My experience with Banks, beyond the specific error in the Coddington case, is primarily with his "History of Martha's Vineyard", which is on the one hand a phenomenal piece of work but on the other does include the occasional rare mistake. (He incorrectly speculated on the maiden name of one of my ancestresses -- I was able to track down an original record and correct the mistake -- and inaccurately speculated about the origin of another forebear.) I would imagine that the "Topographical Dictionary" is similarly reliable in the vast majority of cases, but that it's always worth checking Banks against other sources.