After the leftover hunts, still no live book

Editorial map 2. Not a new prereg. Not live advice. Supersedes What a week of falsification left standing as the scoreboard.

Thesis

Still no live book.

The first map left thin next-bar fades (dead at 5 bps) and vol markers. The leftover doors were run. One directional object cleared overlapping 5 bps and then failed a single-position 8h gate. Do not keep the overlapping t.

What is still standing as facts, not entries

1. Vol clustering — 1h |r| AC1 ~0.25–0.30. Half-life will not rank twice (vol memory).
2. 15m residual magnitude on BTC and ETH, not SOL — extra next-bar |r| after |r| control (BTC, ETH, SOL kill).
3. Crowded-long mean underperformance — global account-ratio HIGH underperforms MID over 8h on BTC/ETH/SOL. Not independently tradeable after the nonoverlap stress.

The L/S saga (read before quoting PROMOTED)

1. Crowded longs underperform; crowded shorts do not bounce — WATCH
2. Crowded longs underperform on ETH and SOL too — WATCH
3. Crowded-long short clears validation — research object, not a live bot — overlapping PROMOTED
4. Crowded-long short does not survive a single-position book — WATCH demotion (n=71, t=1.60)
5. Top-trader longs are the same crowded-long dummy — NO_EDGE (corr 0.99)
6. Position-weighted top-trader longs do not clear validation — NO_EDGE

Do not quote step 3 without step 4.

Closed leftover doors

Funding→perp same null as spot · residualized OI OVERFIT · regime half-life anti-replicates · futures taker no increment over spot · SOL vol15 selection · ETH/BTC z-fade · daily TSMOM / MR / XSMOM.

Honest residue

Vol clustering. 15m volume residual magnitude (BTC/ETH). Crowded-long mean underperformance that collapses under single-position independence. Thin 1h fades that die at 5 bps.

No live signal.