Death-Over Leverage: The Model That Measures T20 Pressure — and the Auction's Wrong Price

At 18.4 overs on a March night, a spinner with a tournament economy of 6.2 — among the league's best — stood at the top of his mark. The chasing side needed 38 off 24. The stands were quiet, the run-up a shade slower. But the number burning in my notebook was not economy. It was leverage. Before that single delivery, the match's leverage index stood at 4.1 — that one ball carried roughly four times the weight of an ordinary over's ball. Everyone saw the result: a full toss, a six over long-on. The match did not turn on that ball alone, yet my model said it was a pattern, not an accident. A pattern the auction is not pricing. In football, PPDA — passes per defensive action — measures pressing intensity. That number does not sit directly on cricket, because cricket has no contest for ball possession. But the logic survives if you swap the variables. What is a pass in football becomes a ball in cricket; what is a defensive action becomes a dot ball plus wicket probability. I built the xG notebook precisely to see which truths survive the math and which only survive the noise of the stands. In a T20 regular season, the question matters more. Group-stage tables rarely reveal real strength, because one or two results move the table without moving the underlying ability. The true strength of a mid-table side hides not in match reports but in the pressure moments. I began scoring those moments separately and called it the Death-Over Pressure Index, DOPI. My first rule was strict: judge no bowler by aggregate economy. Economy is an average, and an average does not know pressure. A bowler with a 6.2 economy may have conceded 4.1 in the powerplay and 9.8 at the death; collapsing both into 6.2 hides his real skill. That is the first information gain: in T20, a bowler's skill is not a fixed number — it is a function of pressure. DOPI works in three layers. The first is leverage index: how many win-probability points a ball can swing, which is its weight. A wicket in the 19th over can move win probability by 30-40 points; a single in the eighth moves it by 3-5. The gap between two balls' weights is tenfold. Economy treats them as equal — that is the core error. The second layer is economy delta: a bowler's economy at high leverage minus his economy at low leverage. In my two-season sample of roughly 240 death overs, bowlers with an average leverage of 3.0 or higher posted an economy of 9.4; the same bowlers in low-pressure overs posted 8.1. Pressure costs about 1.3 runs per over. Some paid 2.5. Re-ranking the top ten by raw economy with DOPI shifts six positions. The third layer is matchup: bowler's hand, batter's hand, and phase. A leg-spinner against a left-hander at the death is one equation; an off-spinner against a right-hander is another. In my notebook, one leg-spinner's death economy was 7.9 against right-handers and 10.6 against left-handers. The captain still bowled him to the left-hander, because that was who was at the crease. The raw number blames the bowler; the fault lies in the decision. I also measure the wicket-versus-dot trade-off separately. At high leverage, a wicket and a dot ball are not equal in marginal value. A wicket brings in the next batter, who carries his own risk; a dot ball only burns a delivery. So DOPI weights a wicket above a dot, but the ratio is phase-dependent. At the death, the wicket's weight peaks. The same model runs for batters. The gap between raw death-overs strike rate and high-leverage strike rate often reaches 40-50 points. A batter who strikes at 150 overall but drops to 110 in the last five overs is called a finisher only by habit. The finisher is whoever holds strike rate under pressure; the name matters less. Everyone knows the names at the death — Jasprit Bumrah, Rashid Khan, Wanindu Hasaranga. But look at their leverage splits, not their names. Now turn to the auction. Franchises bid on strike rate and economy. But the marginal value of a death bowler who saves 0.5 runs under pressure exceeds that of a middle-overs bowler who saves 1.5 in low leverage, because the win-probability pool differs. In my model, two bowlers with identical raw economy but DOPI of plus 1.2 and zero should differ by a 25-30 percent premium. The market still does not pay it. That is the inefficiency. From the transfer market administrator's chair, the arithmetic gets harsher. If a side's death-overs budget is limited, which bowler do you fund? Raw economy does not decide it; DOPI does. Last window, one side released its most expensive death bowler because his raw economy was poor — even though his DOPI was the league's third best. The captain had bowled him in the hardest overs. The side punished the number, not the culprit. At the 2026 World Cup I tracked France's PPDA at 12.4 and Kylian Mbappe's 0.18 xG per shot, and learned there that a tournament sample is a pricing laboratory. Applied here, the logic says: the death-overs sample is small, so two or three sixes in one tournament cannot judge a bowler. In my 2026 empty-stadium study, home advantage fell by 0.27 goals; the lesson was about sample and context. Cricket's death-overs sample is smaller still, so I attach a confidence band to every DOPI rating. Here lies the danger. DOPI measures a correlation, not a cause. A bowler's high-leverage economy can worsen because the captain bowls him in bad conditions — short boundary, wet ball, into the wind. The number then shows the bowler's fault and hides the decision's. When analysts walk into the dressing room with clean tables, those tables detach from the match's actual rhythm. Measuring pressure and understanding pressure are not the same. Just as gegenpressing in football has been solved by mid-table sides through athleticism, T20 death-overs pressure is often solved by raw physical capacity — the yorker, the slower ball, the strong wrist. Body beats brain. My worry: if a DOPI-style metric becomes the auction's main tool, sides will buy the athlete with the clean number, not the intelligent bowler who reads the situation. The metric then picks the bowler, not the captain. So what I will watch next window is clear. Bowlers whose raw economy exceeds 8 but whose DOPI is positive should be cheap — that is my buy list. And sides that can separate captaincy decisions from bowler skill will buy more return for less money. I will publish no death-overs rating without a leverage index. My pre-deadline forecast: in the next auction, at least three death bowlers will earn better than their raw economy implies, and at least one worse — on DOPI alone. I have kept the arithmetic in the notebook. At season's end I will check whether the math held, or the stands did.