Interesting logic. Of course people's resumes don't reflect the entirety of the candidate, so it is likely that basing decisions to move further based on race will eventually cause you to pass up the better candidate. Sounds like a fault to me.

The alternative of ignoring race gives you two identical candidates. If choosing a particular race consistently gives you worse results than choosing randomly would have, that's still evidence supporting racism as a general rule - you're just being racist in the wrong direction.

I just don't buy that it does. We're not picking two random people off the street here. These are people who have passed all the filters needed to have their resumes sitting in front of you. From education, to experience, to HR, to phone screen, etc. If, at this point, the two candidates still look identical, I don't think race gives you any meaningful signal. The black candidate has already proven that he has risen above any disadvantages he might have statistically faced. In fact, this should even be a plus in his column because he had the determination (or raw intelligence) to overcome his disadvantaged lot in life and be sitting in front of you, on equal footing with a white candidate who (statistically) had many advantages given to him.