I'm familiar with the gambler's fallacy. I wasn't being very clear about what I had in mind when I referenced regression to the mean, so here's my model. If someone has been to a few good therapists (better than the typical therapist) and they haven't been able to help, then "find a better therapist" would be tough advice to follow. But if they've been to a few lousy therapists (worse than the typical therapist) then finding a better therapist should be doable, since they'll have a good shot at doing that even if they just pick a new one at random.

Alternatively, if they've been to a few good therapists but haven't found the right one, that's evidence that "the right therapist for me" is a small category, but if they've been to a few bad therapists and haven't found the right one, "the right therapist for me" could still include a fairly large subset of therapists.

Edit: The more general point is that being unlucky is a reason for optimism, since it means that things are likely to get better just from your luck returning to normal.