Death-Over PPDA: A New Pressure Index for T20 and the Mispricing of Bowler Value

Kensington Oval, Barbados, June 29, 2026 — in Dhaka the clock had already turned to the small hours of June 30. India 176/7 on the board, South Africa at the point in a chase where a single misjudgment shuts the door. The number that stayed with me that night was not a batter's strike rate. Jasprit Bumrah finished the tournament with 15 wickets at an economy of 4.17 and the Player of the Tournament award. That economy over four overs works out to six or seven runs an over, 83 across a full twenty — a figure that barely exists in the death-over economy of T20. But when I sat down in the following weeks to match numbers like that against franchise auction prices, the disorder surfaced. The bowler with the highest death-over pressure is not reliably the most expensive; and half the value of the men who do fetch the biggest fees is created in the powerplay, where pressure means something else entirely. I have walked into this trap before in football. In Russia in 2026 I was tracking France's PPDA — passes allowed per defensive action — which settled at 12.4, while the same notebook carried Kylian Mbappé's 0.18 xG per shot. PPDA drew the pressing lines, and Mbappé broke them, showing that shot location and progressive carries were a more valuable asset than raw speed. I wrote that he was a €200m asset inside eighteen months — Root: 2026 — World Cup PPDA and the Mbappé Value Call. The lesson there was that a tournament is a pricing laboratory. Then in 2026, working through empty-stadium Brasileirão data, I found home win percentage had fallen from 52.1% to 42.6%, home goal difference was down 0.27 per match, and distance covered was flat — Root: 2026 — The Empty Stadium Home Advantage Study. Since then every piece I write opens with a confidence interval rather than a declaration. In cricket I built the xG notebook to see which Paulistão truths would survive the math, and which were television highlights dressed as evidence. Cricket has no passes, so PPDA will not transplant cleanly; that has to be admitted before anything else. In football, PPDA counts how many passes an opponent completes before you take a defensive action. The honest T20 equivalent is how many free balls a batting side receives before the bowling side forces a pressure event — a dot, a mis-hit, a false shot, a wicket. In the powerplay, pressure comes from the new ball and the field restrictions. In the middle overs, from spin and field placement. From the 16th over onward, pressure comes only from execution: the yorker, the slower ball, the hard-length cutter, and line discipline. Wicketkeeping for Udity Club in the Dhaka league taught me something no tracking software has: from behind the stumps you read the batter's feet and grip pressure and know which length he intends to attack. Data makes that reading measurable. It does not replace it. My ball-by-ball model rests on five inputs. First, pressure-event rate: the weighted sum of dots, false shots and wickets per ball from overs 16 to 20. Second, free-ball ratio: deliveries where the batter receives the ball in his strong zone with nobody covering the corresponding field position. Third, execution fidelity: how often the planned delivery actually arrived, read off pitch maps, especially for yorkers and slower balls. Fourth, batter suppression: how far an opposing batter's strike rate in that phase drops below his own season baseline, adjusted for venue par. Fifth, leverage: the win-probability swing attached to each individual ball. Combining those five gives what I call DPI — the Death Pressure Index — expressed as runs saved above phase par per twelve balls, because like PPDA it should be a rate, not a residual. In my season logs, bowlers above 0.8 DPI carry a death economy near 7.2; bowlers below 0.2 sit at 10.4. That gap is six to seven runs a match, seven or eight percent of an innings. The real work starts after de-adjustment. In the 2026 World Cup wicket charts, Bumrah was not top. The joint leading wicket-takers were Fazalhaq Farooqi and Arshdeep Singh with 17 each. Wickets and pressure are not the same commodity. Farooqi did his best work in overs one to six, where new-ball swing and the left-arm angle manufacture uncertainty at the top. Buying him as a death specialist is a category error. Arshdeep is worth roughly double, because he matches Farooqi in the powerplay and then, from overs 18 to 20, uses that left-arm angle to push batters off the crease, effectively deleting the slog-square boundary. Rashid Khan and Wanindu Hasaranga are simpler accounting. Their pressure is built in overs seven to fifteen, before the obligation to accelerate has fully arrived. Rashid's dependence on leg-spin and googly means that if a batter simply refuses the slog-sweep, his wagon wheel contracts — his DPI is a function of the opponent's decision more than his own execution. That is the central difficulty of measuring pressure in T20: so much of what follows a delivery is decided by the batter, not the bowler, unlike a defensive action in football. Mustafizur Rahman shows the venue dependence plainly. Where the surface grips, his cutters put him in the top quintile of DPI; on a slick pitch the same bowler slides into the fourth quintile. Same cricketer, two faces, and the difference is made by conditions. Valuation follows from there. In my model, runs saved per match through DPI are multiplied across a 14-match franchise season, then translated through win-probability sensitivity into playoff probability, and from there into revenue. Root: Occupation as Transfer Market Administrator means I spend my days watching how much of a fee comes from data and how much from an agent's narrative. Put DPI quintiles beside auction base prices and the spread is widest in the middle group — which is where the inefficiency lives. That is exactly where my suspicion sits. DPI does not measure a bowler's quality in isolation; it measures the joint output of bowler, fielders, captain and pitch. Move the same bowler behind two different slip cordons and the same delivery produces two different numbers. Second, the death-over sample is brutally small. A bowler may send down 60 to 80 death balls in a season, a thin basis for a quintile. I therefore carry a confidence band of ±0.22 DPI and never present a price built on it as settled. Third, gym culture. The yorker-machine nightmare is young bowlers abandoning slower balls, cutters and length variation to rehearse pace and a low line; DPI may tick up while bowling intelligence drains away. Fourth, analysts in dressing rooms. A number describes what already happened; it cannot describe what is happening in a bowler's wrist before release. Root: Transfer market administrator plus Data Monk archetype keeps the final ten percent of the decision with the bowler. So what is my pre-registered call for the next window? In my notebook there is a bowler whose powerplay data sits in the top quintile, economy under seven, but whose death-over DPI is second quintile on a sample of only 41 balls. He will certainly be priced in the top five at auction. My valuation range puts him outside it, because the confidence band around 41 balls cannot justify a 1.5 million premium. The trigger is simple: if he bowls at least 90 death balls across his next 15 matches and holds a DPI of 0.6, I revise the range upward, write the date on it, and log the error if I am wrong. From the administrator's chair certainty is not available. A calibrated published price is.