I'm using v6.28 and I've tried to reproduce one result of the walk-forward optimization with the optimization target to the "L. Avg. Bars Held" manually, my system is a simple two moving average cross over, so I've two parameters named period_ma1 and period_ma2, and I've done the following steps:
Step0. Run a walk-forward optimization with the following date range:
from 2022/2/7 to 2022/2/18, as the In-sample date range,
from 2022/2/21 to 2022/2/25, as the Out-of-sample date range,
then save the reports for comparison.
Step1. Run an Optimize within the date range 2022/2/7 to 2022/2/18, as the In-sample results. then sort the results by "L. Avg. Bars Held" ascendingly to get the corresponding parameters (period_ma1 and period_ma2) for Out-of-Sample backtesting.
Step2. Run a Backtest within the date range 2022/2/21 to 2022/2/25, as the Out-of-sample results. Firstly apply parameters period_ma1 and period_ma2 from Step1, then run a Backtest.
then I've found the report of this Backtest is totally different (such as Net Profit, Profit Factor, etc.) from what I've seen on the walk-forward backtest report with the same date range.
Do I lose any step or make any mistakes?
Thank you very much!