It doesn't matter. It isn't first case when 2 trainings of FT aren't enough to pop, so it's logical, that one training shouldn't give more than 0,49. At the same time I never heard about case when 3 training aren't enough, so it's obvious, that one training of FT shouldn't give less than 0,34. And because in most cases we see pop after two trainings, so it means that accurate value should be closer to 0,49 than 0,34. Maybe it's 0,41 or 0,48 - who cares, I think 0,45 is more than sufficient estimation. It is enough, important is to know, that one training of FT gives little less than 0,5 level.