Your link describes how security lockouts are probabilistic, yes, but it doesn't get into what the probability is. The article we are commenting on does try to get there, by looking at how often ending what scenarios HN users report getting locked out.
Your link is also talking about the no 2FA case, while the article is recommending 2FA with (multiple!) hardware security tokens.
Your link is also talking about the no 2FA case, while the article is recommending 2FA with (multiple!) hardware security tokens.