A Parameter Sweep tests multiple combinations of selected strategy settings on the same historical dataset.
Instead of choosing one value for each setting, you define ranges for two parameters. The resulting combinations are tested so you can compare how the strategy behaved across those settings.
Start with two parameters
The current Parameter Sweep is two-dimensional.
This means two parameters change during the Sweep while the other settings remain fixed.
For example, a Bollinger strategy Sweep may compare different combinations of:
- Bollinger Length
- Deviation Multiplier
If several values are included for each parameter, every valid pairing becomes a configuration to test.
How the ranges create combinations
Each parameter range is defined using:
- From
- To
- Step Size
For example, suppose the first parameter contains 4 values and the second contains 5 values.
The Sweep creates:
Each combination is then tested on the same historical dataset.
This lets you compare the settings under consistent data conditions.
Fixed settings still matter
Only the selected Sweep parameters change.
Other settings remain fixed for that research run.
Depending on the strategy, these may include:
- other strategy parameters
- stop loss or target settings
- candle handling
- trading-cost settings
- other configuration values
When reviewing the Result, remember that the comparison reflects both the changing parameters and the fixed settings used during the Sweep.
Why run a Parameter Sweep?
A Parameter Sweep is useful when you want to see how a strategy responds to a range of settings instead of relying on one manually selected configuration.
It can help you:
- compare many configurations efficiently
- see how performance changes as parameter values change
- identify configurations worth reviewing more closely
- observe whether nearby settings behave similarly or very differently
This gives you a broader view than a single Backtest.
One dataset, many configurations
Every tested combination uses the same dataset for the Sweep.
Keeping the historical data consistent makes it easier to compare the configurations with one another.
If you change the dataset or historical period, that becomes a different research comparison.
What the Result shows
A completed Parameter Sweep brings the tested configurations together in one Result.
You can use:
- ranked configurations
- the heatmap
- Top Configurations
- performance metrics
to explore how the tested combinations behaved.
The aim is not only to find the highest number. The wider comparison can show how performance changes across the parameter ranges.
Where to go next
Continue with: