The Middle-Overs Dot Ball: Why BPL Auction Prices and Match Results Speak Different Languages

Sylhet International Cricket Stadium, January 2026, 9:40 pm. The scoreboard said the chasing side needed 94 off 48. In the twelfth over a part-time off-spinner bowled — a man bought at base price. Facing him was the batter his franchise had purchased for the second-highest sum of the auction. Four balls, four dots. A six off the fifth. The crowd exploded over the six. The match was lost in the four dots.

Back home at 1:40 am I opened my laptop. The number was already sitting in my file: across his previous fourteen innings, that batter's dot-ball rate between overs seven and fifteen was 52.8 percent. Nobody at the auction table asked for it. Nobody said they would. What nobody bothers to buy is often exactly what the result is calculated from. I built a spreadsheet of 412 cricketers that nobody requested, and three seasons later it turned into a witness.

Method: what I count and what I refuse to count

Since 2026 I have kept a file — not an office database but a ledger I compiled myself. Across the 2026, 2026 and 2026 BPL seasons, 138 matches were played. I could verify ball-by-ball detail for 96 of them against two independent sources. The other 42 I discarded because the delivery counts did not reconcile. My ledger now holds 412 players, each with three separate columns: auction price, currency-normalised value, and over-by-over ball accounting for every innings.

My primary measure is simple, and I named it myself: Middle-Over Dot percentage (MOD) — of all the balls a batter faced between overs seven and fifteen, what share produced no run. A second measure follows: Balls Per Boundary (BPB), the number of deliveries a batter spends per four or six. Minimum filter: 120 balls, meaning only those who faced or bowled at least 120 deliveries in the middle phase qualified. That left 187 batters and 143 bowlers.

The two-source rule is strict. If one scorecard called a delivery '1 run, leg bye' and the other called it a dot, I dropped it rather than forcing a reconciliation. Across 96 matches, 397 deliveries were discarded — about 1.1 percent of all balls. Remember the price of that discipline: 397 lost balls and 42 lost matches.

At the 2026 World Cup I logged all 64 matches and 1,912 on-ball events and learned something durable: a narrative can always be written afterwards, but a model written first saves you the trouble. I brought the same method here.

And I write one paragraph against my own model before anyone else can: what this number cannot tell you. MOD cannot say whether the dot came from good bowling or poor batting. It cannot say what the pitch did. It cannot say what the team management instructed, or which way the wind blew in a small ground. My arithmetic says only this: the ball was bowled and no run came. The rest is the video's job, not mine.

Evidence: 6.7 percentage points and one lonely number

Across three seasons, the teams that reached the play-offs account for twelve team-seasons. The bottom four account for another twelve. While batting, play-off sides averaged a MOD of 38.4 percent. The bottom four averaged 45.1. The gap is 6.7 percentage points.

That looks small, so let me break it. The middle phase is nine overs — 54 balls. A 6.7-point gap is roughly 3.6 extra dot balls per innings. The average margin between winners and losers across these three seasons was 14 runs. The opportunity cost of a dot between overs seven and fifteen sits between 1.2 and 1.4 runs, because the big hitters are being held back for exactly that period. Run the arithmetic and 3.6 dots equals four to five runs an innings, and eight to ten across two innings. Two-thirds of that 14-run margin is hiding right there.

The number that decides matches is 6.7. No highlights package has ever shown 6.7. Highlights show sixes. The lonely number sits under the noise, and the noise is always the six.

Now the real question: does the market pay for the 6.7? Splitting my 187 batters into ten deciles by auction price, the most expensive decile posted an average MOD of 41.9 percent. The cheapest decile posted 43.6. A gap of 1.7 percentage points — against a play-off-to-bottom gap of 6.7. The market is paying for roughly a quarter of the difference that actually shows up in results.

The Pearson coefficient between price (log-transformed) and MOD is just -0.11. That is nothing. Between price and boundary rate (fours and sixes per ball) it is 0.34. The market buys boundaries; matches are lost in the silence between them.

Look at the spin-pace split. Across three seasons, spinners bowled 54.2 percent of all middle-over deliveries. Against them the dot rate was 44.8 percent; against pace, 41.2. Teams fielding two frontline spinners visibly cut their middle-over dots — yet in auction budgets, the third spinner routinely sits at base price while top-order batters and death bowlers absorb the money.

The counter-case: the batter who sells the middle to buy the end

One finding cut against my own thesis. Among the batters with the lowest MOD, a large share also had an elevated dismissal rate in overs sixteen to twenty. They consume balls in the middle, preserve the wicket, then take risk at the death and frequently depart. On the other side, batters with high middle-over tolerance often struck above 180 in the last five overs but failed to finish the innings fourteen times.

Which is more valuable is not something a coefficient will settle. Some matches need a foundation, some need an explosion. But auction prices lean sharply one way, because a six sells as a clip and the absence of a dot sells as nothing at all.

Falsification: three findings that would break this story

Before filing, I keep a falsification file.

First, whether the gap survives a control for team quality. In my current estimate, roughly 2.1 of the 6.7 points is explained by team strength; the rest holds. But that 2.1 means an earlier version of this work, where I reported 8.9, was wrong. The number shrank.

Second, Sylhet and Mirpur are different surfaces. In my 96 matches, Mirpur accounts for 51 percent — against 47 percent of the season's total fixtures. Spinners bowl more at Mirpur, so dots rise. Remove Mirpur and the gap falls from 6.7 to 5.4. The thesis survives; it no longer stands loudly.

Third, whether one season manufactures the entire effect. By year: 7.9 in 2026, 6.4 in 2026, 5.9 in 2026. It is present every year, and shrinking every year. Why it shrinks is next season's question.

The people behind the price tag

Unpaid wages were not an outlier; they were the baseline. Of the players I have tracked across these three seasons, at least eleven have said the final month of their contract arrived late. One was a 31-year-old leg-spinner from Chattogram, sold at base price, who that season bowled 41.6 percent dots in overs seven to fifteen. He finished among the top three spinners in the tournament and was paid roughly two-thirds of his franchise's second-most-expensive batter — three months late.

I counted 1,240 empty-stadium matches, and then I counted three months of unpaid wages, and I had to put both numbers in the same file. At the auction table a cricketer's price is his market value. A wages ledger never recognises market value.

The steelman, and where my model is genuinely weak

Franchises are not being irrational. Tickets sell on stardom. A six-hitting batter moves jerseys; a dot ball has no jersey. Overseas quotas, age balance and a captain's idiosyncratic preferences all make the market inefficient by design. And yet one small problem remains: the points table is not compiled from star sales. It is compiled from deliveries.

Dots are also not always the batter's fault. Good length, slow surfaces, seam and swing with the new ball — for spinners especially, this is real. A player dotting at 45 percent is often the beneficiary of excellent bowling. Judging a career on MOD would be incomplete and unfair. I am auditing the buying decision, not the individual.

There is also reverse causality. The team that is already strong buys good players cheaply from weak teams, so cause and effect blur. The more I have tried to untangle this, the weaker my model looks — and I have left that weakness in the text. My sample is also over-weighted toward one venue, which I have disclosed rather than buried.

The spreadsheet was never the story; the silence around it was.

Three signals I'll watch over the next eight matches

First, the middle-over dot count of the top price decile. If it sits under ten, the theory weakens and I will write that down too. Second, the spin usage and dot rate of the table-toppers — and whether the third spinner bought at base price turns out to be the most profitable purchase of the tournament. Third, the timing of dew. If spinners are being bowled out before the tenth over in the first innings, that is a board's instruction, not the pitch's — and it is a limit of my own method that I am conceding before a reader finds it.

The question is this: when the six arrives after four dots, whose hand do we raise? The batter's, or the number's, the one that never reached any table? If someone asks for my file next season, I will hand it over — on one condition. Do not leave the wages unpaid.