Blank Cells and Middle Overs: What a Hand-Coded Model Taught Me About Bangladesh's Asia Cup Silence

I opened one scorecard three times during the last Asia Cup. First with my eyes, then with over-by-over run rates, and finally with a blank spreadsheet. The scorecard itself was almost clean — a fifty-plus opening stand, 45 for 1 in the powerplay. Then, between overs seven and fifteen, Bangladesh made 58 runs and lost four wickets. Across those nine overs, 54 balls were bowled and 29 were dots. The scorecard told me nothing new. The blank cells confessed everything.

What I was hunting for was a number no broadcast graphic carries: runs off the first ball of each over. The first ball of an over tells you who is controlling it. In that middle phase, Bangladesh scored just 19 runs off 154 first-balls of overs — which means three out of four middle-over overs began with a dot. That is not batting failure. That is structural silence. Silence is not zero; it is a new baseline with its own residuals.

I write this because I have seen the pattern before, in a different sport.

In 2026, at forty, I audited rice-mill accounts in Rangpur by day and hand-coded an expected-goals model by night. Not cricket — football. New media was expanding, so I published a 4,000-word breakdown of the Bangladesh Premier League season on a Dhaka football site: 132 matches, 3,410 shots, my own distance-and-angle weights because no public xG then existed for that league. Abahani Limited's title run showed a 9.4-goal gap between my modelled xG and their actual goals. Three betting syndicates emailed me within a week.

That night my writing changed. I stopped writing match reports and started writing methodology notes — every claim now carries its sample size, its weighting choices, and a stated error margin. My sentences got shorter, my footnotes longer. I now split every number into three labels: measured, modelled, or guessed. Without those labels, I cannot make a claim.

That discipline matters more in the Asia Cup, because the format itself makes data scarce. A bilateral series gives you six matches on the same square; information accumulates. An Asia Cup gives you six matches at most, across changing opponents and changing venues — and against associate nations, ball-by-ball data often never reaches the public domain. I opened a blank spreadsheet and let the Bangladesh Premier League teach me the first time; this time I had to map the event structure first, because cricket has no xG. I average overs, phases, venues, and matchups instead.

The model was crude. I admit that. But a crude model's missing cells confess more than the goals do. Three gaps stood out: ball-by-ball data against associate nations, data on Bangladeshi batters against spin in the middle overs, and second-innings dew effects in day-night matches. Those gaps are not accidental. They are mirrors of scouting bias.

Now the numbers. In Colombo, spinners concede roughly 3.2 runs per over in the middle phase while seamers concede 4.1. In Kandy it flips, because the surface is slower and cutters grip. In Dubai, both sit near 3.5. Bangladesh's problem, though, is not run rate — it is boundary rate. Their middle-over dot-ball percentage is near par, but their boundary-per-ball rate is roughly 25 percent below peers. They are not wasting balls. They are also not using them.

My eye said Bangladesh bat too defensively in the middle. My modelled run-value index said something else: they rotate strike almost every ball in that phase. The real deficit is roughly two fewer boundary attempts per over. A team that takes singles but cannot take boundaries does not hold momentum — it performs the appearance of holding it. Three singles, a dot, and the powerplay-shaped run rate collapses.

By Russia 2026, I was watching Germany twice: with my eyes and with PPDA. Their pressing had decayed from 8.9 in qualifying to 12.6, a decline invisible on the scoreboard. I wrote before the tournament that their press was already gone. They went out in the group stage. But my model still ranked them third-favourite, so I hedged in the text and lost the argument anyway. That failure gave me the two-track habit: a loud public thesis and a quiet appendix listing every error my model made.

On Bangladesh's middle overs, my appendix now contradicts my public thesis. The public thesis says intent is missing. The appendix says role mismatch. In the BPL, Towhid Hridoy's strike rate between overs seven and fifteen is far higher, because risk-takers surround him. For the national side, that risk becomes his alone — but his development never built that role. The cost of role translation does not appear on any scorecard.

This is not about a lack of squad depth, and not about laziness. We like our narratives black and white, so we call a weak outcome a weak intent. That is unfair and statistically wrong.

Wanindu Hasaranga bowls four to six of his eight-to-ten overs in the middle phase. Rashid Khan, Mujeeb Ur Rahman, the Indian wrist-spinners of the Dinesh Karthik-era setup — all play the same trick: force a single off the first ball, then hold the off-side line to block the boundary. That is a plan, not an accident.

My eye read 'impatience' in the middle overs. My ball-tracking matrix shows Bangladesh defending full-length deliveries almost every ball, with a short ball roughly once in nine. That is not impatience; it is unequal distribution. The real crisis is standardisation, not execution. The team is trying to play the middle overs like a powerplay, and those are two different games.

A caveat is essential here. I know my own bias: an INTP brain loves systems, and a data-monk identity drifts into model worship. So I argue against myself — the low middle-over rate may simply follow early wickets. The batter walking in at four arrives into a different situation. Correlation is not causation.

I tested it against base rates. In recent Asia Cups, when Bangladesh's opening stand falls under twenty, middle-over run rate averages 4.2 per over. When the opening stand passes twenty, it rises to 5.1. The gap looks small, but on a three-to-four-match sample it is vast. I cannot claim this is purely a batting problem, because boundary rate is shaped by the opening platform, the opposition's spin distribution, and venue pace.

So I returned to my crude model, just to look at the blank cells. My middle-over spin segment is incomplete — some matches were never tagged for spin, and some powerplay overs were tagged as spin when they were not. That tagging error itself tells me we must look at our information system before passing judgement on Bangladesh's middle-over tactics. We are collecting description, not cause.

This is where my database experience helps. I began looking at the ball that was never bowled — how often a spell's opportunity existed but the bowler never got it. The same missing cell runs through Asia Cup data: how many Hasaranga or Rashid spells fell in the middle overs and how many did not. That unknown number drives Bangladesh's plan more than any visible one.

My long-run position is this: under tournament pressure, the language of calculation loses its purity and the language of patriotism loses its patience. Judgement has to sit between them. A single six or a single dot ball, judged alone, is a lie. Only ball-by-ball context sustains it.

I still trust what I see — after checking it. In Bangladesh's middle overs, one thing is visible: batters stay at the crease, play the ball, rotate strike, yet over after over the scoreboard stalls. The eye says lethargy; the scoreboard says otherwise; the spreadsheet says the most.

Next round I will not watch boundaries. I will watch one number: Bangladesh's first-ball runs between overs seven and fifteen. If that figure climbs to 0.7 to 0.9 per over, a big total becomes plausible. If it does not, the biggest question becomes simpler — when will this team redefine its own middle-over role on international pitches? Nobody else will fill in that spreadsheet. Only the team can.