Grid Search (Parameter Sweep)

OptionForge grid search helps you optimize options backtests by sweeping strategy parameters and comparing outcomes across all tested combinations.

Grid search runs the same strategy across a range of parameter values and returns a report so you can compare performance across combinations.

Subscription required: Grid search is only available on the special Grid Search subscription tier. If you do not see the Run Grid Search button, your account is not enabled for this feature yet.

Syntax

Add grid parameters using a standard Lua assignment followed by a comment in the format [min:max:step] This means the grid search code is also value for a single back-test. It's strongly recommended to make sure that the single back-test is running without issue before using a grid-search.

dte = 45 -- [30:90:15]

Examples

width = 10 -- [5:20:5]
short_delta = -25 -- [-40:-10:5]
qty = 1 -- [1:4:1]

local profit_target = 150 -- [50:300:50]
stop_loss = -400 -- [-800:-200:100]

dte = 21 -- [7:63:7]
entry_hour = 10 -- [9:15:1]
entry_minute = 30 -- [0:45:15]

This lets you sweep non-integers and mixed types by mapping a simple integer index.

local idx = 1 -- [1:5:1]
local val = ({23.5, 99.4, 123.6, 298.1, 333})[idx]
local str = ({"Mon", "Tue", "Wed", "Thu", "Fri"})[idx]

Notes

It is highly recommended to debug with a normal backtest first, then run grid search once the script behaves as expected.

Only integer values are supported for grid parameters.
Grid search currently supports at most three varying parameters.
The range is inclusive: min and max are both tested.
step must be positive.
Each grid parameter line can be local, but the variable name must be a plain identifier.
Every combination of parameters is run, so total runs grow quickly as you add parameters.

Running Grid Search

After adding grid parameters, click Run Grid Search in the UI. A grid report will open when the runs finish.