`ROCP`#

Description#

ROCP(k) is the rate of change over k steps, expressed as a fraction:

\[ \text{ROCP}[t] = \frac{x[t] - x[t-k]}{x[t-k]} \]

This is exactly Return(k). It exists as a separately-named class because TA-Lib calls this indicator ROCP. Internally ROCP is a thin subclass of Return -- the implementation (delay buffer + subtract + divide) is shared, not duplicated.

You want...	Use
TA-Lib parity (writing `ROCP` in ported code)	`ROCP(k)`
Any other context (returns workflows, log-returns family)	`Return(k)`

Parameters:

window_size (int, positive): the lookback k.

NaN handling: NaN for the first k samples; NaN when x[t-k] == 0.

NaN handling#

Policy: propagate. Input NaN values are stored in the lookback. Output is NaN at any index where the function's positional formula references a NaN input; recovery happens once the NaN slides out of the lookback.

Examples#

Usage example#

import numpy as np
import plotly.graph_objects as go
from plotly.subplots import make_subplots
from screamer import ROCP

np.random.seed(0)
price = 100 * np.exp(np.cumsum(np.random.normal(0.0005, 0.02, size=300)))
rocp = ROCP(window_size=20)(price)      # fractional change over the last 20 bars

fig = make_subplots(rows=2, cols=1, shared_xaxes=True,
                    row_heights=[0.55, 0.45], vertical_spacing=0.08)
fig.add_trace(go.Scatter(y=price, mode="lines", name="price"), row=1, col=1)
fig.add_trace(go.Scatter(y=rocp, mode="lines", name="ROCP",
                         line=dict(color="red")), row=2, col=1)
fig.update_layout(title="Fractional rate of change over 20 bars (ROCP)",
                  margin=dict(l=20, r=20, t=60, b=20),
                  legend=dict(orientation="h", yanchor="bottom", y=1.02, xanchor="right", x=1))
fig.update_yaxes(title_text="price", row=1, col=1)
fig.update_yaxes(title_text="fraction", row=2, col=1)
fig.show()

Reference#

Equivalent to talib.ROCP(x, timeperiod=k). Bit-exact match (cross-validated in tests/test_third_party_alignment.py).

ROCP#