The Dew Excuse vs the Ledger: What 70 Ball-by-Ball Matches Say About the Second-Innings Advantage

A regular-season night in the IPL. At the toss, the captain ran a hand over the grass beside the pitch and said into the mic: the dew will come, we'll chase. The dew never came. The air stayed dry, the ball gripped, and spinners took four wickets between overs seven and fifteen. In the dressing room afterwards, the explanation collapsed into one line — we didn't win the toss.

I wasn't at the ground that night. I was in front of a laptop, with a spreadsheet that kept an empty column beside the ball-by-ball card. The spreadsheet remembered what the stadium forgot: at that venue, in that window, there was no measurable dew-driven change at all. Spin's economy in the second innings rose by 0.3 runs per over. As an explanation for a nine-wicket defeat, dew is a hypothesis, not a verdict.

I work as a sports data analyst, based in Bengaluru, rooted in Dhaka. That is the vantage point of this piece — I cover cricket for the Indian market, yet the first habits in my logbook were built watching Bangladeshi cricket. The friction between those two viewpoints is not a liability to me; it is a lens. My entry into this work came in 2026, scraping 12,400 event records from a Bengaluru FC season to build an xG model, and since then every match report has started with a metric, not a story.

This regular season I logged ball-by-ball events from 70 matches. Powerplay run rate in each innings, separate spin and pace economy between overs seven and fifteen, death-over boundary risk, the elapsed time of third-umpire reviews, and rough temperature-humidity estimates for a handful of games. Eighteen venues in all. Logging match after match, the noise eventually turned into a signal. I watched a few of those games from the stands, trying to read what the camera never shows — the condition of the ball itself.

Let me state the limitations plainly: this is my own logged dataset, not the official scorecard. Field placements, injury updates, dressing-room pressure — none of that is in my columns. I keep a column for what the broadcast never shows, and most of its cells stay empty.

One citable fact up front, because the question grows out of it. The highest successful chase in IPL history is 262 — on 26 April 2026 at Eden Gardens, where Kolkata Knight Riders had made 261, and Punjab Kings won it on Jonny Bairstow's 108 off 48 balls. So a big target is not impossible in the second innings. The question then flips: how much is the second-innings advantage really worth, and what exactly is it an advantage in?

In my ledger, 46 of 70 toss-winning captains chose to field — 65.7 percent. In the language of the dew theory, that is a rational decision. But the results column speaks differently. Across the 70 regular-season matches, chasing sides won 40 of them, or 57.1 percent — there is an overall second-innings edge, but it does not match the confidence implied by 65.7 percent.

If the edge really belongs to dew, it should vary by venue. That is what I tried to measure. At the six venues where my logs most often recorded the ball feeling damp in the second innings — coastal, high humidity in April and May — chasing sides won 62 percent. At six dry, inland venues, that number fell to 51 percent, essentially a coin flip.

Spin is the cleanest proxy for dew. A wet ball hurts spinners — if dew were the main cause, spin economy should rise in the second innings everywhere. At humid venues it did: 8.4 in the second innings against 7.6 in the first. At dry venues the gap was close to nothing — 7.1 against 6.8. So dew's effect is real but venue-conditional; and the bulk of that 57 percent win rate cannot be explained by dew.

Where does the rest come from? Two patterns kept returning in my logs.

The first is the powerplay decision. Sides batting first show an average gap of 11 to 14 runs per 100 balls between their strike rate in the first six overs and their strike rate afterwards. That is, they start comparatively slowly in the powerplay — keeping wickets in hand — then attack after the 14th over. But when they finally attack, there is no fixed number in front of them, only an estimate. The side batting second is denied that estimate, because it has a number. The real second-innings advantage is not in the damp grass; it is in the clarity of the scoreboard — with a definite target in hand, the entire risk calculus changes.

The second pattern is middle-over wicket gambling. In the first innings, spinners between overs seven and fifteen are used in a control role; economy matters, wickets are a bonus. In the second innings, the chasing side hunts boundaries against those same spinners, because the run-rate demand is written on the scoreboard every over. Spin's wicket-taking rate rises in the second innings, but so does its boundary-conceding rate. The net effect is roughly zero across venues, tilting only slightly toward spinners at humid grounds.

This is where transplanting a football model helped and failed at the same time. In football I measure pressing intensity with PPDA — passes allowed per defensive action. Cricket has no direct equivalent, because pressing in bowling is not pressing in defence. I built a proxy: average fielder distance from the batter per powerplay delivery, and the share of rough-length bowling in the middle overs. The model could measure powerplay aggression, but it stayed weak at predicting middle-over wicket falls — because in football, pressing is broken by a pass, while in cricket it is broken by a batter's rhythm, and that is not in my columns.

I also keep a rhythm calculation. I log the run rate in the five overs after any review that goes to the third umpire. In matches where those decisions dragged on, the combined run rate of both sides over the following five overs fell, on average — the cricket version of a goal celebration going cold. That is not an accusation, it is a time-cost calculation, and it helps show where the pressure inside a match actually went.

Now the counter-argument has to be turned on my own numbers. Chasing sides winning more — is that a cause, or a selection effect? Captains choose to field when they believe the pitch will ease later. The toss decision itself carries information: the side confident about the surface is the side that chooses to chase. So part of that 57.1 percent is not a pitch advantage at all; it is the fingerprint of decision quality.

The sample is small too. Seventy matches, ten teams. A 57 percent win rate carries a 95 percent confidence interval of roughly 11 to 12 percentage points — meaning the true rate could be 45 percent or 69 percent. Anyone building a toss theory on the back of that sentence should keep one column permanently empty.

And one match deserves recording, the one where the model lost. My calculation gave that side a 71 percent win probability in the second innings. They lost five wickets for 40 runs and let the game slide. The model could read the pitch and read the wickets, but it could not read the rhythm of the man at the crease — the first ten balls of a batter returning from injury are not logged by anyone. The eye test is a hypothesis, not a verdict; and logging something does not make it true, it only makes it recorded.

In the next cycle my eye will be on toss decisions at the dry venues. At those six grounds, where chasing sides win 51 percent, continuing to field out of habit would be the most easily exposed silent gap — invisible on the scoreboard, but a shove to the table once you look.