Twistweight is on the order of 14 kg•cm^2. If you're measuring the swingweight and "90° swingweight" to integer precision in kg•cm^2, you could get results of, for example, 325 and 340, when the actual values are 325.49 and 339.51. You'd calculate 340 - 325 = 15 for the twistweight when the actual value is 14.0 kg•cm^2. That's a potential error of ±7% (before accounting for other errors in the measurement), which is quite large. It might tell you something, but it's not precise enough do anything.

Other swingweight machines may have better precision, or you may be able to devise a method to get better precision from an RDC (assuming Babolat RDC) by adding known weights and averaging several measurements.

The method is valid, though. If you know the moment of inertia in two perpendicular axes, the moment of inertial of an axis perpendicular to the other two axes is the difference between the known values.

My Microgel Radical MP's (re-issues) measure a difference of 12.7 to 13.0 kg•cm^2 using this method and a swingweight machine I'm developing.