Skip to main content

Studio · Product guide

Compare AI optimization candidates

Compare strategy parameters using return, drawdown and validation.

Before you begin

Use a tested version with searchable parameters and available optimization allowance.

Step-by-step instructions

Screenshots are examples. Your balances, limits, availability and status may differ.

  1. Open Backtest results and select Optimize. Choose an objective and maximum drawdown.

    Step 1 · Compare AI optimization candidatesActual app screen · English interface · Illustrative demo data
  2. Compare training and held-out validation results. Candidates above the drawdown ceiling cannot be selected.

    Step 2 · Compare AI optimization candidatesActual app screen · English interface · Illustrative demo data
  3. Use a candidate as a new draft, then run a fresh backtest before saving it to the bot library.

  4. For a Python script, declare the values you want searched as typed inputs (integer or number with a minimum and maximum). Only inputs with bounds can be optimized.

What you should see

A selected candidate becomes a new draft version; the original version and running bots remain unchanged.

If something goes wrong

If no parameters are available, review the strategy inputs. If no candidate qualifies, review the limits and evidence rather than assuming a winner exists.

Contact support
Each optimization run uses one of your optimization runs. A candidate that fits the training period can still fail on new data, so judge it on the validation period.
Open StudioProduct manual · Follow the current status and rules shown in your account