Randomize starting bar

3 replies


Subscriber, bbp_participant, community, 4 replies.

Visit profile

6 years ago #113988

I have some doubts about Randomize starting bar robustness testing.
This test is applied for each trade separately? For example, if the strategy is applied in hourly period and a buy signal is generated at 15/07/2015 9:00, the test will generate random entry delays at 15/07/2015 9:00 (without delay),10:00 (delay of 1 hour), 11:00 (delay of 2 hours), 12:00 (delay of 3 hours)…?
And this test is applied for delayed exits too?
What means the parameter (max change), is the max possible delay for the period?


Matusiak Adrian

Customer, bbp_participant, community, 308 replies.

Visit profile

6 years ago #131600

Good question




Customer, bbp_participant, community, 542 replies.

Visit profile

6 years ago #131605

This one is a really easy one, it just starts the backtest X bars (the “Max Change”) later from your current backtest starting date. This is to avoid strategies that relay on previous trades and only work if they are started at a specific date. E.g. if the strategy is started 1. January 2000 it makes profit, but if started 1. February 2000 it makes constant losses, which can happen with strategies that use “Reverse Existing Position” market order entries. Hence this robustness test changes the starting date of the backtest by X bars randomly to make sure the strategy makes profits regardless of when the backtest starts.


Mark Fric

Administrator, sq-ultimate, 3 replies.

Visit profile

6 years ago #131636

Lorenz thank you for the explanation, it is exactly as you said 🙂

StrategyQuant architect


Viewing 4 posts - 1 through 4 (of 4 total)