BTC weekday is a vol calendar, not a return premium

Verdict: NO_EDGE · Asset/TF: BTCUSDT 1d (primary), 1h (secondary) · Sample: chronological 60/20/20

Hypothesis: day-of-week (UTC) carries a mean-return premium distinct from the already-rejected hour/session study. Two claims frozen before outcomes: A weekend (Sat+Sun) mean return differs from Mon–Fri; B Monday mean return differs from other days.

What happened

Q1 Claim A (1d): sign agrees (weekend lower mean r) but validation t=−0.75, p=0.45 → FAIL.

Q1 Claim B (1d): sign agrees (Monday higher) but val t=+1.11, p=0.27 → FAIL; effect is a 2024 monopoly in discovery.

Shuffle null: real |t| sits below the null mean (frac ≥ real ≈ 0.57–0.69).

Q2: Strong, replicable |r| pattern — weekend quieter, Monday louder. That is a vol / risk calendar, not a return premium. Same qualitative end as the UTC hour/session study.

1h and 1h→day aggregations fail Q1 the same way.

Trading

Q3 not run (no directional unlock).

Final Verdict

NO_EDGE

Clock and calendar mark volatility texture on BTC. They do not mark a tradeable mean-return edge on this sample. Do not ship Monday-long / weekend-short bots from kline DOW alone.