Summary
SmoothedMovingAverage is RTTA's streaming implementation of: Wilder/SMMA/RMA smoothed moving average seeded by an initial SMA window.
Update API
result = rtta.SmoothedMovingAverage().update(value)
The update(...) call consumes one observation using value. advance(...)
uses the same inputs when the caller wants to update state without materializing
a Python return value.
Theory Of Operation
SmoothedMovingAverage is the classic Wilder smoother used by RSI and ATR,
also known as SMMA or TradingView's RMA. The first full window is a simple
average; subsequent values use Wilder's recursive form with effective alpha
\(1/n\).
Recurrence
Let \(z_t = value_t\) and \(n\) the window length.
For the first \(n\) samples, seed with a simple average:
Thereafter:
which is equivalent to \(S_t = \alpha z_t + (1-\alpha)S_{t-1}\) with \(\alpha = 1/n\).
When fillna=True, partial simple averages are returned before the first full
window; when fillna=False, those samples are NaN.
The return value is the current scalar indicator value.
Implementation Notes
The recurrence is implemented in src/rtta/indicator.cpp in class SmoothedMovingAverage.