BTC Regime Classifier / ETF Structural Edge — DORMANT

Five-signal BTC regime classifier (FBDI/SSR/VCS/EAR/SSI → DAMPENING/STRESS/UNWIND) failed to detect BTC’s 30% drawdown — documented as fundamentally broken in KB277. Do not import or act on until redesigned.

Thesis Health

SignalReadingDetail
Evidence Balance⊖ Net negativeModel failure documented; no rehabilitation evidence
Evidence VelocityZeroNo new work since KB277 failure analysis
Consensus Breadth0 active contributorsNo member engagement on this signal
Contestation LevelN/ANot contested — acknowledged as broken
Market ValidationFailedMissed 30% BTC drawdown — the one event it was designed to detect

Status: Dormant — This signal framework was built as part of Etta’s infrastructure. It produced five composite signals (Flow-Based Demand Indicator, Supply Shock Ratio, Volume Confirmation Score, Exchange Activity Ratio, Stablecoin Strength Index) that mapped to three regime states. It missed BTC’s 30% drawdown entirely. KB277 documents the failure analysis. The framework remains in Etta’s codebase but is flagged as DO NOT IMPORT into Inner Loop.

What would reactivate this theme: A complete redesign of the regime classifier with backtesting against the missed drawdown, proposed by a member with quantitative expertise. Not a priority — deferred to post-MVP (B33 in build order).

Thesis / Overview

The original thesis: BTC market microstructure signals (exchange flows, stablecoin movements, volume patterns, supply dynamics) could be combined into a regime classifier that identifies DAMPENING (accumulation), STRESS (distribution), and UNWIND (capitulation) phases with enough lead time to be actionable.

The failure: the classifier remained in DAMPENING throughout a 30% drawdown. The signals were either too slow, too noisy, or structurally unable to distinguish between healthy consolidation and genuine distribution at scale.

This is kept as a dormant theme rather than deleted because: (a) the underlying data sources are still valuable, (b) a redesigned classifier could work, and (c) the failure analysis itself is useful intellectual capital for future quantitative signal design.

Key claims

  • ⊖ Missed 30% BTC drawdown — the defining failure (→ kb277-etf-edge-failure)
  • ⊖ Five signals showed no degradation signal before or during the drawdown
  • ⊖ Exchange flow data may be fundamentally unreliable (OTC desks, wrapped tokens, cross-chain bridges all create noise)
  • ⚖ Individual component signals (esp. stablecoin flows) may still have value as inputs to a simpler model — not tested
  • gold-case-for-and-against — BTC/Gold correlation (Hard Money Ratio) is a separate, functioning signal
  • ma-hard-assets — BTC positions (WBIT) exist in portfolio but not based on this regime classifier

Counter-arguments & data gaps

  • Was the failure in the signals, the combination logic, or the threshold calibration? KB277 doesn’t fully resolve this.
  • Exchange flow data quality has improved since the original build — a rebuild might work with better inputs.
  • The 30% drawdown may have been driven by factors outside the model’s scope (macro contagion, not BTC-specific microstructure).

What would change this view

  • Complete redesign with backtesting against the missed drawdown period
  • Alternative regime classification approach (e.g., on-chain analytics, options-implied, or cross-asset momentum) that demonstrably would have caught the drawdown
  • Member with quantitative expertise proposing a rehabilitation plan

Sources

Events reckoned with

  • BTC 30% drawdown — missed by classifier — reckoned 2026-01-15 (KB277 written)
  • Decision to defer to post-MVP (B33) — reckoned 2026-04-05