Volume matching in A/B comparing is important. Usually, louder will sound "better," even by 1 dB.
However, I'm of the opinion that one should choose volume levels to match what one thinks the source probably sounded like. Tool? 100 dB is about right. Mahler at 90 dB? Sure. For big Mahler finales, with a gong on stage for instance (e.g., in the Eighth), even more than 90 dB would not be too loud.
But if you crank it up for small chamber orchestras, or string quartets, or solo piano, or acoustic guitar, etc., you misrepresent the source. To my ears, fine tuning the volume as I begin to listen is as important as fine tuning the balance (which is important for getting instrumental placement and soundstage right).