The Skeleton of an Empty Gallery: A Provenance Audit of Home Advantage in Asian Cricket

May 2026, Signal Iduna Park. Dortmund versus Schalke, the Revierderby. Not one person in the stands, only sponsor boards and camera cables. I watched that match with a spreadsheet open, every press timestamp in its own column. Dortmund's PPDA was 6.8, Schalke's 14.2. Dortmund ran 113.4 kilometres. In xG terms the match was 2.7 versus 0.4, the result 4-0. But the number that kept me awake was not on the scoreline. Across 83 empty-stadium Bundesliga matches, the average home advantage fell from 0.42 goals per game to 0.18. The crowd had vanished, yet the advantage had not fully gone. The empty stadium did not erase home advantage; it exposed its skeleton.

That skeleton is what I want to carry into Asian cricket. The Mirpur gallery, the Premadasa deck, a Sharjah night — here the crowd was never merely an atmosphere. The crowd is a variable. And where variables are tangled this tightly, a sentence like Bangladesh plays well at home is not data, it is inherited lore. My years of watching matches tell me that in the subcontinent the small box beside the scorecard lies the most — home advantage, dew factor, pressure. I logged 1,842 shots before I trusted the pattern, and this piece begins with the same restraint.

What does data provenance mean in Asian cricket? It means asking where a scorecard actually comes from. An international match keeps a ball-by-ball log through two separate channels — the on-field umpire's signals, and the broadcast feed with scoring software. Most days the two agree. On the day they do not — a leg-bye recorded as a bye, a wide as a bye, a third-umpire review call — small discrepancies accumulate and quietly rewrite run rates, a bowler's economy, even the story of the match. I do not start a fielding analysis until the scorecard has been reconciled against the ball-by-ball log. It is slow, tiring, and boring to most editors. It is still the only path.

The problem is not confined to scorecards; it lives in selection culture too. In Bangladesh cricket, decisions are often made on career averages, which are really a cushion — a comfortable sofa where you avoid the hard chair of current form. A 40-match career average cannot tell you that over the last 10 matches this batsman has dragged his strike rate against spin down to 70. A career average speaks of the past; selection needs the future. And the language of the future is the rolling window.

Asian match context adds three more variables that Europe lacks. Pitch curation — how spin-friendly or seam-friendly a home side can make a surface — is home advantage directly. Dew — once the ball gets wet in the second innings of an evening, spinners become ineffective and the winning equation shifts. Scheduling — travel, rest, the pressure of back-to-back fixtures. All three operate outside the crowd, and mixed with the crowd they manufacture the thing we call home advantage.

This is where the Crowd-Absence Coefficient (CAC) I built earns its keep. The idea is simple: if removing the crowd reduces the advantage, how much of it was the crowd and how much was pitch, schedule and umpire? In 2026 Europe was forced to ask that question. Asia was not, but we have the opportunity — treat limited- or no-attendance series as natural experiments. What remains when the crowd is gone is the real structure.

Now the actual work. I pre-commit three windows — 10, 20 and 50 matches. The reason for pre-committing is simple: choosing a convenient window after the fact, what I call window-gerrymandering, is the easiest self-deception there is. If someone says he has been weak over the last 7 matches, I ask — why 7, why not 10? That question alone kills half the argument.

Metric selection is pre-committed too. For batsmen at home I look at four things: dot-ball percentage between overs 11 and 30; strike rate against spin; post-powerplay run rate; and a pressure index, meaning the gap between chasing and defending strike rates. For bowlers I look at: the ratio of economy to wicket-taking rate; dot-ball percentage in the death overs; a finger-spin versus wrist-spin split; and review success rate.

Say a middle-order batsman's strike rate against spin at home is 82 across a 50-match window, but over the last 10 matches it has fallen to 64. The career average still looks fine, because the older innings are doing the cushioning. This is where the stability score comes in — how close the three windows sit to each other, and which way the trend points. An average that says the same thing across three windows is trustworthy; an average that survives only in the 50-match window is a historical document, not a present-tense one.

The stability score is not just a number; it is a warning. The man stable across 50 matches and the man collapsing across 10 are two different patients needing two different treatments. One is in a form slump, the other is the victim of a structural change. Feeding both the same medicine is not analysis, it is laziness.

The same logic holds in bowling. A spinner's 50-match economy is 4.4, but over 10 matches it is 5.2. Two possibilities: his finger-spin split has shifted, or he has been bowled more often in dew-heavy evening matches. Without data these two cannot be separated, and if they are not separated the decision will be wrong.

Here the Bangladesh thread returns. Shakib Al Hasan is the only cricketer with more than 7,000 ODI runs and more than 300 ODI wickets — a rare career record. But that record does not tell you in which window he should be bowled how much, in which position he should bat, which match he should be rested for. A record is history; a decision is the future. Confusing the two is our oldest disease.

Sensitivity is a rule of my writing. I never write about a single window. I place the 10-, 20- and 50-match results side by side, and if the three tell different stories, I say so. That transparency is a contract with the reader. Hide it and the numbers look prettier, but the bet goes wrong.

This is why the empty-stadium data is so valuable. The no-attendance matches of 2026-21 showed that some sides still won at home with the crowd removed — because their advantage lived in the pitch, the schedule, the familiarity, the umpire relationship. The crowd was one layer of the advantage, not the whole house. An analyst who treats the crowd as the entire cause is selling a roof as a foundation.

The biggest trap sits right here: correlation is not causation. Bangladesh plays well at home in Mirpur, and Mirpur draws big crowds — the two happen together, but one is not the cause of the other. Home results come from pitch curation, spin-friendly conditions, a travel-light schedule and players' familiar routines. The crowd is one input in that equation, not the only one. Remove the crowd and home advantage shrinks, but it does not reach zero — because the rest of the structure is still standing.

Home advantage is, in truth, a curated product. The home board builds the pitch, arranges the schedule, sets up the camp, and draws on umpire familiarity. The crowd is that product's packaging. Strip the packaging and the product looks less attractive, but the product still exists. Those who claim home advantage dies in an empty stadium have failed to tell packaging from product.

The second trap is system-fit fatalism. If a player does not fit the current template, discarding him forever is not data analysis, it is comfort. In cricket roles shift, teams shift, conditions shift. A batsman may fail at the top but see his strike rate leap as a finisher at number five. If the data judges only by career position, the possibility of an alternate role is never even seen. System-fit does not mean a permanent verdict from the system; it means calculating the cost of a role change.

The third trap is provenance paralysis. Chasing proof, many never write at all, because every number carries doubt. I solve this with a pre-registered evidence threshold — if three windows point the same way, I write with confidence, and I write the uncertainty level alongside. Staying silent is not neutrality; it is also a decision.

Another trap is treating the empty stadium as a pure laboratory. Even without a crowd there is attendance pressure, TV pressure, family pressure, player workload. So I never pull a conclusion from a single variable. Attendance figures, noise levels, umpire decisions, player load — I read all four together. A claim standing on one of them collapses, and it collapses at the worst possible moment.

Provenance itself has blind spots. The ball-by-ball log is human-written, so it carries errors. The broadcast feed is bound by sponsor shadows. The selection record is not always public — how much was injury, how much politics, how much rest never reaches the archive. I admit these blind spots in the writing, because an analyst who does not know the limits of his own model is not running a model, he is floating on faith in one.

What will I watch in the next round? Across Asia's coming series I will track three things — whether the pitch-curation trend is shifting, which innings dew is hitting hardest, and whether career average or rolling window is winning in selection. The day a selection committee picks a squad by window, our home-ground arithmetic becomes far cleaner.

A bet is a hypothesis with a scoreline attached. A good hypothesis tends to produce a good scoreline, but certainty never arrives. So I love the process, not the result. The process is patient; the result is moody.

I do not chase narratives; I archive them until they confess. In Asian cricket, the story we keep telling about home advantage may well be true — but being true and being proven are not the same thing. So the question matters: when the next empty or half-empty gallery arrives, will you see the absence of a crowd, or the clarity of a structure?