The Dot-Ball Mirage: Why the Scoreboard Lies in T20 Middle Overs

On the fourteenth over of a T20 match last season, one number stuck in my notebook. The chasing side was 89/2; between overs seven and fourteen they had scored 52 runs and lost a single wicket. The commentary box said they were building a platform, that the explosion would come in the last five overs. The scoreboard said the story was under control. My ledger said the opposite: 41 dot balls in those eight overs, and only four boundaries. They collapsed to 34 in the final five and lost by 22 runs.

That night I started a separate ledger for middle-overs dot balls. The scoreboard tells one story; control of a match tells another, and confusing the two is one of the most expensive mistakes in T20 cricket. The first xG ledger began as a private argument with the scoreboard. This is another instalment of the same argument.

The ledger the public record does not keep

T20 analysis in our region has a specific obstacle. For Bangladeshi and Sri Lankan domestic cricket, ball-by-ball and over-by-over public data is not as easy to obtain as in the western leagues. Franchise scorecards exist, but who played how many dots, where the field stood in a given over—that information is scattered, incomplete, and often survives only in local commentary memory. So I had to build the ledger myself.

While watching, I note four things by hand: runs per over, wickets per over, dot balls per over, and boundaries per over. Then I split the raw data into three phases: powerplay (1–6), middle (7–15), death (16–20). Each phase yields four metrics—runs per over, wickets per over, dot-ball rate, boundary rate.

I did not trust the table until it survived a season of variance. A single match fluctuates; only a number that holds across a season earns the right to change a decision.

Why the middle overs? Because the quietest eight or nine overs of the game live there. In the powerplay the field must come in; at the death everyone attacks. But between overs seven and fifteen the scoreboard barely moves, and that is exactly where the real clock of the match runs. The lesson I learned in football arrives in cricket disguised as a dot ball: possession is not control. Spain completed 1,029 passes, and the goal disappeared into the possession.

What the dot-ball table does not say

Across two seasons my ledger shows a clear pattern. The sides with the highest middle-overs dot-ball counts are not automatically the sides with the best win rates. Over a full season, the relationship between middle-overs dot-ball rate and win rate is weak—yet we force the two together by sheer insistence. Strong bowling attacks do produce more dots, which is true; whether those dots cause wins is a separate question.

Here is the real trap. A side that squeezes the middle overs usually has a large reason for it—its top order has already posted a big score, so the bowler can afford an attacking field. The reverse happens too: a side 30 runs behind forces its batters to take risks, so its dot count falls while its wickets climb. The dot-ball rate is often a result of the scoreboard situation, not its cause.

The anchor illusion

The most celebrated character of the middle overs is the "anchor"—high average, rarely dismissed, strike rate around 120. Commentary calls him a steady hand, a foundation stone. My ledger calls him something else.

One domestic season: a top-order batter played 16 innings, averaged 47, but his middle-overs strike rate was 118. For every 100 balls he faced, the team got 118 runs, while the league's middle-overs benchmark was 131. In every innings he cost his side somewhere between 9 and 14 runs, and the security he offered was worth no more than that. At the end of the match a handsome score sits beside his name—that is what commentary reads. What the scorecard never prints is the damage built inside the innings.

Score 47 off 40 balls and everyone calls it a success. But if 22 of those 40 balls are dots, and the team loses five wickets in the last five overs trying to repay the debt, then the 47 on the card and the result of the match are two entirely different events. That is why I keep a second number beside a batter's average: his middle-overs output per over. The average hides what the output reveals.

The economy-spinner trap

The bowling side shows the same picture. The spinner who finishes with 6.8 runs per over becomes the emblem of "control" in commentary. My ledger has him at 0.6 wickets per match. In T20, a bowler's final value is ultimately measured in wickets, because a middle-overs wicket brings a new batter to the crease, and a new batter naturally begins slowly.

In my calculation, between a spinner at 6.8 an over and a spinner at 7.4 an over who takes 1.3 wickets a match, the second wins his side more matches across a season. The first bowler's dots come from safe fielding, not from wickets; a strong batting line-up plays him out, and the death-overs burden then lands on someone else's shoulders. Wicket-taking legspinners in the mould of Rashid Khan or Wanindu Hasaranga are precisely the bowlers who deserve a separate column outside the economy table.

Auction prices and on-field accounts

Another calculation sits beside this one, outside my ledger entirely—the auction price. A batter who has not yet played fifty top-level matches can go for a fortune. That price does not come from a season of data; it comes from one innings, one viral clip, one expectation. The risk football carries when it pays a hundred million euros for a player with fewer than fifty appearances is mirrored exactly in a cricket auction. The data monk's job is to see the gap between price and output—when a young talent has no seasonal account beside his name, every match is really an estimate.

And when I read news of player rest, I always ask one question: is the rest protecting a body from injury, or arranging a commercial calendar? Workload management is a beautiful phrase, but in my ledger it is often a synonym for the tour schedule. I learned that lesson while matching fitness data off the field, and it is visible in the shape of bowling changes.

Where the scoreboard does tell the truth

Over two seasons, two numbers correlate best with winning, and both are deeply unfashionable. The first is powerplay wickets—how many wickets fall between overs one and six. Each extra powerplay wicket raises a side's win probability markedly. The second is the opposition's boundary rate at the death—how many boundaries are conceded per over between overs sixteen and twenty. A side that grips both ends of an innings that way can afford a quiet middle and still not lose the match.

And where do we look instead? At the middle-overs economy table, because the numbers there look tidy. Variance does not care about your narrative. So in my previews I keep two columns: one for territory, one for danger. Territory is who faced how many balls; danger is who scored how many runs.

Lessons that came from Burnley

My first ledger came from Burnley's seventh-place finish in the 2026-18 English Premier League. They took 54 points but their expected points were 45.1; they conceded 39 goals against an expected goals conceded of 49.7. Burnley was a mirage, and xG kept the receipt.

In cricket, the middle-overs dot ball sits exactly where Burnley sat. What actually happened goes in one column; what should have happened goes in another. Otherwise we sell a stroke of luck as strategy, and then act surprised when it breaks the following season.

Empty stands and the lesson of context

In 2026, when sport stopped, I was modelling the effect of empty stadiums. In the Bundesliga's May restart, the home win rate fell from 43.3% to 33.8%, and home goals per game dropped from 1.74 to 1.29. That lesson entered my cricket ledger too: no number is neutral, and context is welded to it.

In cricket, context is named venue, toss, the length of the break, travel, and crowd. The same bowler's dot-ball rate differs at home and away; the same batter's death-overs strike rate shifts with the toss. So before comparing a dot-ball column, it has to be stratified by venue, phase, opposition quality, and the conditions of the day. Otherwise we merge two kinds of dot ball into one—one the product of an attacking field, the other the product of a safe one.

The danger of the counter-reading

Now the most dangerous spot. Everything above has an inverted reading—and that reading is also wrong. Throwing dot balls out entirely is just as mistaken. The problem is cramming two different kinds of dot ball into one cell using a single number.

So I use notes built on what I actually saw: where the ball went, what shot the batter attempted, whose decision produced the dot. This note-taking has deliberately slowed my conclusions, because my ledger has a hard rule—any contrarian claim must first beat a simple base-rate model. Base rate means the ordinary accounting of venue, toss, and team strength. I ran the middle-overs dot-ball rate against that base-rate model; for explaining wins, the column added almost nothing.

That is the difference between correlation and causation. Dot balls do not win matches; the state of the scoreboard, an attacking field, and wickets do. The dot ball is often the shadow of those three, and we mistake the shadow for the cause.

The signal for the next round

In the next round, keep two columns in your ledger: powerplay wickets, and the opposition's boundary rate at the death. Cover the middle-overs economy table. Then ask one question—which spinner takes wickets while the field is up? Whoever does is the real control. The rest is noise, and a handsome print on the scoreboard.