It appears that range ratings are dervied from a version of (PO+A+E)/G.  This works fine for many players, but leads to very misleading results for players who were used as defensive replacements.  Ironically, it often makes the very best defensive players in baseball look like the very worst.

A first baseman who makes 3 plays in 3 innings as a defensive replacement should be considered to have a range of 9, not 3.  Examples of players who are unduly penalized by this include:
Tony Muser (especially 1974-76)
Mike Squires (1982-84)
Mark Belanger (1977-81)
Rafael Belliard (his whole career, but especially 1991-92)
And many others.

www.baseball-reference.com posts range per 9 innings, in addition to range per game, for many players.  This corrects the problem.  Even if the data is not available for every player in history, it makes sense to use the best data we have for the players for whom it is available.