Status Card — V0

This backtest study examined the Flux Signal model on historical market data covering the period from January 2022 to March 9, 2026. The test framework was intentionally designed in a simplified structure, consistent with the earlier Flow Observatory methodology, and was organized into four core stages: data synchronization, signal engine construction, historical backtesting, and consolidated reporting. Historical Binance data was used for BTC, while Alpha Vantage daily price data was used for GOLD and SPX.

During the initial test phase, the model produced an imbalanced regime distribution, particularly on the GOLD and SPX side, where signal states were overly concentrated in upper bands. On the BTC side, the CRITICAL regime did not appear at all. To address this, the volatility, stress, and sentiment scales inside the signal engine were recalibrated. Following this revision, the regime distribution became more balanced, allowing the model to produce a more interpretable separation between STABLE, PRESSURE, TRANSITION, and CRITICAL states.

The recalibrated results indicate that the model demonstrates a meaningful regime discrimination capacity, especially for BTC and SPX. In the BTC series, the TRANSITION regime was associated with larger forward price movement and a weaker forward return profile, suggesting that higher λS states may correspond to structurally more fragile market conditions. In the SPX series, higher signal intensity was clearly associated with broader future movement ranges, particularly over longer horizons. This suggests that, in its current form, Flux Signal is not merely behaving as a directional signal, but more fundamentally as a detector of market energy, structural tension, and regime intensity.

The GOLD results improved after recalibration, particularly in terms of signal distribution, but the post-regime behavior remains less consistent than in BTC and SPX. For that reason, the GOLD component should currently be interpreted as promising but not yet sufficiently stable for strong inference.

Overall, this study establishes that a simplified but functional historical testing framework for Flux Signal has been successfully constructed, and that the model already shows encouraging early evidence of regime-reading capability in BTC and SPX. At this stage, the present version may be regarded as the V1 calibrated baseline of the Flux Signal backtest framework.