Empty Stands, Wet Ball: IPL 2026's Dew Residual and the Quiet Rewriting of Home Advantage

Hook

November 10, 2026, Dubai International Stadium. The stands were almost empty — a few hundred masked people scattered across rows kept apart by social distance. In the second innings of the final, Mumbai Indians needed 157. The chase ended in 18.4 overs. Seeing that 18.4 on the board, an old page in my notebook turned over on its own: the same number my 2026 Russia World Cup model had given France as a title probability, 18.4 percent.

That night I was not watching a match. I was watching a hypothesis. When the ball seamed, when it skidded, when dew arrived, when a spinner changed his grip — I logged all of it. The question was not simple. If a crowd is the primary carrier of home advantage, then when the stadium empties, who inherits it?

Context

In May 2026, with world sport paused, I ran the numbers on 56 Bundesliga matches played behind closed doors. The result was blunt: home advantage fell from 0.42 goals per game to 0.17, and home teams' PPDA worsened by 1.3. When the crowd leaves, a team does not only lose goals — it loses the shape of its pressing. That study went to 15,000 subscribers, was cited by two European clubs, and produced my Euro 2026 live-analysis commission.

Could the same test be run in cricket? IPL 2026 handed me a natural experiment: outside India, at three venues, without spectators, across 60 matches. No side had a genuine home ground; every squad moved hotel to hotel. The home and away labels survived only as scheduling paperwork.

I first saw the pattern in a Delhi newsletter, long before the data had a name. In 2026, at 51, I started Expected Delhi, applying xG and PPDA to the Indian Super League. Its first lesson: Bengaluru FC scored 27 goals from 22.4 xG in the 2026-17 I-League, a 4.6 overperformance. Nobody knew it before the season ended, because nobody kept a model. I want the same discipline in cricket.

So I wrote the variables down. Dependent: win percentage of the home-designated side, and win percentage of the chasing side. Independent: powerplay dot-ball rate, powerplay strike rate, middle-over spin economy, death-over run rate, toss decision. Control sample: the 60 matches of 2026, where home venue and crowd both existed.

Core

In 2026 the home-designated side won 33 of 60 matches, roughly 55 percent. In 2026, once venues became neutral, that fell to 29, or 48.3 percent. A seven-point drop was my first suspicion.

Look at the other side. In 2026 the chasing team won about 46 percent of matches. In 2026 the chasing team won 58 percent. The gap between seasons is close to twelve percentage points. The rise was not uniform by venue: the chasing edge was largest in Dubai and Abu Dhabi, smaller in Sharjah. The later the evening dew arrived, the better the ball came onto the bat in the second innings, and the less certain a bowler's grip became.

The real carrier of home advantage is dew and pitch inheritance, not the noise of a crowd. In 2026, home-labelled sides' powerplay strike rate fell 1.3 percentage points below their own 2026 baseline, and powerplay dot-ball rate rose 2.1 points. At the same time, home-labelled spinners conceded 0.31 runs per over more than their own 2026 level. There is a trap here: in 2026 the words home and away had almost no functional meaning. That decline was not a decline in home advantage; it was a decay of the variable's label.

Take Sharjah. Small boundaries, the same strip reused — pitch inheritance became an independent variable in 2026. On the same surface, batters in the second innings attacked with more confidence, because the boundary was close enough to blur the line between an edge and a mis-hit. That cause has nothing to do with a crowd, and it explained more of my model's variance than anything else.

My 2026 Pedri study is oddly relevant here. At Euro 2026 I tracked Pedri's 65 progressive passes and 92 percent pass completion — zero goals, yet his 8.3 progressive carries per 90 rated elite in my model. At the Tokyo Olympics he played six matches in 18 days; my workload model had already said so. It taught me that output numbers and process numbers do not measure the same thing. IPL 2026 is the same: win percentage is output, dew-skid is process.

The final is the last example. Delhi made 156 for 7; Trent Boult's new-ball spell in the powerplay set the tempo of the whole match. A neutral venue does not mean neutral ball behaviour — that was the most honest fact of the night. Death-over run rate rose, because a dew-wet ball was sliding out of the grip of the seamers.

That lesson has a market price. Franchise auction models now buy spinners for a home pitch and middle-order batters for a home venue. If the home-pitch variable is genuinely weak, that valuation model can misallocate two to three crore rupees every season. Metric-to-market translation starts there — not on the scoreboard, but at the auction table.

Contrarian

Here I have to stop. Sixty matches is a small sample, and the confidence interval on a seven-point swing is wide — I wrote it myself, roughly plus or minus 4.2 points. On top of that, 2026 was a Covid-protocol season: bio-bubbles, hotels without families, a compressed schedule. Venues were neutral; fatigue was not.

When the stadiums emptied, the home advantage stayed and stared back — but what stared back may not be a reflection of the crowd. The 2026-to-2026 comparison pits different squads, different form, different pitch preparation. Correlation is not causation. The crowd was absent, yes; but one-bounce pitches, new-ball seam movement and toss luck all changed too. Blaming the crowd alone is as wrong as blaming xG alone.

So I set a pre-registered threshold: I will report an effect only if it survives above five points across at least two independent seasons. The Big Bash season had limited crowds; the Pakistan Super League playoffs were played before partially empty stands. Across three leagues the pattern has held — but it is still not proof. Remove the India-market lens and this is my biggest caution.

One more thing. Who bears this decision? If a franchise reads my model, concludes home advantage is gone, and then drops a home-pitch spinner at auction, the loss lands on the analyst's desk. Wrong model decisions are paid for by people, not by numbers.

Takeaway

At sixty, I have learned that the quietest spreadsheet often has the loudest story. Next cycle I will watch two things: whether home sides' powerplay dot-ball rate rises by more than two points across their first six home matches, and whether home win percentage stays below 50 even after match 30. If it does, the empty-stadium lesson corrected us — home advantage does not live in a crowd, it lives in a pitch and in time.

The 18.4 percent model did not predict France; it predicted my next five years. I wait 900 minutes before judging a rising player, and I will do the same for Ravi Bishnoi and his cohort. Before judging a season's home advantage, we should wait at least three. A rising star is a culture — and so is a single season of data.