zeonta.ift_rsi() — RSI compressed toward -1/+1 through Ehlers’ Inverse Fisher Transform.
What it measures
Rescales rsi toward zero, smooths it, and squashes the result through Ehlers’ Inverse Fisher Transform — a curve that passes the middle of its input through almost unchanged but compresses everything else hard toward -1 or +1, trading RSI’s gentle 0-100 curve for a near-binary reading.
Formula
v1 = 0.1*(RSI-50); v2 = WMA(v1, smooth); IFTRSI = (exp(2*v2)-1)/(exp(2*v2)+1)
Parameters
Required inputs: close
| Parameter | Default |
|---|---|
length |
14 |
smooth |
9 |
Returns
| Column |
|---|
IFTRSI_14_9 |
Usage
Examples run against the 300-bar OHLCV fixture in tests/data/ohlcv.csv, loaded as df. The output shown is the real output.
import pandas as pd
import zeonta
df = pd.read_csv('tests/data/ohlcv.csv', parse_dates=['date']).set_index('date')
zeonta.ift_rsi(df['close']).tail(3)
date
2024-10-25 -0.616901
2024-10-26 -0.687237
2024-10-27 -0.764766
Name: IFTRSI_14_9, dtype: float64
Accessor form: df.zta.ift_rsi(...)
How to read it
Readings pin close to -1 or +1 far more often than RSI pins near 0 or 100 — that compression is the entire point, giving very clear (if less nuanced) turning-point signals.
Pitfalls
The compression means small, genuine changes in the underlying RSI can vanish once squashed toward an extreme — this trades resolution for clarity, not a free improvement on RSI.
Reference
Formula source: https://www.mesasoftware.com/papers/TheInverseFisherTransform.pdf