← Back to Promethium Labs
Methodology · Updated

Research · How the system works (and how we know)

We learn in public. Methodology first, performance second.

Most crypto-signal services post precision numbers without showing the math. We do the opposite — strategy components, backtest methodology, the regime framework, and the iteration log. Live performance numbers publish automatically once a meaningful sample of closed trades accumulates. Until then, expect transparency about what we’re measuring and why.

Current strategy

Loading…

What backtests can’t tell us

Backtest ≠ live performance

Forward-pick replay assumes perfect order fills, no slippage above stated fees, and that historical liquidity reflects what we’d actually face. Live trading exposes real microstructure cost.

Selection bias is real

Strategy components were tuned on the same data we backtested. Some of the measured edge is curve-fitting. Out-of-sample live performance is the genuine test.

Regime conditioning isn’t proven yet

The regime-conditional bear-trap detector hasn’t had a chance to fire in our 30-day window because we’ve been in a sustained bull market. Its calibration is theoretical until a regime shift validates it live.

Back to main sitePromethium Labs · transparent by default