BRS glitch tag shouldn't be applied when that BRS isn't used for sensor correction

This is also something that Niko and Jim and I already talked about, I'm writting it here just to help track all the things that are going on.

At the end of O3a BRSX was railed, and removed from sensor correction. It started glitching and some locklosses were marked as BRS_glitch. Niko has an idea of how to change the analysis to prevent this in the future.