I Reopened 462 Hand-Coded Matches: Where BPL Home Advantage Actually Lives

In April 2026, while the Bangladesh Premier League was returning to empty stands, I opened a three-year-old spreadsheet in my Chattogram study. Four seasons. 462 matches. Every ball hand-tagged — shot location, body part, defensive pressure, and the number of people in the ground. I reopened the hand-coded season, and the margins disagreed.

With crowds, home teams won 43.7% of their matches. In the spectator-free season, that fell to 37.9%. Five point eight percentage points. The easy reading is that noise wins matches. After reconciling the columns by hand, I am fairly sure the easy reading is wrong.

═══ CONTEXT AND METHOD ═══

I made my ODI debut for the national team in 2026 and played until 2026; sixteen years of broadcast work in Chattogram followed. In 2026, as Facebook Live and YouTube highlights swallowed the evening television wrap, I was hand-counting every shot of Chattogram Abahani's 22 matches — 588 attempts, 197 on target. That table reached 44,000 people and three club analysts. It fixed my voice permanently: no claim without a denominator.

For every match I logged the date, the start time (day or evening), the toss winner and decision, the first-innings score, a dew proxy (Mirpur matches starting after 6pm), pitch usage counts, attendance (club announcement versus my own count), venue, and margin. The ledger has seven versions. Version three contained an eleven-ball counting error across two Chattogram matches; I corrected it in version four and never deleted the old file. A ledger that hides its own mistakes is not a ledger — it is a press release.

My rule on empty rows is simple. Eight matches have no recorded attendance anywhere. That is not zero; it is unknown. Two matches were washed out. Also not zero. Fourteen months of silence taught me that empty rows are not zeros. When the league stopped mid-season in March 2026, I did not write opinion — I spent fourteen months re-coding all 462 matches. Silence is a dataset, and I read it for fourteen months.

═══ THE EVIDENCE CHAIN ═══

I broke home advantage into three parts: the clock, the dew, and the noise.

The clock first. In Mirpur evening matches, the side batting second has won 58.4% of the time; in day matches that drops to 49.1%. A large share of home advantage does not sit in the stands — it sits in the clock. The team handed the evening slot gets a wet ball, a ruined spinner's grip, and a punished fielding side. I have watched this from the Mirpur floodlights for years, but watching is not evidence. Columns are evidence.

The noise second. The gap between full and empty grounds is 5.8 percentage points, which looks meaningful. But across four seasons the home win rate wandered between 41.2% and 46.4% — a natural spread close to four points. 5.8 points is a signal, but it is not larger than the ordinary wobble of four seasons.

The toss third. Home captains have won the toss 51.2% of the time — a fair coin. The coin does not hand the home side anything. For anyone who equates home advantage with toss luck, that number is uncomfortable.

The timeline fourth. The script bends at the 17th over. In my log, home sides trail through most of the first sixteen overs at home, then the dew and a bad death over rewrite the arithmetic. The bend is sharp in Mirpur and mild in Chattogram.

At bowling level the picture sharpens. Home spinners average an economy of 7.1 in home evening matches against 8.0 away; home quicks concede 8.4 at the death at home against 9.1 away. I can split this player by player, because every ball in the ledger carries a bowler's name — Taskin Ahmed's death overs, Mehidy Hasan Miraz's powerplay spells, Mustafizur Rahman's cutter-reliant overs each sit in their own column. I still will not declare a trend from one name; small samples have proved me wrong more than once.

There is a layer the table never shows. In franchise cricket, bigger leagues increasingly treat small-league prodigies as satellite assets — not to fill a home quota, but to raise a price. For names like Litton Das or Mushfiqur Rahim, much of the load they carry is structural, not personal. The ledger misses that, but the over-allocation patterns do not.

Squad construction is flattening too. Almost every franchise assembles the same template — one left-right pair, one finisher, two spinners. Pitch usage counts make that sameness visible.

Some history belongs here. On 22 March 2026, Bangladesh lost the Asia Cup final at Mirpur to Pakistan by two runs; in 2026 they lost the T20 Asia Cup final at the same ground to India. Two major finals, two home venues, two defeats. Dew gives you an edge; it does not hand you a trophy.

═══ THE COUNTER-CASE ═══

That crowds restore home wins is still unproven. It is still an estimate. The confounders are tangled: bio-bubble scheduling, neutral umpire appointments, DRS protocols, a change of ball supplier, and a different mix of day and night fixtures all arrived together. The 37.9% figure rests on just 34 matches, where the confidence interval is roughly three percentage points wide. The true effect could plausibly sit anywhere between two and nine points.

There is a second trap: treating the pre-COVID baseline as clean. It was not. Neutral venues, rescheduled fixtures, half-written attendance fields — all of it was already there. I sort every gap three ways: a true zero (the ground really was empty), missing-at-random (nobody wrote the attendance down), and unobserved (the part of the ground off camera). Stitching those three together and calling it a baseline is how bad baselines are born.

═══ WHAT TO WATCH ═══

Next season I will be watching two things: chase-conversion rate after the 16th over in evening matches, and the dew proxy. I will not publish this after the tournament ends. I will file the chart before the final, so the timestamp can testify for me.

The question left is simple, though not comfortable: are we actually trying to bring the crowds back, or are we only trying to fix the clock?