The Hidden Channel at the Death: Twelve Zones and the Making of a Bowling Decision

Over my last three matches I coded 87 death-over deliveries separately — each one's release point, line, length, and the instant the batter moved his feet. One ball is still stuck in my eye. The fourth delivery of the nineteenth over: the bowler bowled a full yorker, and it still went for four. On the stump camera the delivery looked perfect. But just before release the batter's front foot had moved to the left, and his backlift was finished before the ball arrived. I went frame by frame and found that the previous two balls had landed in nearly the same channel. The batter had memorised the zone. This is not a question of the bowler's skill; it is a question of repetition. A large share of runs conceded at the death comes from channel repetition, not from a missed yorker. The broadcast stops right there. The camera shows the ball, not the channel. In a coaching staff's eye the delivery is therefore not the end but the beginning — which zone was chosen before the ball was bowled is the real information. The T20 death over is really a game of space. Between the sixteenth and twentieth overs a bowler has thirty balls, but very little room to arrange them, because once a batter reads the line, every variation looks the same. These five overs usually produce 45 to 55 runs in a match, and that is where most results are decided. In 2026, working with Sheikh Russel KC, I built twelve pitch zones around set-pieces, coding 47 corners and 31 free-kicks. The database had twelve zones before anyone asked for one. Later I pulled the same structure into cricket. For a right-arm bowler the twelve death zone can be laid out as: wide yorker, stump yorker, leg-stump yorker, top-of-off, hard length, slower length, wide slower, cross-seam, chase-the-bend, leg-side blockhole, bouncer, and body-line. In cricket, one of Mustafizur Rahman's cutters falls into two of these zones; Taskin Ahmed's pace sits in another. To know the zone is to know the bowler. The table is easy; the decision is hard, because every zone has a cost and that cost depends on the environment. With dew, a wide yorker loses its grip and slips out of the hand. In an empty stadium a slower ball works less, because the batter, spared the noise, simply sees the ball longer. I coded empty stadiums until silence became a coordinate. I never seal a trend until I can place at least three prior matches in the precedent table. The precedent table does not predict; it remembers. Now the real arithmetic. Of the 87 death balls coded across three matches, zone repetition occurred in 29 cases — that is, on two or more consecutive balls the bowler chose the same channel. Those 29 balls produced 41 runs, roughly 1.41 per ball. The other 58 balls, where the channel changed, produced 63 runs, 1.09 per ball. The difference looks small, but at the death it is the difference. On average two balls per over repeat; four death overs in a match mean about eight balls. That is what turns a seven-run over into a nine-run over, and in a tight match, that is the result itself. I know this is a small sample. So I do not claim repetition is the only cause, but the pattern points the same way in all three matches, and three matches is my rule. If there is an exception, it stays in the file too, because every match leaves a precedent; my job is to file it correctly. Which of the twelve zones is cheapest? My coding says the wide yorker — 19 balls, only 14 runs, 0.74 per ball. There is a trap here. Fourteen of those 19 came with a right-handed batter standing on the leg side. Against a left-hander the same wide yorker drifts into his natural swinging arc, and there the cost rises to nearly 1.8. So whether a zone is good or bad is not a property of the zone itself — it is a function of the matchup and the environment. The second thing my coding caught was the entry channel. Bowlers usually set the line they use early in the innings, and often start the death over from the same line. But those who begin the death over directly from a wide-off angle concede less on their first two balls, because the batter has not set his line yet. This is exactly where broadcast statistics and coding statistics split. On broadcast you see economy, yorker percentage, dot balls. In coding you see the order in which the zones are chosen — where the bowler entered, where he stayed, where he exited. The bench sees what the broadcast never shows. The structure is not confined to the death overs. For spinners in the middle overs the same arithmetic holds — if a leg-spinner keeps three straight balls on the same length, the batter is already set up for the cover drive. There the change is not in the zone but in the length. The coding structure is one; only the variable changes. The biggest blind spot sits in coaching decisions, not bowling skill. Teams now chase yorker percentage as though the percentage were proof of success. My coding says the relationship between yorker percentage and economy is weak unless channel variety is added to it. A bowler can succeed with 70 percent yorkers if the remaining 30 percent are scattered across different zones. Another bowls 80 percent yorkers and still gets hit, because the remaining 20 percent is predictable. The question is not how many yorkers, but how many zones, and in what order. One more dimension is needed. I do not finalise a decision without checking it in two different environments. The same bowler, the same zone, in dewy and dry conditions — the result often flips. So any death-bowling claim should carry a counterfactual: would this zone hold without dew? Without asking that, analysis stays only a story. There is a caution here too. If zone-mapping sprawls, every match disappears under a vast table and the analysis weakens. So in each piece I pick three decision-defining zones, not all twelve. The rest stay in the file, unpublished. In the next three matches I will watch one thing — which zone the bowler enters in the seventeenth over, and whether he stays there on two consecutive balls. If repetition falls but economy does not, then my three-match rule was wrong, and that too must be accepted. A precedent table is useful only when it also remembers its own errors. The question in the end is simple. If someone wants to hold the seventeenth over next match, does he raise his yorker percentage, or his zone variety? My arithmetic points to the second. The next three matches will write the answer, and I will file it in the right place.