The Robin Hanson blog post points out the exact same thing. Why is a positive correction for women useful, but a negative one "not adding anything to the discussion"?

If you want to use gender as a predictor, it could be positive or negative.