4,112 Balls in a Silent Mirpur: Which Room Home Advantage Actually Hides In

The first page of the notebook was not blank

November 2026, Sher-e-Bangla National Stadium, Mirpur. 7:20 pm, 22 degrees Celsius, 64 percent humidity, zero spectators in the stands. I sat at my desk in Rajshahi filling a ball-by-ball notebook: dot balls per over, who was bowling, which delivery opened a death spell, and how many seconds an umpire took before raising a finger after an appeal. Across 33 matches the book accumulated 4,112 balls and 274 death-over deliveries.

From that book a number surfaced that I had not been looking for. In the Bangabandhu T20 Cup, the side treated as the designated "home" team saw its death-over wicket-taking rate fall from 38 percent to 24 percent. Same squad, near-identical bowling rotation, same pitch. One thing had changed: the sound.

This piece is about that number. And about my own model, which I have sat down to break.

Context: 34 degrees, 81 percent, 88 overs

August 30, 2026, at the same Mirpur ground. The thermometer in the press box read 34 degrees Celsius; humidity was 81 percent. Australia's innings was in its final phase and Shakib Al Hasan was walking into his tenth wicket. Bangladesh won that match by 20 runs, and Shakib's figures were 5/68 and 5/85.

Check the Test record books: the rare double of a century and ten wickets in a single Test had, before him, been achieved only by Ian Botham (2026) and Imran Khan (2026). That is the numbers angle.

What I was measuring that day was something else. Bangladesh's bowlers sent down 88 overs across four days at 34 degrees. I logged every bowler's over-beat before and after each spell. Among the spinners, bounce inconsistency that normally ran at 2.9 feet in an eight-over spell climbed to 4.1 feet once the temperature crossed 32.

I filed that as the match report for my daily newspaper.

Then I sat down and wrote something separate, 4,200 words long, asking one question: after absorbing 88 overs of thermal load, how did Bangladesh's spinners still hold their bounce in the final innings? I did not write it for the paper. I wrote it for my own newsletter, The Half-Space.

The daily's editor never ran it. Nine hundred subscribers arrived in eleven days.

Since then my opening line has stopped being a score. It has become a load: overs bowled, minutes played, degrees Celsius.

Core: three hypotheses and one number

The Bangabandhu T20 Cup ran entirely at Mirpur in November and December 2026, with no spectators. I coded all 33 matches. 4,112 balls.

What the number said: home-side death-over wickets fell from 38 percent to 24 percent. What does that actually mean? When a stadium empties, where does home advantage go?

I had three hypotheses. I tested all three.

Hypothesis one: pitch preparation. Home advantage comes from conditions fitted to the home side. If a zero-crowd tournament forces the curator into less time and a slower surface, death-over wicket-taking should fall.

The data did not support it. In 29 of the 33 matches, the gap in boundaries per over between the first and second halves of an innings stayed under 1.2. The pitch did not change. Neither did the seam of the ball.

Hypothesis two: bowler rhythm. Crowd noise lifts a bowler into an extra over. This is the most comfortable argument.

But when I lengthened the spell-level data, I found that across the seven spinners who regularly bowled a fourth over, the standard deviation of their strike rates barely separated between matches with a crowd and matches with none. 0.4 of strike rate, on a 22-ball sample. That is noise, not signal.

Hypothesis three: umpire decision latency. I hesitated here, because this is ammunition for trolls, not serious data. I counted anyway. Of the 94 lbw and wide decisions timestamped in my notebook during the 33 spectator-free matches, mean latency was 6.3 seconds. In the four playoff matches, where minimum staffing created some presence, it was 5.1 seconds.

A small difference. On 94 decisions, half a second may be an artefact. But then something clicked.

I had set out to measure a bowler's death-over performance. Yet the biggest weapon a bowler has at the death is not a bouncer or a yorker. It is a fact that cricket writing almost never states: death-over wickets come from umpire decisions, not from deliveries.

Where the bat went is not measured by the ball. It is measured by a person, whose range of hearing and vision sits in a specific spectrum. Twenty-five thousand voices in a full stadium manufacture an illusion that an empty ground does not.

Among my 4,112 balls, I sorted every marginal lbw appeal and wide in a death over. In the spectator-free matches, 31 of 142 appeals had to be logged out as doubtful. With a crowd present, that figure was 23.

The difference is not in the ball, it is in the decision

I am not the best person to evaluate three pitch-wicket counts. I spent twenty-seven years in a cricket newsroom guessing who was favourite and who was not. That is long-service labour.

Not this time. When I carried one notebook through sixty matches in Kazan in 2026, one thing became clear. In football I logged 1,200 pressing sequences. In cricket the number is smaller and the field narrower, but the mechanism is identical. The game does not change; the context changes.

I could argue that the very basis of calling a side a "home team" in a domestic tournament is a category error. Playing at home in a franchise trophy does not mean playing in your own seat. Eleven overs into a match at the Wankhede, the home side takes a wicket, and every away team plays to the same arithmetic.

So where is the difference manufactured?

Here I return to a line I wrote in my notebook from that 2026 Mirpur match: a tactic is a hypothesis; a match is its peer review. Bangladesh's spin tactics in 2026 were built on the supposition that, in hot air, Australian batters would block more than they would slog. Australia's entire innings became a laboratory in which the supposition held.

In 2026 there was no laboratory, because a spectator-free ground had cut the number of live variables. Three variables dropped out: noise, sweat and inhibition.

The first two are measurable. The third is not, but it leaves a print on an umpire's fingernail.

Contrarian: the mistake I had made for years

Cricket writers have an easy habit. When a match has no crowd, we write: the atmosphere is ruined, the contest has faded. It is a comfortable sentence, because it needs no evidence, since an empty seat proves itself on sight.

In August 2026 I made exactly that mistake, and a year later an indicator landed in my hands. The biggest note of my career was never measured. It was decision latency as a constant.

The idea is not easy to convey, so I will slow down. Some cricket decisions are not really taken on the field; they are taken outside the metric. A half-volley is one. A front-foot no-ball is one. Technically, these are decisions. But a match holds more decisions that have no technical basis at all: dressing-room setup, umpire confidence, the songs of a stadium, the timing of a bowling change. That is where the spectator-free match becomes a different game.

So my third conclusion is this. The cause was not an illusion; the cause was a real problem. Noise influenced the umpire, and that influence had a measurable outcome. In my notebook that outcome is called a doubtful death-over lbw.

Now let me cut off my own hand. This data has three weaknesses, and I will list them.

First: across 33 matches there were only 57 death-over wickets. On 57 events, a 14-point difference is not statistically durable. Until a second season of data arrives, I will call this an observation, not a thesis.

Second: the idea of a "home team" was meaningless in this tournament. Five sides, every match at one ground. Nobody was genuinely away. In a franchise system, home and away means a fixed venue, not a fixed team. In that arrangement home advantage cannot be measured.

Third: I was not a spectator. I was on a 32-inch screen in Rajshahi, two seconds behind the live ball.

The third point matters most, because it puts my method on trial. If I track the ball from a screen, I invert the role of the spectator. I am not hearing the ground; I am hearing numbers.

Which means the umpire's decision is not my data. It is my interpretation of my data.

What I did not find in my own data

Here comes the least comfortable part of this piece. Across those 33 matches in 2026, there was one thing I never measured: what does a leg-spinner think before bowling a death over in an empty ground?

I did not think about it for years. I thought about conditions, pitches, over rotation, thermal load. Consider it: 88 overs of bowling at 34 degrees, 81 percent humidity. The variable I measured in 2026, I looked for in the bowler's body, not in his head.

My newsletter has a small recurring line, written after 2026: the 2026 silence was not an absence — it was a variable with a pulse. I believe it, because I measured it.

And because I measured it, I understand something else: cricket's largest datasets will never arrive, because measuring them would require shutting the game down for good.

What I will watch in the next match

So I move the conclusion forward. If this is right, then when crowds return in the 2026 domestic season, home-side death-over wicket percentages should return towards 38 — and the measure of that return should be the count of doubtful lbws, not bowling strike rate.

I have a rule in my filing. I write the date of a number. I write who measured it, on how many samples, under what conditions. Next to today's 24 percent, my notebook records a date: December 18, 2026. Mirpur.

At sixty-four, I still trust the anomaly more than the average. The average tells you everything is fine. The anomaly gives you a question.

And what gave the question this season was an empty seat in the stands.

Sources and conditions: Bangabandhu T20 Cup, November 24 – December 18, 2026, Sher-e-Bangla National Stadium, Mirpur; all matches played before no spectators. Match and ball-by-ball coding by the author; sample of 33 matches, 4,112 balls, 274 death-over deliveries, 57 death-over wickets. 2026 data: Bangladesh v Australia, first Test, August 27–30, 2026, Mirpur; Bangladesh won by 20 runs; Shakib Al Hasan 5/68 and 5/85. Temperature and humidity readings taken by the author on site.