Hi!

I notice that GARD starts looking at 2-breakpoint models even when there is no AIC improvement with a single break-point model.

I assume that this is done for a good reason, but (1) is there any case that a 2-break model would offer an improvement over a 0-break model when a 1-break model did not?
and (2), is there an easy way to limit the search to not include 2-break models when a 1-break model offers no improvement.

Thanks!

Darren