on the leg side but carries no data at all in the off-side channel, that is not proof of strength — it is a hole in the scouting. A fielding captain senses that hole; so should a director of cricket. The difference is only this: on the field a mistake is punished immediately, in the transfer market it is punished two seasons later.

Every number needs a comparison context. Whether a batter averaging 35 is elite depends on era, format and pitch. An economy above eight is poor in the death overs but close to irrelevant in a first-innings Test session. When that context is itself missing from the database, comparison becomes impossible, and guesswork walks out wearing the clothes of evidence. My habit is simple: before I reach any conclusion, I ask what this number was excluded from. It is the question analysts ask least and should ask most.

On 15 July 2026, at Luzhniki Stadium in Moscow, the World Cup final. France beat Croatia 4-2, and the most important line in the scorecard was missing entirely — Croatia had arrived having played three consecutive matches into extra time, more than 240 additional minutes. Before kick-off I built a minute-by-minute timeline, and it said this: after the 60th minute Croatia's midfield line would drop roughly eight metres. Fatigue is a formation, not a feeling. Fatigue then takes tactical shape — the gap that opens between midfield and defence is exactly the half-space's address. Antoine Griezmann received the ball there; the penalty and the assist came from there. The box score at ninety minutes read 1-1; the fatigue ledger was in nobody's column.

That is the lesson. Croatia's fatigue was not a bad attitude; it was a geographic fact — who could still stand in which part of the field, and who could not. An analysis that counts only goals and assists never sees that geography, and so makes every post-semi-final forecast blind. A method that treats fatigue as mood will never treat it as formation, and that difference is worth two points in a table.

On 26 May 2026, the first major match after the pandemic pause — Bayern Munich beat Dortmund 1-0 in a near-empty stadium. Digging through fifty behind-closed-doors Bundesliga matches, I found something no scorecard carries: home advantage had fallen from 0.36 to 0.22 goals per match. Bayern's pressing intensity dropped twelve per cent in the first fifteen minutes. The reason is not mysterious — pressing is a social act; crowd noise pulls the press trigger earlier. With the crowd gone, the trigger arrives late. They do not erase pressure; they relocate it.

The easy error here was to log 'crowd = zero' when the correct entry was 'crowd = absent'. That distinction decides whether a pressing model is right or wrong. Building a silent-press model cost me three days and one missed deadline, after which I imposed a forty-eight-hour cap on myself. The lesson is plain: naming a missing variable correctly matters far more than avoiding a wrong model.

On 11 July 2026, the Euro final at Wembley — Italy and England 1-1, Italy winning 3-2 on penalties. Jorginho's pass completion was 94 per cent, alongside twelve pressure regains. From my years of watching matches I can say this: read separately, 94 per cent completion is close to meaningless. A midfielder who drops off on every attack and takes the safe pass will post a high number; a midfielder who takes responsibility for winning the ball back and breaks his line will post the regains. Writing a story off one number is not faith in statistics; it is laziness wearing statistics' name.

On 6 August 2026, the Tokyo Olympics women's football final — Canada and Sweden 1-1, Canada winning 3-2 on penalties for gold. I analysed that match with exactly the same method I used beside the men's tournament: the same press map, the same workload accounting, the same fatigue timeline. Women's football is not a separate tactical category; only its coverage budget is separate. In 2026 I began working in pairs with a woman data scientist in Dhaka, swapping pressing models and trying to break each other's assumptions. That habit taught me something: the biggest missing dataset is usually not on the pitch, it is in the coverage decision.

In a transfer window this gap is at its most valuable. One league has full ball-by-ball tracking; another has only a scorecard. So two equally talented players arrive with fourteen filled columns and six. The market then mistakes a difference in numbers for a difference in ability. A side that signs three years on six months of data is buying a probability, and the risk inside that probability is written down nowhere. That is where the largest cost hides — not in the fee, but in the difference between understanding and not understanding.

In Bangladesh's domestic game the problem is sharper. A bowler in Sylhet works a whole season, yet his ball-tracking data is stored nowhere. On the selection table his column is blank, and a blank column is naturally read as 'no evidence' — which means no case at all. So those with data get more opportunity, and those without drop out for want of proof even when the talent was never missing. This is not deliberate discrimination; it is the discrimination of inertia, and inertia is harder to fix than a plan. In 2026 I was dropped from a television panel on the grounds that 'women don't read formations'. That day I understood that the easiest way to close a door is to never open it.

There is a parallel in tournament football. When a side produces an upset, its best player is taken by a bigger club in the very next window — the upset then becomes evidence about the player, not the team. In domestic cricket, if a spinner produces a dazzling season, the data built around him arrives just after he has already left. The cost of missing data is always paid by the weaker side; the profit is collected by the bigger one. This is not moral drama; it is a rule of accounting.

Contrarian angle: the reflex is to buy more data — more cameras, more subscriptions, more scouts. I think that route leads to the wrong address. The damage does not come from a lack of information; it comes from passing off the lack of information as information. An honestly empty cell costs nothing; the cost arrives when it is auto-filled with a zero, and that zero paints the model green. Buying more data shrinks the empty cells, but it creates new ones — a new league, a new format, a new match situation. Without discipline, the problem simply scales.

The second error is fear. Many analysts believe that writing 'no data' makes their work look weak, so they bury the empty cell. The opposite is true: the only report you can trust is the one that states plainly which questions it cannot answer. A report that answers everything has probably asked nothing.

The most dangerous dimension is procedural. If a system automatically passes analysis downstream, a report full of blank cells can arrive as 'no risk'. Nobody then knows that a decision was never actually made — it appears to have made itself. In cricket this is a familiar scene: at a selection meeting someone says 'there is no bad report on him', when the truth is that there is no report on him at all. Silence is not evidence; silence is only silence.

Takeaway: watch one specific thing in this transfer window — how many club and board profiles are published with an explicit note on which information is missing. At the next selection, check whether a decision is accompanied by verifiable metrics, or whether the columns are filled again with 'long experience' and 'team requirement'. The first club to admit its blank cells are blank will price risk better than anyone over the next three seasons — even if its data budget is smaller than its rivals'.