run-study
command run-study
Run a parameter study defined in a configuration file.
Reads a .toml, .yaml/.yml, or .json file containing a config
mapping with the shared experiment settings and a study mapping with the
parameter sweep. Every candidate is run and persisted under one study
record. Optional walk-forward validation selects candidates on training
windows and evaluates them on the test windows that follow.
Read more in the study guide.
| Parameters |
Path to a study configuration. The top-level
--log_level, -l : str, default="warn"config field accepts an
ExperimentConfig. The top-level study field accepts
parameter_space, optional saved strategy, min_trades,
max_drawdown, and optional walk_forward settings.
Minimum log level to emit. Choose from:
--verbose/--no-verbose, -v : bool, default=Trueerror, warn, info or
debug.
Whether to display progress while the study is running.
|
See Also
Run a study with a parameter sweep and optional walk-forward validation.
Run a backtest experiment defined in a configuration file.
Launch the Backtide UI in a local web browser.
Example
Create study.toml using the name of one saved strategy:
[config]
metrics = ["sharpe"]
[config.general]
name = "SMA parameter study"
[config.data]
symbols = ["SPY"]
full_history = false
start_date = "2012-01-01"
end_date = "2025-12-31"
[config.strategy]
strategies = ["My SMA strategy"]
[study]
min_trades = 30
max_drawdown = 0.25
[study.parameter_space]
fast = [10, 20, 30]
slow = [100, 150, 200]
[study.walk_forward]
training_days = 1095
test_days = 365
Run the study:
backtide run-study study.toml