The Middle-Over Wicket Tax: Why T20 Models Overpay for the 17th Over

On 22 June 2026, at Arnos Vale in St Vincent, Afghanistan batted first and made 148 for 6 — in the current grammar of T20 cricket, a score usually described as already lost. Thirteen overs into the chase, Australia were 106 for 3, Glenn Maxwell and Marcus Stoinis at the crease, needing 43 from 24 balls. The win-probability window on my laptop had Australia at 78 per cent. The scoreboard and my model were saying the same thing in the same voice. Then Gulbadin Naib took the ball, and nine overs later Afghanistan had won by 21 runs. I started with the expected run, not the final score. The expected run told the truth that night, but only half of it. The column my model was discounting most heavily was the wicket column between overs seven and fifteen. Context: How the model gets built, and by whom In April 2026 I started a one-man newsletter called The Expected Goal from the back room of a Fitzroy share house. Sydney FC had just created 1.94 xG against Melbourne Victory's 0.61 and drawn the match; I posted a chart at two in the morning and 300 people opened it. By December there were 4,200 subscribers. The share house taught me every dataset has a kitchen table — who is cooking, for whom, and on whose money. Cricket's table is more crowded. Football has expected goals; cricket has a whole family of proxies — phase-based par scores, venue-adjusted expected run rates, batter-bowler matchup matrices, ball-by-ball win probability. Every one of them uses the past to describe the future. Every one of them is built by people who hate being wrong. My own roots sit awkwardly inside that. I opened the batting and kept wicket for Udity Club in the Dhaka league in 2026, where the only accounting that mattered lived in the scorebook. Later, in Melbourne suburban cricket, I found a different accounting: who was willing to take risk, and who had the licence to. Two club cultures, two appetites for risk. The model does not measure the appetite. Core: The death-over premium and the middle-over tax The arithmetic is straightforward. In T20 markets, a ball bowled in overs 17 to 20 is priced at roughly 1.7 to 1.9 times a ball bowled in the middle phase. The reason is obvious — sixes live there, boundaries live there, the scoreboard jumps. But the wicket is priced on a different schedule, and that is where I object. I sat with my own sample of T20 internationals and tournament matches since 2026. Teams that took two or more wickets between overs seven and fifteen won better than five of every seven matches in that period. I checked the inverse as well: for sides that failed to take a wicket in that window, whether the death overs produced boundaries made far less difference to the result than the commentary suggested. I am not mistaking that for causation — correlation is not causation is not a slogan, it is the first line of my job description. Teams that are ahead set attacking fields and take middle-over wickets; teams 60 runs behind also attack and lose wickets. Reading victory out of a wicket count alone is dangerous. But one thing is reasonably clear: the market prices death-over risk very well, and it prices middle-over wicket risk with assumptions that are close to a decade old. The matchup matrices show the same gap. Spinners generally strike in exactly that window. Wanindu Hasaranga was the leading wicket-taker at the 2026 T20 World Cup with 15, most of them in the middle phase. Maheesh Theekshana's action and pace force batters into strokes in that same phase, and the over is gone before they reset. Sri Lanka's bowling unit has done this section of the game well for years. Australia, by contrast, has been searching for a second spin option behind Adam Zampa, because their wicket-taking structure runs on fast bowling and death-over yorkers. And that is the death-over muddle. At the 2026 ODI World Cup, on 15 October in Delhi, Afghanistan beat England by 69 runs. They did not win it with sixes. They won it by breaking England's middle order in the middle overs — a 50-over match and a 20-over match running on the same engine. I sit with the numbers until they confess their bias. This dataset's bias is simple: the ball that goes for six in the 18th over stays on television, and the ball that hits the pad in the 14th over never reaches the highlights. The information leans towards what can be seen. The franchise market is folded into the same problem. In the IPL auction a 34-year-old fast bowler sold for 24.75 crore rupees, and in the next auction the same bowler went for 11.75 crore — the price tracking visibility more than performance. Leagues that borrow a city's name while their ownership and audience sit somewhere else slowly move their real investment away from quiet places like the middle overs and towards the loud ones. The market is growing. The middle-over ledger is still barely being written. Contrarian: Information leans towards what is easy to see I am not arguing that middle-over wickets matter more than death-over sixes. I am arguing that our instruments are weighted unevenly. T20 reinvents itself roughly every eighteen months — a new ball, an impact-player rule, new dew patterns, new fashions in the slog sweep. A model trained on an older format commits confident errors in a newer one. One more thing needs saying. Sri Lanka is not one thing — the club grounds of Colombo, school cricket in Kandy, and the Sunday tape-ball games of the Sri Lankan community in Melbourne are three different cricket economies. Sri Lankan pitches are not Indian pitches, and daylight is not a dew-soaked night. Applying one national model to a country is imagining a dataset without its kitchen table. Takeaway: What to watch next tournament The 2026 T20 World Cup runs from 7 February to 8 March in India and Sri Lanka. The thing I will watch most closely is not a squad list. It is the wicket column from the 12th to the 16th over — who brings their best bowler back into that window and who does not, and who takes wickets there rather than merely containing. Your team's scoreboard may show you the sixes. Those five overs are probably where the match was actually priced. What do you want to watch — the six in the 18th over, or the wicket in the 14th that stopped it?