Regime Lab
Single Sharpe ratio hides truth że most edges są conditional: great w calm, trending markets i dead w turbulence (lub reverse). Regime Lab breaks strategy's history do volatility regimes i shows jak to robiło w każdy — więc wiesz gdy Twoja edge rzeczywiście works.
Otwórz cBots → Regime Lab (/quant/regimes).
Co robi
Biorąc return series (lub equity curve, oldest first), to:
- computes trailing realized volatility na każdy point i splits history do Calm, Normal i Turbulent regimes przez terciles tego volatility;
- reports per-regime performance — observations, mean return, volatility i Sharpe — więc możesz see gdzie edge lives;
- estimates Hurst exponent via rescaled-range (R/S) analysis: above ~0.55 series to trending / persistent, below ~0.45 to mean-reverting, i around 0.5 to close do random walk.
POST /api/quant/regimes
{ "returns": [...], "window": 10 } // lub { "equity": [...] }
Dlaczego jest niezawodny
Pure, deterministyczne domain code (Core.Regimes) z żadną infrastrukturą dependency i no external calls
— unit-tested dla regime separation (calm vs turbulent volatility) i dla Hurst direction
(anti-persistent series score poniżej 0.5, persistent trend scores above). Same regime signal feeds
autonomous agents' reflection loop, więc agent może lean do regimes gdzie jego edge to real.