Say if i was backtesting using:



But then adjusted this to:



Why do my backtesting results differ?

I would have though the second version amounts to a BarPeriod of 15 also and so the backtesting results should be exactly the same as each other?

Or am I not understanding this correctly?

The bar period determines the duration of a bar, the bar offset determines the time when the bar starts.

Thanks JCL, i knew it was simple

So the second example would simply be a 12 minute bar that starts 3 minutes ahead of the standard 15min bar period... Basically it's a customised bar, and you can do this with any figures to suit your strategy?

This is cool, and for some strange reason the backtesting results are much better with the second version.

Would you encourage this non standard way of picking Bar Periods and OffSets if the backtesting results significantly increase?

Thanks again for your help and for Zorro