The Auction Ledger: Which Numbers Actually Speak in the T20 Franchise Market

The Auction Ledger: Which Numbers Actually Speak in the T20 Franchise Market

It was two in the morning and I was staring at the screen. Jeddah's auction stage was burning money, and I had my own ledger open in a flat in Mumbai. Rishabh Pant went to Lucknow Super Giants for INR 27 crore — the highest price ever paid for a single franchise season, as recorded in ESPNcricinfo's post-auction report. In the same room, on the same bidding list, a death bowler went unsold — a bowler with an economy of 8.9 across his last two seasons.

In my ledger, that bowler's price should have gone up, and the gap between the market's number for that batter and my own was roughly threefold. This piece is an explanation of that gap.

Context: why an auction looks like a risk portfolio to me

When I joined Mumbai City FC as a junior analyst in 2026, my assignment was specific — build an xG model for 18 Indian Super League matches. Five years later, auditing Morocco's low block at the 2026 World Cup in Qatar, I was writing the same idea in a different language: defence is not absence, defence is a budget. How much space you spend to buy how much risk — that is a calculation, and the language of calculation is the same across sports.

I kept an ISL xG ledger, and then the World Cup asked for real-time confession. In cricket the same instrument becomes ball-by-ball. A delivery has four inputs — type of ball, line and length, field setting, and the batter-bowler match-up. Together they give me Expected Runs Added, or ERA: what a normal batter should have extracted from that ball.

Then comes the second layer, which I call leverage. A dot ball in the 20th over and a dot ball in the 7th over are never the same currency. A six when your side is 5 runs behind is not the six you hit when your side is 60 ahead. A model that ignores this difference will misprice an auction.

So I split every innings into three blocks: powerplay, middle overs, death overs. Then I weight each ball by match state. Sitting in empty stadiums taught me that a model can hear its own assumptions — and unless those assumptions are written down, the model is a number factory, not a decision tool.

Core analysis: four places where the market misprices

One: powerplay strike rate is an incomplete truth. In my ledger I split powerplay runs into two blocks — balls in the first three overs, and balls in the last three. In the first three, the ball is new, the field is in, two slips are standing. In the last three, the ball is softer, the field spreads, and strike rates jump. A batter who faces the first three overs will always show a lower powerplay strike rate. A batter who walks in at number three and only bats the last three overs sees his strike rate artificially inflated.

At the December 2026 auction, Mitchell Starc went to Kolkata Knight Riders for INR 24.75 crore and Pat Cummins to Sunrisers Hyderabad for INR 20.5 crore. In both cases the market was essentially buying one job: wickets with the new ball. In my ledger that job is called powerplay wicket value, and it is worth far more than powerplay economy. Bowlers who concede at 7.2 an over in the powerplay are common; a bowler who takes three wickets in 42 balls at that same economy is rare.

Two: the price of a middle-overs spin match-up is written nowhere. I regard the middle overs as the most undervalued zone in T20. At 7.5 an over, a team thinks it is fine; at 10 an over, it thinks it is a disaster. But the real calculation is the match-up. A leg-spinner turning the ball into a left-hander and an off-spinner turning it away from a right-hander can show almost identical economy, yet one may be 35 percent better than the other on runs per wicket.

I build a match-up matrix. Rows are batter types, columns are bowler types, cells hold balls faced and runs over the last three seasons. Any cell under 30 balls I treat as uncountable and mark 'insufficient sample.' The biggest mistakes in the franchise market are born in exactly those small cells.

Three: economy is the wrong denominator at the death. To price death bowling I do not use economy, because the denominator is not constant. If the batting side is already 140 at the 17th over, that over is dead; conceding 12 costs nothing. If the batting side is 110 at the 19th over, the same 12 runs can flip the match.

So I compute a pressure-weighted economy, multiplying every ball by its capacity to change the likely result. In my ledger, a celebrated death bowler bought for INR 1.5 crore shows a pressure-weighted economy of 9.6, while an unfashionable bowler who was released shows 8.1. That is 1.5 runs an over, roughly 90 runs a season — often the difference between two results.

Four: the overseas quota is an invisible tax. To an overseas player's fee I attach an availability tax. He may cost INR 3.2 crore a season, but national duty means he is available for only 8 of 14 matches. The real cost is then around INR 40 lakh per available match. Add the cost of the replacement player, who draws his own salary.

I read transfer rumours the way I read variance: loud early, rarely significant late. The name you hear most before an auction very rarely delivers the best return. The real signal comes from squad construction: how many left-arm spinners a side has, how many can bowl the pressure overs, how many are comfortable on their home ground.

The replacement-level calculation everyone avoids

An auction price is not the price of runs; it is the price of runs above replacement level. In my ledger I set a replacement level for each role. Roughly 128 strike rate for openers, 145 for finishers, 8.4 economy for a death spinner. The further above that line a player adds runs, the better the return per crore.

This is why a INR 7 crore batter who scores at 170 instead of 145 can be a worse investment than a INR 2 crore batter who scores at 142 instead of 130. The second player's surplus runs are identical; his price is a third. In January 2026, when I was screening 14 targets for a Mumbai-based agency and an ISL club, I used exactly this calculation — progressive passes, xG chain, and pressure-resistance indices.

What the ledger cannot see

My ledger has no dressing room. A number cannot tell you whether a 22-year-old is exhausted after three straight seasons, or whether he is still learning. I have a specific objection to how development-age players are handled — bodies are pushed into senior rhythms before they are finished, and the model then prices them on their best season. That is not the model's fault; it is the user's.

The ledger cannot see which bowler stays calm under pressure, or who breaks after conceding 40 in his first four overs. That can be measured, but I do not yet have a clean way to measure it. Where I cannot measure, I write 'uncountable' and quote no price.

Contrarian angle: a finisher's numbers are interest on someone else's labour

Now the difficult part. The thing my own ledger has got most wrong over three seasons is finisher valuation. I first assumed death-overs strike rate was an independent skill. Later I saw it is often a dependent number.

Consider it. If a side's openers bat the first 12 overs at 5.5 an over, the finisher behind them never faces a chase of 180; he is not required to drag 145 to 175. Conversely, if the top order posts 120 in 12 overs, the finisher bats on an open ground, against a softer ball, into a spread field. His strike rate touches 200, but the job was easier.

In other words, a finisher's death strike rate is largely a function of the top order's efficiency. When I corrected for top-order performance and re-ranked finishers, the list almost inverted. Of the three finishers paid the most in the market, two were not in my corrected top five. That is correlation against causation — and in the auction market the difference is worth about INR 2 crore.

The second trap is sample size. Eleven matches a season, 20 to 25 balls each — that is not proof of a skill, it is a wobble. Fifty-two off 31 balls looks beautiful in a ledger, but the same batter can strike at 96 the next season with nobody surprised, because the model never made a claim. Structure is not bureaucracy; structure is the shortest path to a repeatable decision. So every scouting report I write carries the sample size and the confidence level separately.

The third trap, and the market's biggest inefficiency: injury risk. Clubs apply a discount multiplier to injury-prone profiles, but the multiplier is usually too small. For fast bowlers over 30, my ledger assumes available matches at 75 percent of the two-season average, and prices that. Most boards never apply the correction — so they pay full price for a player who will be missing for a quarter of the season.

Signal for the next window

So what does my ledger say for the next auction? I will watch three things. First, middle-overs spinners — especially those bowling to left-handers; the market still prices them near replacement level. Second, bowlers who take the first three overs of the powerplay, because that is the most underpriced 18-ball block in the match. Third, anyone whose every innings can be measured by leverage — a player who bats the same way whether his side is 60 ahead or 30 behind.

The rest is a year of waiting. The ledger stays open; and I am stating now that if I am wrong, I will write that down in this same column.