Awesome Oscillator in Python — Coding It from Scratch (pandas)

The Awesome Oscillator is a two-line indicator in disguise — a difference of two simple moving averages of the bar midpoint, as the explainer covered. Here it is in pandas, including the momentum bar colours and a saucer signal.

What you’ll need

• Python 3.9+, pandas
• high, low series.

From scratch

import pandas as pd


def awesome_oscillator(high, low, fast: int = 5, slow: int = 34):
    median = (high + low) / 2
    return (median.rolling(fast).mean()
            - median.rolling(slow).mean()).rename("ao")

That’s the whole indicator: median price (H+L)/2, then the 5-period SMA minus the 34-period SMA. No EMA, no signal line, no close — the three things that make it an AO rather than a MACD.

Bar colours and signals

The classic display colours each bar by whether AO rose or fell:

ao = awesome_oscillator(df["high"], df["low"])
rising = ao > ao.shift(1)          # green when momentum increased, else red

# Zero-line cross.
df["ao_bull"] = (ao > 0) & (ao.shift(1) <= 0)

# Bull saucer: AO above zero, two falling bars then a rising bar.
above = ao > 0
two_down = (ao.shift(1) < ao.shift(2)) & (ao.shift(2) < ao.shift(3))
df["ao_saucer"] = above & two_down & (ao > ao.shift(1))

ao_bull and ao_saucer are concrete signals to test in AlgoGen — the saucer especially benefits from a backtest, since it fires often.

Gotchas

• Median price, not close. (high + low) / 2. Using close turns it into something closer to a plain MACD variant.
• SMA, not EMA. The AO uses simple moving averages; swapping in EMAs changes the line.
• Fixed 5/34. These are the indicator’s identity; changing them makes it a generic SMA-difference oscillator, not “the AO.”

Bonus: the Accelerator Oscillator

Williams’ companion indicator is one more line — the AO minus a 5-SMA of itself:

def accelerator(high, low):
    ao = awesome_oscillator(high, low)
    return (ao - ao.rolling(5).mean()).rename("ac")

The AC measures the acceleration of momentum (the change in the AO), coloured the same green/red way. Rising AC above zero is Williams’ strongest “go” condition.

The lazy one-liner

import pandas_ta as ta
df["ao"] = ta.ao(df["high"], df["low"])   # 5/34 by default

Same indicator elsewhere: MQL5, Pine Script, EasyLanguage, NinjaScript.

---

This post is educational, not financial advice. Indicators describe the past; they don’t predict the future. Backtest anything before you risk real money on it.