The Dataset of Negative Results: Finding Bangladeshi Cricket's True Signal in Khulna's Uncounted Matches

A first-class scorecard stopped me in my tracks. In November 2026, a young left-arm spinner bowled 34 overs in a National Cricket League match at Khulna's Sheikh Abu Naser Stadium, 27 of them maidens. But that data was never recorded. There is no ball-by-ball log in the BCB official database, no video, and the dressing-room scorebook itself sits gathering dust in a file cabinet at the Khulna divisional office. I collected that scorebook by hand, because I know that Bangladeshi cricket's real signal lives in these uncounted matches, not in the Mirpur press box. The story of Bangladeshi cricket is usually a story of heartbreak. The 2026 World Cup loss to Afghanistan, the 2026 T20 World Cup exit in the Super Eight against South Africa- these narratives suggest our cricket is a cyclical tragedy. But that narrative is written before the evidence arrives. My job is to avoid conclusions without proof. Since 2026, I have been collecting data by hand from domestic and age-group cricket in Khulna, Rajshahi, Bogra and the Dhaka leagues. Over this time I have built ball-by-ball logs for over 470 matches, 210 of which have no official scorecard at all. This dataset is my real reporting. The problem is that this dataset has a systematic flaw. The matches I have collected are few in number and geographically limited. But that flaw is itself a datum. Around 65 domestic matches are played in the Khulna division each season, yet only 18 of their scorecards are uploaded to the BCB online portal. The results of the other 47 appear only in a single paragraph in local newspapers. This means 72 percent of Bangladeshi domestic matches have no digital record. Without filling this gap, we can never properly understand how our cricketers are made. I recently ran a test. Between 2026 and 2026, I collected ball-by-ball data on 28 young cricketers from the Khulna division, 19 of whom have not yet played for the national team. Analysing their bowling workloads, I found their average overs per match was 14.3, 28 percent higher than the national team bowlers' average of 11.2. This extra load raises their injury risk. But the problem is that this finding is based only on my incomplete dataset. I do not know how rigorous the age-verification process is for national-team bowlers, or how different their physical trajectories are. So I can say cautiously that this difference is a signal, not a conclusion. A big part of my research is the negative result- what is not in the data is my data. Take a 2026 example. A right-arm pacer bowled 12 overs for an HSC team in Khulna, and did not bowl a single ball below 130 kph in any over. There is no video of that match, no speed gun. I spoke to a local spectator who said the pacer was 'throwing fire'. But a spectator's memory is not evidence. So I collected the scorecards from his next three matches- he took 4/28, 2/35 and 0/41 respectively. His average pace had dropped across those three matches, but there is no data to prove it. That gap is my most valuable datum. The numbers were not lying; they were waiting for a better question. When I started this method in 2026, my hypothesis was that spinners' average economy rate in domestic cricket would be better than pacers', because Bangladeshi pitches are slow and spin-friendly. But analysing the data, across 280 Khulna division matches between 2026 and 2026, spinners' average economy was 3.42, pacers' 3.38. A difference of just 0.04. That difference is not statistically significant. My hypothesis was wrong. I published that error, because the first condition of the scientific method is admitting error. If the data says my hypothesis is wrong, I accept it. My method resembles Munich-based football analysis. Just as German football analysts use pass-by-pass data to analyse team systems, I use ball-by-ball data to analyse cricketers' roles. The heatmap is the new astrology- it hides a player's real role. For example, a spinner's heatmap shows he bowls more length balls, but it does not say what he does under pressure. Using ball-by-ball logs, I found that of 18 spinners in Khulna, 14 bowled 23 percent more short balls on average in their third spell. That datum cannot be seen on a heatmap. My biggest challenge is the contrarian reflex. This habit makes me always want to find something opposite to the popular view. But in some cases the popular view may be right. In 2026 I analysed whether the number of fast bowlers in Bangladeshi domestic cricket was declining. Comparing 2026-2026 and 2026-2026, I found that in fact the number of 135+ kph bowlers has remained stable. I published that 'boring' finding, because when the data says something, I say it. The biggest limitation of my dataset is that it is limited to the Khulna region. I have not been able to collect complete data from Dhaka or Rajshahi. So any of my conclusions apply only in the Khulna context. I always disclose this, because reaching a conclusion without knowing the limitations is deception. I can say that spinners' workload in Khulna domestic matches is 19 percent higher than the national average, but I cannot say that is true across the country. A major problem is that the actual peak curve of Bangladeshi cricketers does not match the peak curve of SENA countries. In Australia or England, a pacer's peak usually comes at 27-30 years of age. But in Bangladesh our pacers play the most matches by 23-25, because the domestic calendar has more matches and less rest time. This mismatch is a major structural problem. As of 2026, a pacer in Bangladeshi domestic cricket plays an average of 32 first-class matches a year, while in Indian domestic cricket the number is 21. Those extra 11 matches mean about 300 extra overs of bowling. What is the effect? My dataset shows that among Khulna pacers under 25, 41 percent suffer at least one injury in a season. I do not want to write any 'redemption arc'. 'Bangladesh is finally turning around' or 'we always find a way to lose'- both narratives are written before the data arrives. My job is to stay outside these narratives. Before reaching any conclusion, I state the hypothesis and the expected result. Then I collect data. Whatever the data says, I publish it. The next Khulna match this season will be played on 21 December 2026 at Sheikh Abu Naser Stadium. I will collect the ball-by-ball log of that match. My expectation: no single pacer will bowl more than 23 overs in the first two days. If that is not the case, I will publish that data, and that data itself will create a new signal. Because in Khulna I learned that silence is also a dataset.