Oh, FFS, it's imperfect because virtually no line stays together very long in the NHL. I set the sample so that you'd get 128 lines exactly, which will get you statistically valid rankings 1-32=1st line caliber, 33-64 2nd line, 65-96 3rd line, 97-128 4th (the math was easier when there were 30 teams, damnit). Some teams will have more lines in there, some teams will have fewer just because there are only a half dozen lines in the NHL that stay consistently the same over a season, but this gives you the rough positioning against other consistent line combinations. That way, you get a great idea of where any given line stands against the rest of the league even if it's imperfect. The fact that even our temporary injury lines appear there show that Brind'Amour changes lines far less than other teams, which is interesting in itself, but this is a problem with dozens of variables and I'm just holding a bunch of the less relevant ones constant to get a rough picture of reality.