← All indicators

zeonta.divergence() — Regular and hidden divergences between price swings and an oscillator.

What it measures

When price makes a new extreme but the oscillator does not, the move is being made with less force than the one before it. That disagreement — divergence — is one of the few genuinely forward-looking things in technical analysis.

Formula

Regular Bearish = price Higher High + oscillator Lower High; Regular Bullish = price Lower Low + oscillator Higher Low; Hidden Bearish = price Lower High + oscillator Higher High; Hidden Bullish = price Higher Low + oscillator Lower Low

Parameters

Required inputs: high, low, close

Parameter	Default
oscillator	None
osc_length	14
left	5
right	5

Returns

Column
DIVREGBULL_5_5
DIVREGBEAR_5_5
DIVHIDBULL_5_5
DIVHIDBEAR_5_5

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.divergence(df['high'], df['low'], df['close'], left=5, right=5).sum()

DIVREGBULL_5_5    2.0
DIVREGBEAR_5_5    3.0
DIVHIDBULL_5_5    0.0
DIVHIDBEAR_5_5    4.0
dtype: float64

Accessor form: df.zta.divergence(...)

How to read it

Regular divergence argues the trend is tiring and a reversal is closer. Hidden divergence argues the opposite: a pullback inside a trend is ending and the trend is about to resume. The default oscillator is RSI(14); pass any series via oscillator.

Pitfalls

A divergence is a warning, not a signal — in a strong trend an oscillator can diverge three or four times while price keeps going, and each one looks convincing in hindsight. Wait for price confirmation. Note too that flags land on the pivot bar, which is only knowable right bars later: shift the output before backtesting.