The Ten Overs After: Bangladesh's Test Collapses Have an Address the Scorebook Can't Show

Mirpur, November 2026. The fourth wicket fell at 219. In the next ten overs Bangladesh added 21 runs and lost three more. In scorebook language this is a middle-order failure. In my logged database of 41 home Test innings, it is not a single event but a system state. In the ten overs after a wicket, the team's run rate drops from 2.9 to 2.4, and in that same window a wicket falls every 31 balls — nearly twice the innings average of 56. The numbers didn't break the model; they exposed where the model was blind. When I launched Expected Truth from Khulna in 2026, my first model was football's xG. Abahani Limited's 26 matches produced 34 goals against 26.8 xG; I logged PPDA in a 2-0 win over Sheikh Jamal Dhanmondi. Back then I believed that once a scoring framework was in place, the game could be read. The data from 83 empty-stadium matches in 2026 taught me otherwise: you can read the score, but drop the context and you read in the wrong direction. Bringing that lesson back to cricket, I changed the question — does a batting collapse really happen suddenly, or does a measurable window open before it? Bangladesh's home Tests have three characters across three grounds. Mirpur is slow, low and spin-friendly; Sylhet offers brief pace; Chattogram is slow-turn where the ball starts keeping low from day three. From 2026 to 2026, scoring rates in the key sessions of first innings at these venues ran at 2.8 to 3.1 per ball. Runs come slowly whether or not wickets fall. The real pressure is not losing a wicket — it is keeping two new batters alive together. For batters raised on ODI and T20 tempo, this phase is the most uncomfortable, because it demands role clarity and patient strike rotation more than raw skill. To answer the question I pre-registered a simple baseline and capped extra variables. The index is called the Post-Wicket Recovery Index (PWRI). Runs scored by both batters in the ten overs after a wicket, measured against a run-rate baseline and multiplied by 100. A reading of 100 means baseline performance; below 100 means post-wicket contraction. The sample window runs January 2026 to December 2026, covering Bangladesh home Tests and 41 collapse events. Rain-affected sessions and partnerships shorter than 15 balls were excluded, because small samples make PWRI jump artificially. The results arrived in three layers. First layer — an average PWRI of 68.4. In the ten overs after a wicket, Bangladesh routinely falls to two-thirds of baseline. For comparison, equivalent home events give India 87 and Sri Lanka 81. The first signal is clear: the difference is not talent, it is tempo. Bangladeshi batters are forced to play 12 to 15 extra dot balls in that window, which builds pressure across the next two sessions. Second layer — the non-striker factor. In the 30 balls after a wicket, the new batter's strike rotation rate falls to 38-42 percent. The man arriving at the crease consumes 30 balls without changing ends. The set batter's exposure in front of the ball shrinks, and scoreboard pressure accumulates at both ends. At domestic and international level alike, a large share of set batters are dismissed immediately after this unstable rotation. Third layer — the contagion window. If PWRI drops below 70 in the ten overs after the first wicket, the probability of a second or third wicket in the next 20 overs is 1.6 times baseline. This figure is not just statistical; it matches spin bowlers' planning. An attacking spinner starts with slip and short leg, but once the set batter is gone the field shifts — long cover, deep midwicket, and four or five overs from the same end. The new batter thinks of driving through cover while the ball skids low into an inside edge or pad-bat trap. Watching matches over the years, I notice what the data confirms: a new batter's first ball is usually a block, the second or third a premeditated shot, the fourth a missed half-volley. That three-ball pattern repeats almost identically. At Mount Maunganui on January 1, 2026 — Bangladesh's first Test win in New Zealand — PWRI in the first innings stayed above 90. The difference that day was not talent but decisions: soft hands for the first six balls, and the non-striker taking responsibility to rotate ends quickly. So is the problem psychological? I won't say that, because this is where correlation and causation blur. There are three alternative explanations for a low PWRI that the model does not capture. First, pitch age: wickets tend to fall in the second or third session, when the surface has begun turning. A low PWRI may be a symptom of pitch deterioration rather than batters freezing. Second, the opposition's lead spinner quality factor — facing a frontline spinner raises risk inside the post-wicket window. Third, survivorship bias: we remember the failures to recover, while equally sized successful recoveries sit in the database and never reach the headline. There is another gap I readily admit. PWRI cannot capture dressing-room role clarity, the batting coach's message, or a senior batter's instruction. I don't chase outliers; I follow them until they confess. A batter who cannot rotate strike for 30 balls often has no technical problem — he does not know his role: see off or score? In home conditions this ambiguity hurts most, because no metric is visible other than the value of survival. The limitations deserve honesty too. Forty-one events is a small sample; a single series produces only four or five windows, so the error margin is wide. Second, PWRI is not match-state neutral — a team following on or trailing by 300 faces a different situation. Third, comparing PWRI across different pitch qualities is not comparing like with like. So I treat the index not as a verdict. Expected truth is not a final ruling; it is a working estimate that the next series can test. For the next home series I have three pre-registered checkpoints. One, the 12th-over checkpoint: if the first wicket falls before the 12th over and PWRI in the next ten overs stays below 50, the chance of losing the last five wickets rises above one-third. Two, if the non-striker strike rotation rate drops below 50 percent, a second wicket is more likely within 30 balls. Three, if the match is at Mirpur rather than Sylhet, the threshold gets harder, because low bounce makes end-changing difficult. One thought stays with me. The empty-stadium data of 2026 taught me that when context changes, the meaning of numbers changes with it. Home-condition pressure is nothing new for Bangladesh, but if that pressure can be expressed in a named index, the next collapse will at least not arrive unannounced. The question now is this — in the next series, will we still send two new batters to the crease after the 12th over with the same plan, or will PWRI become a line item in the coaching meeting?