I thought it didn’t work because when you start a new algo you have to fill in risk management “Max position size” and there you can not fill in 0.5 contract.

See Paul’s version 5 posted on page 12 (post #103373). Double-check your time zone

Open the Variable Optimization window (top left of code window). Select walk forward test, change the repetitions to 5 and then select Inactive again to disable walk-forward. You can then rather use the standard optimization if you are looking to optimize for best backtest results. But be warned this will result in curve fitting

@Stefanb, it does that already. Just look at the closed positions list.

It is depended on the reset counter and the values you use for position criteria and how sharp/lose you set your exit criteria.

I am also getting “Your ProBackTest exceeds the limit of repetitions for walk forward optimization”. It works if i put 4 instead of 6. If I put 5 it just changes back to 6 and error again.

Does it effect the trading system in any way having 4 instead of 6? In a negative way?

The results are pretty good and it s my real trading account.

A question: When should we stop with an automated strategy that is winning well and then begins to loose in the future?

At which moment? When the dradwown exceeds the backtested drawndown?? When the win percentage goes below the backtested win percentage?

Thanks in advance

@vince_laur No action indeed. It’s curve-fitted with the trading times. Have a test if you set usetimecriteria to 0 and perhaps adjust the starttime.

thx Paul, but ratio W/L is better with usetimecriteria set to 1  :

3,79 vs 2.32

(based upon 200k backtest)

so, i prefer not to modify usetimecriteria.

thx anyway