A batsman scores 78 off 42 in Tuesday's match. The narrative writes itself. He is in form. He is dangerous. He is, in the language of sport, hot.
The odds for Friday's fixture reflect this. The market prices him as a threat. Punters follow. The narrative compounds.
The data, more often than not, disagrees.
What the research actually shows
The hot hand was first examined formally by Gilovich, Vallone and Tversky in 1985, using NBA free throw and field goal data. Their finding was stark: what observers interpreted as streaks were statistically indistinguishable from random sequences. The perception of momentum was real. The momentum itself was not.
The research has been refined since. Miller and Sanjurjo (2018) identified a sampling bias in the original analysis and argued that a genuine hot hand effect does exist in some contexts, though smaller than intuition suggests. The debate continues. What is not contested is that human observers substantially overestimate the persistence of form.
In T20 cricket, the conditions for overestimation are even more favourable.
Why T20 amplifies the problem
A T20 innings lasts 20 overs. A top-order batsman might face 30 to 50 deliveries in a good game, fewer in a bad one. Across an IPL season, a player might bat in 12 to 15 matches. That is a small number of observations.
Small samples are where narratives take hold most easily. Three good innings in a row is enough to build a reputation. Three bad ones is enough to destroy it. Neither run tells you much about the underlying probability.
Beyond sample size, T20 introduces additional variables that raw form cannot account for:
Pitch conditions vary by venue. A batsman "in form" at Chepauk faces very different conditions at the Chinnaswamy. The same swing that produced 60 runs last week may produce nothing on a surface offering early movement.
Opposition bowling changes. A team in form against a pace-heavy attack is not necessarily the same team in form against quality spin. The matchup matters more than the recent scoreline.
Squad rotation is constant. Rested players return, injured players drop out, tactical changes shift batting orders. The "hot" batsman from Tuesday may bat at a different position on Friday.
The data available to a serious model treats all of this as context. A recent high score is one data point among many — and often not the most informative one.
How the market misprices form
Bookmaker pricing reflects, in part, public opinion. Public opinion is pattern-seeking. When a team wins three in a row, the market drifts toward them in ways that systematic analysis frequently identifies as excessive.
This is not a criticism of bookmakers. They price to balance a book across a wide population of bettors, most of whom weight recent performance heavily. The result is that markets for teams and players perceived as "in form" tend to be slightly shorter than the underlying probability warrants. The mispricing is modest. Across a season, it is detectable.
This is one of the mechanisms through which our machine learning models find value. Not by ignoring form, but by weighting it appropriately alongside five other analytical dimensions: batting cohesion, bowling quality, momentum indicators, venue-specific performance, and inter-season consistency. A recent high score influences the model. It does not dominate it.
When form does carry signal
This is worth being precise about. The hot hand effect in T20 cricket is not zero.
Within a short window of the same tournament, on similar surfaces, against similar opposition, a team that is genuinely performing well will carry some momentum. Player confidence affects batting and bowling outputs in ways that are real, if modest. A team that has solved a tactical problem often continues to execute it correctly.
What the data resists is the version of "form" that commentators and bettors reach for: the narrative version, where a 78 on Tuesday is treated as a reliable predictor of 78 on Friday. That is not what the numbers show.
The signal is narrower than the story. Models are built to find the signal. The narrative is built for consumption.
The practical implication
Before the next IPL match, you will read about who is in form. You will hear that a particular batsman is "unstoppable right now" or that a bowling attack has "found their rhythm." This framing is not useless — there is something in it, statistically, when the context matches.
But the version you read on social media, written six hours before a fixture, weighted almost entirely on the last two or three games? That is not research. It is pattern recognition dressed as analysis.
The underlying data — venue history, bowling matchups, seasonal consistency, resource-adjusted batting metrics — carries far more predictive weight than any single game's scorecard.
Past performance does not guarantee future results. That is true of EdgeXI's track record and it is equally true of Tuesday's 78.
EdgeXI's IPL 2026 recommendations are published on Tipstrr before each match and tracked publicly. Follow on Telegram for daily updates during the season.