Even if you're right, I'm not sure what evidence it's supposed to count as; everybody knows we make bad calls, and I know it better than anyone. But I searched for your incident and could only find https://news.ycombinator.com/item?id=14530877, which is close to what you describe and (I think) would strike most HN readers as pretty defensible.
The point is that if there is a bias, it would be that moderation seems to target those that disagree when people bring up sexism and racism in unrelated article discussions. It seems very safe, even profitable in terms of votes to bring up sexism and racism, which would be contrary to the views of matt4077.
I predict that if moderation was a bit more consistent in keeping out discussion about sexism and racism in unrelated articles (unless exceptional new insight), we would actually see less reader fatigue and lower amount of flagging for explicit articles about sexism and racism.
(Just to be extra clear, I don't disagree with the moderation itself on the two above examples if they had been the ones bringing in the topic of sexism into those threads).
These calls are somewhat arbitrary because when a thread degenerates, you can always trace the root of the degeneration further upthread. But moderation is more like pruning a tree than excising a tumor. The goal isn't to get rid of all malignancy—that would be too draconian and is impossible anyway. It's to detach the bulk of unwanted growth.
People can reasonably differ about individual cases, but the principle here ought to be pretty solid for anyone who has read https://news.ycombinator.com/newsguidelines.html and wants to use the site as intended.
For an article a about work and unrelated to sexism, gender or racism, a commenter wrote out of the blue that all white men live privileged and easy lives. A second person replied and disagreed, arguing that some men don't have everything perfect and easy from the point of birth to death. The reply got moderated for causing a flame war, with third-party people asking dang why he targeted the dissenting view and not the original comment.