Confluence State
A graded confluence state over one representative from each of five independent domains. Read forward, not momentum: bullish = cheap/fearful (favourable forward conditions), bearish = expensive/euphoric — a euphoric top reads bearish by design. Each component fires against its own expanding trailing window (prefix-invariant, look-ahead-free). The effective domain count (participation ratio of the rolling correlation matrix, on the changes) exposes redundancy: "5 of 5" can be effectively 3.x. Extends the primitive bull-market gauge; descriptive, not advice.
- Formula:
- Tercile vote per domain (trailing percentile) → aligned-count + dispersion (std-dev of conviction) + N_eff = (Σλ)²/Σλ² of the rolling correlation matrix
- Originator:
- Backtesting Arena (weather-ensemble analogy, read correctly)
- Data source:
- Compute-on-read über je einen PIT-sauberen Vertreter je Domäne (Mayer/MVRV-Z aus btc_chart_metrics, Macro-Regime PIT via Vintages, Funding, Fear & Greed) — /api/charts/confluence-state
Strategies to backtest
Thematically related strategies from our library — try them in the backtest engine or read up on the methodology.
Trade the market mood, not the noise — average sentiment over a defined cadence window. A unique sentiment-driven strategy for crypto.
Open strategy →The textbook RSI strategy — buy when oversold (RSI < 30), sell when overbought (RSI > 70). Best in sideways markets.
Open strategy →More Bitcoin indicators
Aggregated sentiment score 0-100 (alternative.me). Extreme fear historically marks bottoms, extreme greed marks top zones — e.g. March 2020 (8), November 2022 (20), April 2021 (95).
Current percentage distance to all-time high plus historical comparison of every bear market (depth, duration, recovery time).
Deribit DVOL (implied vol) vs. realized 30-day volatility. The area between is the variance risk premium.