You are creating strawmen of your own degsme. The point of the OP is that, it is hard to argue that an oral exam where the applicant's and the evaluator's race is readily apparent is more race neutral than a written exam designed to be race neutral in which the race/age/gender of the applicant is not known and there is no reliance on subjective opinion.

That is not a radical position. A nearly universal requirement for employment applications is you not ask applicants for protected class information. The reason being that you want the selection of applications to be protected class neutral. Most federal government applications don't even use applicant's names once they get to the selection process, instead assigning them a number. This is because there is a long recognized truth that NOT knowing the applicant's race/gender/age/etc... is less likely to lead to discrimination than when an applicant's protected class status is known to the selector. It is just common sense.

So it is a tough argument that the written test designed to be race neutral is the problem, and that the verbal test in which race is readily apparent is the "good" part of the test.

Your statement that there is testimony "on the record" that the written test was biased is meaningless. There is also testimony "on the record" that it wasn't. And if memory serves the Supreme Court did not find racial bias with the test.

But please enlighten us, exactly which parts of the written test were biased and how?