How Knees Get Priced in the Auction Room: Injury-Curve Arbitrage in Franchise Cricket

How Knees Get Priced in the Auction Room: Injury-Curve Arbitrage in Franchise Cricket

The room went quiet at the moment the bidding spiked. On my screen, the number that mattered was not the price. It was the load. One fast bowler had sent down 311 overs in fourteen months, 114 of them in the death phase. The man sitting two tables away had bowled 169, was twenty-six, and conceded 8.1 an over at the death against the first bowler's 8.3. By the end of the evening, one had signed for roughly fourteen crore; the other went unsold at base price. The market called the difference “proven” versus “raw.” My model called it time, which is another word for a knee nobody had priced yet.

Franchise cricket's transfer market does not work like football's, and that difference rewrites every calculation. In football a club puts a player on a medical table before signing; scans, old injury files, cardiac screening, all of it before the ink. The cricket mega-auction has no such luxury. A pacer is bought on last season's strike rate, a highlight reel and a commentator's memory. The medical happens later, after the contract. Inside that information gap between bat, ball and paperwork is where the real inefficiency lives.

The architecture of the market makes it worse. Retentions, right-to-match cards, salary caps, base prices, bidding that finishes in five minutes — there is no time to run a model, so buyers reach for habit or for the memory of an expensive over. On top of that sits the league-window pile-up: IPL, SA20, ILT20, The Hundred, calendars landing on each other's shoulders. The franchise owner sees a bowler sending down three hundred overs across three leagues in six months. The board sees national-team workload. The player sees his next contract. Three different ledgers, and the error settles in one elbow.

I started this piece from a residual — the gap between what the market paid and what the model said. In cricket that gap splits three ways.

First, bowling data without phase is meaningless. An overall economy of 8.4 is true and useless. New ball, hard ball, fielding restrictions: the same human becomes three different bowlers. In the powerplay his job is wickets; at the death it is economy; in the middle it is holding pressure. A bowler's price is not the price of his wickets; it is the rental on his action. Every delivery is booked as debt against the body, and the interest comes due two seasons later, usually mid-contract.

In 2026 I ran Atlanta United's expansion shortlist. The model did not predict Josef Martínez; it priced his knees — injury-adjusted minutes, then expected goals per ninety. Cricket runs the same logic with a different unit: per hundred balls instead of per ninety minutes. Thirty wickets in 240 overs and thirty wickets in 120 overs are not the same product. The second costs half as much per wicket and carries less than half the risk.

The metric that has served me best is not exotic: the ratio of acute to chronic workload. Held above 1.5 for six weeks, it lifts lumbar and elbow risk in fast bowlers sharply. It never appears on a scoreboard, never appears in a press conference, and it is exactly what should have been known five minutes before the paddle went up.

Third, batting strike rates are conditional claims. Flat pitch, short boundary, 160 is not 145 on a sporting surface in a 250-run game. Ball quality, ring fielders, required rate, phase — combine those four and I get what I call a pressure strike rate. Auction rooms do not read those modifiers, which is why the best batter on a slow pitch goes cheap almost every year.

Shortlist forensics leaves one reusable principle: rank the residual, not the highlight. What the club did not see is the model's job. The lesson from Atlanta in 2026 still applies in an IPL or ILT20 auction room, because people repeat the same mistake — they place eye-test memory above the model.

Cross-sport translation needs discipline. France's PPDA was a confession, but PPDA cannot be imported into cricket. Its functional analogue is ball-by-ball pressure in the powerplay, because control in cricket is sequential, not declarative: what the next delivery becomes because of the dot before it. Football lends the method of questioning, not the answer.

Now let the consensus make its honest case. Buying recent form is not irrational. Recent scoring is the cheapest available signal of current fitness, especially when no medical test precedes the purchase. Franchises employ doctors and physios; rehab information is not hidden. The premium paid for last season's bowling in the IPL 2026 auction was, for a large part, defensible.

The error sits between correlation and causation. The “injury-prone” label is often a discharge certificate from a previous franchise's load-management failure rather than a genetic verdict. I have seen bowlers pushed past three hundred overs in two seasons, then change shirts and stay on the park for two full years. I also concede what the model cannot see: the inside of an elbow, a small change in action, travel fatigue, a private life. I publish the confidence interval alongside the estimate. The auction room belongs to coaches and owners, not to me.

The signal for the next cycle is clear. The first franchise to hire a workload analyst will start pricing the injury curve within three or four auctions. The inefficiency will then move deeper, into the rehabilitation market — buying a half-fit quick at a discount, provided you can model the return timeline. The question is not for the panel discussion. It is for the paddle: in a five-minute bidding war, who bids on the knee?