Adapted from QuantConnect / Quantpedia 'Short Term Reversal with Futures'. The source runs a weekly cross-sectional book over 24 CME/ICE futures. It keeps only contracts whose weekly volume change is in the top half and whose open-interest change is in the bottom half, then goes long the worst weekly loser and short the best weekly winner. The part worth keeping is the conditioning. A price move that comes with MORE volume but LESS open interest is driven by positions being closed (forced deleveraging and stop-outs), not by new directional capital, so it should partly revert. The 24-future cross-section and the dollar-neutral two-leg book are dropped: our venues have no such universe, and lessons L137/L173 show multi-leg neutral books and ranked baskets die here. The mechanism is instead measured and traded on ONE Binance USD-M perp, where the open-interest feed (supplementary key `open_interest`, Binance metrics, 5-minute, history from 2021-12-01) lives on the same contract that is traded (L160, L143). Pre-study, point-in-time, daily close-stamped bars, OI = last 5-minute value of each day, non-overlapping trades. (1) The UNGATED weekly reversal is negative on all 14 names tested (for example SOL -1.4%, XRP -1.3%, TRX, AVAX), so plain short-term reversal is dead in crypto; momentum dominates. (2) With the vol-up/OI-down gate at |z| > 1, SHORT signals (rallies with falling OI) lose on most names, so the short leg is dropped. (3) LONG signals (sell-offs with rising volume and falling OI) are positive on 9 of 10 alts at a 3-day hold. DOGE: 54 longs, +1.84% mean over 3 days, t≈2.3. Pooled 7-day across 10 names: +0.62%, t=1.2. DOGE is the strongest cell of a small grid (10 names x 2 holds x 3 thresholds), so this is NOT an independent test. The pre-registered numbers below are therefore deliberately modest, and the analyst should require the XRPUSDT.BINANCE cross-check with the rule unchanged before any promotion. The strategy is flat about 90% of the time.