The Price of a Knee: Injury-Curve Arbitrage and the Miscounting of Cricket's Transfer Window

December 19, 2026, Dubai. Within minutes of Mitchell Starc's name being read out at the auction stage, the price touched 24.75 crore rupees, the highest ever paid for a single player in IPL history. At the same table, Pat Cummins went for 20.5 crore rupees to Sunrisers Hyderabad. Both are fast bowlers, both play all three formats, both have captained their national sides. Yet in the spreadsheet open on my laptop, those two numbers did not sit on the same logic. One price was built on output. The other was built on availability.

Back at the hotel after the auction I ran a cricket version of an older model, swapping goals for wickets and minutes for overs bowled. The result was oddly simple. In franchise cricket, a fast bowler's auction price is poorly explained by his runs per over. It is explained by how many overs he is likely to bowl in a season. The model does not predict anyone's future. The model puts a price on his knee.

Context: two market layers, and a third one in the gap

The IPL transfer window does not operate in one place. The first layer is retention and auction, where the working parts are purse, base price, overseas quota and board NOCs. The second layer is trade, where an all-cash deal lets one franchise buy a player directly from another. Hardik Pandya's return to Mumbai Indians late in 2026 was exactly that kind of deal, with no player exchanged, only money. Between these two layers sits a third market: the UAE's ILT20, South Africa's SA20 and the Caribbean Premier League. The same fast bowler gets three different prices in the same year, and to his own team each of them feels correct.

There is machinery behind this, not hype. Every franchise sits down with a fixed budget, and every slot carries an opportunity cost. Buy one overseas batsman and less money remains for a fast bowler. So an auction price is never a measure of quality. It is a measure of a squad's deficit. Where three fast bowlers already exist, the fourth pacer's price drops, because he is not needed. Where two pacers are injured, the same bowler's price jumps, because panic has entered the price.

That is why the IPL auction table sometimes looks like an incomplete market. And in an incomplete market, the analyst has one job: find where the error was placed.

Core: injury curves, over-budgets and the discount nobody sees

The lesson from building Atlanta United's first shortlist was this: a player's limitation is not a separate cost, it is an input. In 2026 I did not look at names. I looked at injury-adjusted availability. I ran Atlanta off a spreadsheet, and one habit from there I still keep, which is never trusting raw goals or raw wickets on their own.

In cricket I break the logic into four layers.

The first is expected wickets. Wickets per innings is a good start but not the end. The real number is wickets per innings multiplied by expected innings. A pacer who concedes eight an over and takes 1.1 wickets per innings but plays nine of a possible fourteen matches contributes less across a season than one who concedes slightly more but plays all fourteen.

The second is phase scarcity. In T20 cricket the rarest asset is not just wickets, it is the ability to bowl the last four overs. A bowler who succeeds with the new ball but cannot bowl at the death loses market value, even though death-over overs are a limited and non-substitutable quota for the team. The market often prices such a bowler on two different scales: one value from powerplay economy, another from death skill.

The third is the injury curve. Between 29 and 34, the relationship between pace workload and back or lumbar stress issues is not linear. Some bowlers go long stretches without playing and still avoid injury because their national board holds them inside a disciplined workload agreement. That appears on paper, never on a scorecard. And information that never appears on a scorecard is the information markets neglect most.

The fourth is repricing. A UAE league buyer picks someone up cheaply, decides the injury fear was overstated, and two seasons later that bowler returns to the IPL at several times the price, with the same scout saying we always knew. That is my least favourite sentence in the sport.

When I lay out a pacer list across these four layers, the dispersion becomes visible. Not for Starc or Cummins, who were priced correctly. The dispersion sits one tier below: the bowler who has had one injury, whose name never made a headline, and for whom no franchise wants to raise its paddle past two and a half crore because the question of why so much keeps demanding an answer.

My method starts differently here. I do not begin with economy rate. I begin with an over-budget. A franchise has roughly four hundred overs of bowling work in a season. Powerplay takes twenty-six to thirty of them. Death overs take forty-eight to sixty. The rest is middle rhythm. How much of that budget a bowler can actually fill is his real price. A pacer who is excellent in the powerplay but does not bowl at the death should never be valued above his share of those twenty-six overs.

One caution matters, because my work spans two sports. Football pressing frameworks do not transfer directly into cricket. Minutes run in football and overs bowled in cricket are not the same unit, because cricket workload comes in bursts, not in a continuous stretch. Each over delivers four to six heavy physical impacts, separated by seconds of rest. That is not comparable to sustained running. So I borrow the question from football. I do not borrow the number.

Look at a deal shaped like Jofra Archer's. At the 2026 mega auction Mumbai Indians bought him for around eight crore rupees, and injury limited him to only a handful of matches that season. In market language that is a failed deal. In model language it is a risk-adjusted probability whose supporting structure was never priced because the data was thin. That gap is the whole story.

Contrarian: where the market was actually right

Start by conceding it. The big deals were not pure hype. Starc made a real impact on Kolkata's title run in 2026 and was named player of the match in the final, strangling the scoring rate. Franchise boards value an experienced pacer in a big match because that is a single innings with no substitute available. In Cummins's case there was a further reason, leadership, which shows up in field placement and bowler management but cannot be measured in any spreadsheet. If the market prices only those two names on that basis, the market did not get it wrong.

The problem sits elsewhere, and it comes from observation rather than model loyalty. The market does not measure everyone on the same principle. One player's injury history is punished with a threefold discount while another's identical history is completely ignored. The reason is unequal information. Public data is richer on bowlers from larger boards, so scouts know them. For an equally skilled bowler from a smaller board, the record of an old injury is stored nowhere, so the price is set by headlines and memory.

Correlation versus causation is decisive here. An injury history does not guarantee a future injury. Where the link to age, over-load, bowling action and off-field training discipline is weak, cutting a price purely on injury history is wrong. Equally wrong is bidding to the ceiling on a clean injury record, because a pacer who has never been injured is rare in the market, and that does not make him invulnerable.

Takeaway

In the space between auction and trade, the information priced lowest right now is not the player's name. It is the contract terms. NOC clauses, workload clauses, injury insurance accounting and release-clause structure are the documents that reveal whether a franchise is buying a bowler's season or merely his arm. Over the next two windows I will be watching the franchises that begin buying terms alongside players. Everyone buys players on the field. The team that settles the arithmetic off it collects the discount. So one plain question remains: is your franchise buying a fast bowler's arm, or his entire season?