That's a fair point. My original problem with this is that you'd likely not have the entire copy of breached passwords available for local comparisons. But Troy

That's a fair point. My original problem with this is that you'd likely not have the entire copy of breached passwords available for local comparisons. But Troy Hunt provides that freely so you don't need to use his API. That wouldn't have much latency at all...that fact in combination with the obvious probabilities involved here, yeah I'll concede the point.