AcceleratorOscillator

Summary

AcceleratorOscillator is RTTA's streaming Bill Williams Accelerator Oscillator (AC): the Awesome Oscillator minus a short SMA of itself. It measures the change in driving force of the market (acceleration), not just the force.

Update API

result = rtta.AcceleratorOscillator(ao_slow=34, ao_fast=5, smooth=5, fillna=True).update(high, low)

update(high, low) returns a scalar AC value.

Theory Of Operation

Bill Williams' Awesome Oscillator (AO) is the difference between a fast and slow SMA of median price \((h+l)/2\). The Accelerator Oscillator is AO minus a further SMA of AO:

When AC is above zero and rising, upside force is accelerating; when below zero and falling, downside force is accelerating. Zero-line and color-change rules from Williams' framework treat AC as a confirmation filter for AO and Alligator setups.

RTTA composes the existing AwesomeOscillator with an SMA of length smooth.

Recurrence

Median price and AO (matching AwesomeOscillator; high/low are ordered so \(h\ge l\)):

with defaults \(n_f=5\), \(n_s=34\).

Let \(k=\texttt{smooth}\) (default 5). Feed AO into an SMA (NaN AO treated as 0.0 when updating the SMA so state advances):

Warm-up when fillna=False: require \(\max(n_s,n_f) + k\) samples before a non-NaN return. If AO itself is still NaN, return 0.0 when fillna=True, else NaN.

Implementation Notes

• Implemented in src/rtta/indicator.cpp in class AcceleratorOscillator.
• Internals: AwesomeOscillator ao_ with fillna=True, plus SMA ao_sma_.
• Constructor order is (ao_slow, ao_fast, smooth) matching AO's (window_1=slow, window_2=fast).

Reference

• Investopedia — Bill Williams Alligator / AC context
• Awesome Oscillator background