When the Auction Applause Stops: Price and Value in Cricket's Franchise Transfer Window

Hook

One scene from last season's auction night is still stuck in my eye. An overseas opener, pushing thirty, had a base price of forty lakh. Across the table, three franchises raised their bats in turn. Ten minutes later the price had climbed to two crore ten lakh. In the very next set came a left-arm spinner whose powerplay economy had been among the league's top five for two seasons. Base price thirty lakh. Nobody bid. When the announcer said "unsold," a light laugh rippled through the hall. I wrote that laugh down in my notebook, because it was the most valuable data point of the night — not who went for how much, but which question nobody asked.

After that night I decided I would not write auction news off the headline. I would write about the players who have no price but do have value. Every number is a person who never got to explain themselves.

Context

Cricket's franchise transfer window now makes as much noise as football's. The IPL, the Bangladesh Premier League, ILT20, SA20, the CPL, The Hundred — the picture is the same everywhere. Retention, right-to-match, release clauses, salary caps, overseas quotas, agent commissions. And riding on top of all of it, what the audience sees is a three-word headline: "So-and-so signed so-and-so."

The trouble is that sentence carries no information. Price tells you who is in demand; it does not tell you who is effective. That is true in football, and truer in cricket — because in cricket a player's job splits into fixed phases. Powerplay, middle overs, death overs. A bowler's powerplay economy and death economy are entirely different skills. At the auction table, though, nobody makes that split; there is just one line — "Bowler, right-arm, medium pace."

My work started from that gap. In 2026, at the Khulna District Stadium, I logged 24 matches onto a paper grid and built my own model, because back then no provider offered advanced data for the Bangladesh Premier League. I built the model by hand, because the league deserved to be counted. That was my first lesson: if you wait for the dataset to exist, the writing never happens.

Core

This season I pulled the scorecards of 240 matches across four franchise leagues. No provider would chart it, so the counting became a kind of prayer. The aim was simple: to trace the relationship between price and value.

I kept the model deliberately simple. For each player I produced two numbers. One, "contribution above replacement" — how much more than the league's average player he delivers, weighted by phase. Two, "value per crore" — that contribution divided by his auction price. I set the phase weights by match state: powerplay wickets and death-over run rate both earn a premium, because both change the result directly.

Three things emerged.

First, the average value-per-crore of the ten most expensive buys was 34 percent lower than that of the twenty cheapest buys. In other words, spending more does not buy you more return; the most expensive corner of the market is the least efficient.

Second, of the three bowlers who conceded under a run an over in the powerplay across those four leagues, two were unsold or went at base price. Meanwhile those conceding 1.4 in the same phase cost three to five times more. Nobody asked, "which part of the over?" They only asked, "how many wickets?"

Third, and this is the most uncomfortable — at least six players with an injury history over the past two seasons fetched full price at auction. Injury risk was not priced in, because risk is not measured, it is guessed. Yet that risk can wreck a franchise's entire season.

I know someone will look at these numbers and say, "hand-built model, small sample." Agreed. I am not handing down a final verdict on anyone from 240 matches. But a small sample and a wrong sample are not the same thing. I will also state plainly what my model cannot see: dressing-room chemistry, a player's understanding with the captain, who adapts to home conditions — none of that shows up on my grid.

Contrarian

Now the part where I distrust my own numbers.

There is a relationship between price and performance — but it is not a cause. The market sets price through demand, and demand is set by three things: scarcity, passport and timing. Overseas slots are limited, so an overseas finisher's price inflates almost by default, however good he is. Timing means that if he plays one good innings just before the auction, his price jumps. That is not cricket; that is a market.

This is where the agent enters. A good agent knows when a particular franchise has a particular gap, and that gap sets the price — not the player's real contribution. A transfer is a story wearing a spreadsheet like a coat. The structure of release clauses and retention is really the agent's storytelling space, not cricket's.

And there is one angle cricket journalists write about too rarely. Live data feeds now flow to betting companies, second by second. The same data I use to read tactics, someone else uses to place a bet. A player's form, injury updates, even post-toss conditions — all of it is now raw material for a business. I am not telling anyone to bet; I am saying that where a dataset goes matters as much as its numbers.

So I will not tell anyone, "buy these ten and you will win." Betting football and team-building cricket — I am sceptical about both. What I will say is smaller: when someone calls it "the biggest buy of the year," ask them — in which phase, over how many overs, and how was his body in the six months after the injury?

Takeaway

In the next transfer window I will watch one thing, and it is not auction day — it is the week before. I will watch which franchise uses a release clause in place of retention. Because a side letting go of an old player to lean toward a new one is really saying: "We are changing the structure, not the names."

And I will leave one question I cannot answer. If franchise cricket really is a market for skill, why does the most expensive buy of each season so often turn out to be the least efficient? Either the market is inefficient, or my model is still incomplete. In both cases, there is no option but to keep counting.