Skip to content

Volatility

Volatility targets compute realized volatility over a future window. They have warm_up_period = 0 and forward_period = horizon.

Future CTC Volatility

Research only <300µs/1k bars

\[ \sigma_t = \text{std}\!\left(\ln\frac{x_{t+1}}{x_t},\, \ln\frac{x_{t+2}}{x_{t+1}},\, \ldots,\, \ln\frac{x_{t+h}}{x_{t+h-1}}\right) \]

Close-to-close realized volatility over the next horizon bars. Computes the sample standard deviation of log returns over the forward window. The output name is auto-generated as {input}_future_ctc_vol_{horizon}.

Name Type Constraint Description
input str non-empty Price column, e.g. "close"
horizon int >= 1 Number of bars to look ahead (\(h\))
Column When valid Description
{input}_future_ctc_vol_{horizon} t <= N - horizon - 1 Sample std of horizon log returns starting at t+1
Property Value
warm_up_period 0
forward_period horizon

  • Forward NaN. The last horizon values are NaN because the future window is not complete. The first valid value appears at bar 0.

  • NaN in prices. If any price in the forward window is NaN or non-positive, the corresponding log return is NaN, and the standard deviation returns NaN.

  • Stateless. run_research() has no internal state. Calling it twice with the same input always returns the same output.

  • Implementation. Computes log returns via a pairwise pass, then applies a rolling sample standard deviation with shift to align results to bar t.

  • Label. The standard label for realized volatility forecasting. Captures how much the price moved over the next horizon bars, regardless of direction.

  • Units. Log-return scale, not annualized. To annualize, multiply by sqrt(periods_per_year).

import pandas as pd
from oryon.targets import FutureCTCVolatility
from oryon import TargetPipeline, run_targets_pipeline

t = FutureCTCVolatility(input="close", horizon=3)
print(t.output_names())  # ['close_future_ctc_vol_3']

tp = TargetPipeline(targets=[t], input_columns=["close"])
df = pd.DataFrame({
    "close": [100.0, 101.0, 103.0, 102.0, 105.0, 107.0, 106.0],
})
out = run_targets_pipeline(tp, df)
print(out)
#    close_future_ctc_vol_3
# 0                  0.0150
# 1                  0.0202
# 2                  0.0201
# 3                  0.0199
# 4                     NaN
# 5                     NaN
# 6                     NaN

crates/oryon/src/targets/future_ctc_volatility.rs