When we try to avoid inbreeding in other species I'm involved with (reptile species), we get stock from different sources, raise up those animals until they're sexually mature, and then pair them off with unrelated animals. In this way, the pairs are known to be unrelated (well, at least non-sibling).
I'm not sure how the random pairing is supposed to be working. There are two sibs, another two sibs, and then a lone ranger. The lone ranger has 0% chance of inbreeding, but the other frogs each have a 25% chance of inbreeding, given a random mate choice from the other four frogs. I suppose if a person IDs the animals and keeps track of pairing, this isn't an issue, but it seems an odd way to go about it, especially given the other reasons not to keep these as a group. Not a criticism, really, just hard to see the reasoning.
On a different but related note: selective breeding and inbreeding are two distinct breeding methods, and neither implies the existence of the other in any given case.