Signal Inversion

Signal Inversion takes a chosen trigger stream and trades against it — the definitive clean build of signal inversion as a measured decision rather than a reflex. Pointed at the built-in MA-cross or RSI-threshold demos, or at any other indicator's plot, it grades every original signal on a fixed horizon, charges both directions identical costs, and prints inverted markers only where the numbers back the fade.

How to Trade the Signal Inversion?

• Inverted Long / Short Signal: triangles fire against the original stream while the inversion gate is satisfied — a faded entry needs its own stops and faster targets, since the original's exits do not carry over.
• Negative-Edge Background: tints the stretches where the window shows the original losing beyond its costs; its rotation across history is regime dependence made visible.
• Comparison Dashboard: hit rates, net per signal and cumulative net for both streams under identical costs, plus a verdict row stating plainly when neither side is worth trading.

Signal Inversion Settings

• Signal To Invert (default MA Cross): the stream being graded and mirrored; External Signal points at any other plot.
• Grade Horizon (Bars) (default 10): bars ahead each signal is scored on — keep it near the fade's holding period.
• Rolling Window (Signals) (default 20): graded signals in the sample behind the statistics and the gate.
• Cost per Trade (%) (default 0.1): round-trip friction charged to both streams.
• Inversion Mode (default Adaptive (Edge Gate)) with Minimum Sample (Signals) (default 10): gates the mirror on measured negative edge; Always Invert is the unconditional research view.
• Demo inputs — MA Type (default EMA), Fast Length (default 20), Slow (default 50), RSI Length (default 14), RSI Upper (default 70), Lower (default 30) — shape the built-in triggers.

Frequently Asked Questions

Why doesn't it invert every losing signal?

Costs subtract from both directions, so the two net expectancies always sum to minus twice the cost per trade — a stream that merely bleeds its costs mirrors into another bleeder. The gate therefore waits for losses beyond costs across the minimum sample.

How does this relate to fading breakouts?

Anti-breakout trading is the classic special case: the False Breakout detector hunts one failure pattern at range edges, while this engine measures any stream's failure rate and only then trades the other side of it.

What alerts does the build include?

Seven: inverted long, short and any-direction signals, the original pair for A/B comparison, and Fade Edge Detected / Fade Edge Lost when the gate changes state.