I think what you're missing is that a tiny minority of toxic users (as perceived from your viewpoint) can easily be perceived as a high concentration of harassment from those that are on the receiving end of abuse.
This asymmetry of harassment is a dynamic worth thinking about more broadly; for example, if only one in a thousand men cat-calls women, you might not even have any male friends that have ever cat-called, and yet that might mean a particularly attractive woman might pass a cat-caller many times per day. Small differences in the rate of cat-calling (1:1000 vs 1:2000) won't be detectable to you, but can represent a huge difference (once a day vs twice a day, say) to those sampling far more frequently from the distribution than you are.
More specifically to the thread here, the question is not whether HN is well moderated, or whether toxicity is quite low-frequency from your perspective. The relevant question is whether the base-rate of anti-trans harassment is higher than in the general population.
It's entirely possible (I think it's true) that people in this community are in general more civil to each other than on most of the internet, and I think the thoughtful moderation is a big part of that. But it's also possible (and I don't have any direct experience to claim whether it is or not, so this is just numbers to illustrate what's possible, not a claim of how things are) that at the tails of the distribution there's 10x more anti-trans harassment in the community too, and we'd simply not be in a good position to observe or measure it unless we were on the receiving end of it or auditing lots of comments very closely. Honestly, based on the tenor of responses when this specific topic comes up (quite different to the normal HN vibe IMO), that wouldn't surprise me either.