[!todo] Seed note. A starting point, not a finished note yet.
Right-sizing is the practice of breaking work down so that individual items fall within a roughly similar, small size range, rather than assigning each a story-point estimate. It matters because it is the assumption that makes count-based forecasting honest: when items are similar in size, counting how many finish per week is a meaningful signal, and a Monte Carlo simulation over that count produces a trustworthy distribution. When items vary wildly, a single six-month epic hiding among fifty one-day tickets corrupts the throughput history and the forecast built on it. The discipline replaces “how big is this?” with “is this small enough to start?”, which is a faster and more reliable question. Seeded from The Probabilistic Delivery Forecaster.