The Silent Arithmetic of Death Overs: What a Bowling Load Ledger Says, and What It Refuses To

Last month I was sitting on a balcony in Rangpur watching a T20 match. The seventeenth over was on screen. The bowler sent down four cutters, two yorkers and a slower ball; the over cost two runs. The commentary box erupted. I wrote nothing that night. I have a rule: no notes before the match ends, and not even on the night after it ends. The next morning, tea in hand, I went back six overs and found that the same bowler had conceded forty-one runs in his previous six, and that his average pace in the first two overs of the spell was 3.4 kph slower than in the later ones. One brilliant over, six ordinary ones — and the headline went to the over. My ledger points the other way.

This piece is about that ledger. Midway through a regular season we look at the table — points, run rate, net run rate. The real signals of a season are built quietly underneath it: bowling load, back-to-back spells, a fielding position drifting two steps, and an umpire's invisible patience. The things that are not headlines today are the headlines in two months.

Context: the notebook, the rule, and the ten-match gate

In 2026 I was twenty-two, a student in Rangpur, logging every shot of the Bangladesh Premier League by hand. After Abahani Limited Dhaka drew 1-1 with Sheikh Russel KC I worked out that Abahani's run expectancy was 2.7 against Sheikh Russel's 0.6. That is when I wrote myself a rule: no claim without ten matches of data. The rule made me slow. It also made me trusted.

In cricket that rule is harder to keep than in football, because cricket samples shift fast. In one innings a batsman faces six different bowlers, the pitch does not behave the same at ball one and ball 120, and dew in the second innings overturns the whole calculation. In football I measured pressing with PPDA; my cricket equivalent is dot-ball pressure percentage — what share of a bowler's deliveries forced a genuine stroke, and what share merely forced a defensive push. In the regular season that distinction is the most valuable thing there is. In a knockout, teams pick their best eleven; in a regular season they pick who to rest. That choice is the real data.

Core analysis: load, pressure and performative running

Over the last two seasons I have kept a hand-written ledger of sixty-eight T20 matches, domestic and international. For every spell I record four things: overs bowled, dot-ball pressure percentage, boundary-conceded rate, and average pace in the first and last overs of the spell. Put the four columns together and the answer is not comfortable.

First finding: death-over economy alone tells you nothing. Of the five bowlers with the best death-over economy last season, three bowled those overs slower than at any other point in the season. The success came from reducing pace, not raising it. Mustafizur Rahman's cutter is an established weapon in international cricket, and Shakib Al Hasan has changed pace for years — he is Bangladesh's leading wicket-taker in both ODIs and T20Is. What commentary calls courage, the ledger usually calls restraint.

Second finding is worse. Bowlers who sent down the most overs in a season saw their economy rise by an average of 0.7 runs the following season. But the trap is here: the relationship is not linear. Those who played four matches in a row declined sharply; those with a two-match break in between showed almost no decline at all. The damage comes from density, not from the total.

Third finding concerns fielding. I began recording how far fielders move from their starting positions on every dot ball. The interesting part: the sides known statistically as the best fielding units cover more ground on average — but a large share of that running is fruitless. It is exactly like football, where distance covered produces beautiful numbers while opening the defence. In cricket the biggest victim of fruitless running is the last four weeks of a regular season, when the body is tired and the mind is already on holiday.

Fourth finding is about home conditions. In 2026, when stadiums went quiet, home advantage lost its voice — across eighty-three Bundesliga matches the home win rate fell from 43.3 percent to 33.1 percent. When crowds returned to cricket, I reconciled my domestic logs and found that home teams' win rate came back, but their death-over economy did not. The crowd returned, but the second of breathing space before a bowler releases the ball did not. That small gap is what a home side hides best in a regular season.

Fifth finding is about pitches. I have tried to reconcile the language of pitch reports with actual ball turn, recording the difference in turn between the first ten overs and the last ten since the 2026 season. About a third of pitches announced as spin-friendly actually turned less in the second innings, because they were slow pitches, not turning pitches. Slow and turning — we use the two words together, yet they are separate weapons.

Sixth finding is about the toss, and it is my least favourite. Everyone knows the toss-and-chase decision correlates with winning in domestic leagues. My ledger has that correlation, but it is weak — because the teams that chase well are usually the better teams, and better teams choose to chase when they win the toss. The toss here is an effect, not a cause. The real signal is when the dew arrives. Over two seasons I have logged how often the ball's grip had to be changed in the twentieth over. In matches where it changed more than three times, second-innings death-over economy rose by an average of 1.8 runs.

Seventh finding is in batting, where the data is thickest and the patience thinnest. Everyone praises strike rotation in the middle overs. My ledger shows that sides which raised strike rotation purely by cutting dot balls scored more in the last five overs — but only when two set batsmen were at the crease. Without them, the same patience becomes stagnation. The ledger does not lie, but the ledger alone does not tell the truth either.

The contrarian case: correlation is not causation

This is where I have to argue against myself. Every relationship above has a question hanging beside it, and honesty means admitting it.

For bowlers who succeeded by dropping pace at the death, the question is whether they reduced pace by design or simply bowled the overs in which the batsman was weak. Sort the data by bowler and I get one answer; sort it by batsman and the answer changes. The overs are few, the opposition quality is uneven, and separating cause from outcome here is close to impossible.

My second objection is aimed at myself. I like the ten-match rule, but ten matches is not always enough. If a bowler bowls six death overs a match, ten matches means sixty overs — fine. If he bowls two, ten matches means twenty overs, and over twenty overs luck and skill weigh roughly the same. I now run two thresholds: at least thirty overs for an individual bowler, at least ten matches for a team pattern.

The third objection is league-specific. Domestic pitches, balls and fielding standards differ from international ones. Moving from a domestic ledger to an international call, I add one variable — context adjustment. Without it we simply write our favourite story into the notebook.

The biggest objection is about my own trade. I work in a world where numbers build forecasts, and those forecasts end up imprisoned inside their own model. A model is a confession, not a prophecy. Forget that and the distance between the ledger and the truth keeps growing.

Forward signal

For the next five matches of the regular season I am watching three things. One, how many death-over bowlers are being used in the same role three matches running — density is the real danger, not the total. Two, how much spinners' boundary-conceded rate rises in the second innings compared with the first — that is where slow pitches separate from turning pitches. Three, how far fielders drift from their starting positions in the middle overs — if fruitless running keeps climbing, the win column will look fine for a while and then turn suddenly.

I recalibrate because the world does, not because the model is fashionable. The notebook is still open, the pen still uncapped — but before the next innings begins I want an answer to one question: are we measuring a bowler's ability, or his luck?