Cross-sectional crypto momentum does not clear validation

Verdict: NO_EDGE · Universe: 10 liquid USDT pairs · TF: 1d · Sample: chronological 60/20/20

Hypothesis: long top-K / short bottom-K by past R-day return beats chance on a fixed liquid crypto panel. Distinct from the rejected 2-asset alt/BTC spread and from single-asset BTC TSMOM.

What happened

Frozen R=20, H=5, K=2 (K=3 report-only). Panel from data.binance.vision: BTC, ETH, SOL, BNB, XRP, ADA, DOGE, AVAX, LINK, DOT — 2165 complete days (2020-09-22 → 2026-08-26), splits 1299/433/433.

Q1 (val, non-overlap): E[LS]=+0.164%, t=0.23 ≪ 2, hit=46.5% → gate fails. Discovery t=1.77 (soft scent only); OOS t=0.46. K=3 does not rescue (val t=0.54).

Q2: diagnostic only after failed Q1 — val α t=0.13, residual vs equal-weight universe t=0.29. No market-disguise salvage.

Trading

Q3 skipped.

Final Verdict

NO_EDGE

Do not ship a top-2 / bottom-2 crypto momentum book from this 10-name daily panel. Cross-section does not rescue what single-name TSMOM already failed.