Inverse Fisher Transform

Inverse Fisher Transform compresses a smoothed, scaled oscillator through Ehlers' S-curve so it pins near +1 or −1 and crosses the middle quickly. This build applies the inverse Fisher transform to a selectable base — a 5-period RSI by default, with Stochastic, SMI, and CCI alternatives — scales it by 0.1, smooths it with a 9-period weighted moving average, and plots the result against ±0.5 thresholds and ±1 rails.

How to Trade the Inverse Fisher Transform?

• Cross up through the Lower Threshold: the conventional buy trigger — the output leaving the lower saturated zone.
• Cross down through the Upper Threshold: the conventional sell or exit trigger.
• Crossings into the outer zones: a cross through +0.5 or −0.5 outward commits the output to a rail — a regime label, not a fade.
• Time at a rail: extended saturation reads as trend persistence; between thresholds the tool is deliberately brief.

All four events carry alerts — trigger markers are off by default, the alerts fire regardless.

Inverse Fisher Transform Settings

• Base Oscillator (default RSI): the input fed to the transform; every base is centered before scaling.
• Length (default 5): base-oscillator lookback, short enough to actually visit the extremes.
• Source (default close): the series feeding the base.
• SMI Smoothing (default 3): double-EMA smoothing for the SMI base only.
• Scale Factor (default 0.1): how often the rails are reached — larger saturates sooner, smaller lingers in the middle.
• Smoothing Length (default 9): weighted average applied before the transform; 1 disables it.
• Upper Threshold (default 0.5) and Lower Threshold (default -0.5): the trigger boundaries.
• Gradient Fill (default enabled), Trigger Markers (default disabled), and the colors are style options.

Frequently Asked Questions

How does it relate to the Fisher Transform?

They are opposites: the Fisher Transform expands the tails of a bounded input so turning points stand out, while the inverse compresses its input into −1 to +1 to create saturated states. One sharpens turning points; the other holds a regime read.

Why does the output sit at +1 for so long?

Saturation is the point: in a persistent trend the output stays committed to one rail rather than oscillating — a +1 print labels a bullish regime, not an imminent reversal. The information lives in the threshold crossings, not the level.

What does the Scale Factor change?

It multiplies the centered oscillator before smoothing and transforming, setting how much input strength reaches the rails. Different scale factors can disagree on the same data — 0.1 is Ehlers' published value.