This [[Repertoire statistics|repertoire statistic]] measures the objective strength of your repertoire. It does this by evaluating all the positions in your repertoire within your coverage goal, then taking a weighted average of those positions, with the weight being how common those positions are expected to be at your level. By doing it this way, you get rewarded more on this measure for high-eval positions that are also likely to happen. Having a bunch of sound computer lines that won't happen often, won't get you a higher score. This raw score is then converted into a percentile. We do this by looking at your peers (people of your same rating range on Lichess) and seeing how your repertoire stacks up to theirs. Chessbook users have *very* good repertoires vs their peers, so don't be too surprised if you surpass 90%+ of your peers pretty quickly.