The author of that blog seems to give advice based purely on statistical correlation without making any effort to distinguish between correlation and causation. See this post for example.
Well the author of the blog is trying to get people interested in the website, so it should be viewed as marketing more than advice. Advice just sells better than pure stats.
It is still the only source of statistics on what goes on behind the scenes of okcupid, so it should provide some useful information, if you are interested in mainstream dating. It is also interesting purely from a sociological point of view, which races message the other races.
It also provides the people on okcupid a common topic of conversation, if they are moderately interested in statistics.
This thread is for the discussion of Less Wrong topics that have not appeared in recent posts. If a discussion gets unwieldy, celebrate by turning it into a top-level post.