The Empty-Stand Ledger: Which Parts of Cricket's Home Advantage Were Ever Real

31.4. That is not an economy rate, and it is not a strike rate. Across 217 international cricket matches I scraped ball by ball between 1 March 2026 and 31 December 2026, that was the home side's win percentage in fixtures played in empty or near-empty stadiums. Across the five years immediately before, in the same formats and comparable match sets, the same figure was 52.7 per cent. Home advantage in cricket decays slowly; it does not lose twenty-one points in a season. The spreadsheet began to hum, and I knew the broadcast was over. I keep returning to 8 July 2026, the Rose Bowl, Southampton. The first international cricket match after lockdown, England against West Indies, not a spectator in the ground, cardboard cut-outs taped to the seats. Jason Holder's West Indies won by four wickets, and the unease in the commentary box is not the sort of thing that reaches a scorecard. What I wanted to measure lived outside the scorecard: how many of those quiet overs quietly rewrote an lbw, a review, a death-over decision. I started the work because “home advantage” has spent decades operating as a black box in cricket. Pitch dryness, grass cover, an umpire's instinctive lean, crowd pressure, travel fatigue — all of it goes in, and a single percentage comes out. After I modelled Russia's pressing at the 2026 World Cup through passes allowed per defensive action, one belief hardened in me: a single metric does not change reality, it only changes the angle from which you see it. I ran the PPDA numbers again, and the flat in Moscow started to feel real. Cricket has no PPDA, because pressure here is not built through defensive actions but through ball-by-ball state value. So I built a three-tier index. Tier one is visible: results, run rates, the shape of wicket clusters. Tier two is the shadow of officiating: umpiring decisions, lbws, caught-behinds, review outcomes. Tier three is the bolder one: the gap between home spinners' economy and the scoring rate of home batting units at the death. The data came from four broadcast feeds and three scoring databases, each delivery carrying its own timestamp. I kept the older dataset of 1,200 football matches from 2026 beside it, because in the ghost games the crowd disappeared, but the pressing lines left fingerprints. The question is the same in both sports: does the crowd actually drive the game, or merely our explanation of it? Since taking a BCB advisory role in 2026, I have seen fragments of internal scheduling and venue documents, and the picture sharpened — much of home advantage is a decision-making advantage. The answer is not simple, but the data carries one obvious fracture. The collapse was not uniform. On Asian soil home advantage held almost intact; in England, Australia and New Zealand it fell apart. In Asia the home win rate slipped from 52 per cent to 47 — five points. In England, Australia, New Zealand and South Africa it dropped from 58 per cent to 33. That gap is the real story. Home advantage in Asia is largely pitch curation: dry, slow, spin-friendly surfaces cut to the home spinners' measurements. An empty stand does not empty the pitch. In England and Australia, by contrast, a large share of the advantage is atmospheric — thirty thousand people push a fast bowler's roar into a batter's ear and tilt an umpire's instinct. Tier two made my hands shake. In tight lbw appeals, the share of decisions going the home side's way fell from 58.1 per cent to 52.4. Umpires did not become more accurate on average; the direction of the error moved. Home noise, a slow replay on a giant screen, a stadium's collective shout — these create a measurable lean that no law book records. My ethical guardrail applies here: the number does not call umpires incompetent, it says humans are human and lean in measurable ways. Tier three produced the most unexpected finding, at the death. In T20 ghost matches, the home side's death-over strike rate fell by roughly 7 per cent while away bowlers' economy improved slightly. Away batters' death-over strike rate barely moved. The damage, in other words, landed on the attack that had thirty thousand people standing behind it. To a visiting batter, an empty stadium is neutral. The model did not predict the wicket; it predicted the regret of ignoring it. The spin numbers are subtler still. In Asia, home spinners' economy worsened by about 0.5 runs per over, because the rhythm of using a slow pitch becomes far more legible without crowd pressure. Away spinners' economy hardly changed — nobody took the home label from them, only part of the audience. IPL 2026 was staged entirely in the UAE, 19 September to 10 November, and no team played at home. On 10 November in Dubai, Mumbai Indians beat Delhi Capitals by five wickets in the final. Home-ground advantage was zero, yet one advantage was visible: sides that arrived early and trained longest, with stable camps, moved ahead. Home advantage had quietly become acclimatisation advantage. The lesson that stuck: the advantage does not vanish, it changes address. This is where my ethical kill switch engages. However precise, the model conceals a cost ledger. In the 2026-21 season fewer spectators returned, but the expense landed on other shoulders — stewards, ground staff, canteen workers, many of them migrant labour, whose night shifts appear in no column of my file. There is a monastery in every dataset, and its silence is not empty. The figure I record as lost home advantage is, for someone, a missing payslip. Now the testimony against my own work. “Empty stands” is a comfortable cause, but it is correlation. In 2026-21 many variables moved at once: bio-secure bubbles, travel restrictions, neutral venues, workload management, even toss and fielding patterns. Home advantage in cricket was never purely about a crowd; it was institutional — which pitch gets prepared, how much grass stays on it, which bowler gets rested. The 2026 World Cup matters here. On 16 June at Old Trafford, India versus Pakistan filled the ground, but it was nobody's home crowd — the audience followed the flag, not the geography. England were nominal hosts with no home crowd of their own. The real variable is not “home crowd” but “partisan crowd,” and my first model confused the two. The true test arrived in late 2026, when partial crowds returned. The crowd came back; the home win rate did not fully follow, and nowhere returned to its pre-2026 level. My model had overfitted — it granted the crowd more power than the crowd possessed. In several cases, travel schedules and pitch planning outweighed attendance. I do not trust the eye test until it can survive a scatter plot, and this eye test did not survive. Then came the real turn. I began building a “crowd value model” that would price a full stadium in runs, or in run-rate saved. I built it for six days and deleted it in one night. A model that prices a crowd hands administrators a number they will use in the next crisis: keep the gates open, it is worth this many runs to us. At precisely that point, an analytical method stops being a method and becomes a weapon. Watch elsewhere next season. The question is no longer whether crowds drive the game; it is which parts of home advantage travel from one place to another. Measure the pitch-curation index first, the crowd decibels second. In T20 franchise cricket, where the index itself is being rewritten as teams buy travel and rest advantages instead of home ones, old percentages kept in stone will make us tell the wrong story. Six months from now, when a home side loses at home and someone on commentary says the crowd is back but the rhythm is not, keep one number nearby. It will not give you an answer. It will only ask which specific advantage is being missed — and who was holding it.