The way I've studied this in the past is to look at ratio compared to peak winning percentage, the per-minute version of my wins above replacement player (WARP) metric. A few years ago, I put together a list of players who had played at least nine seasons and at least 10,000 minutes and ended their careers between 2005 and 2010.
That list had 21 players I labeled point guards. They divide rather evenly into 10 players 6-foot-2 or taller (the tall group) and 11 who were 6-foot-1 or shorter (the short group, and at 5-foot-9 I mean nothing disparaging by that). While those sample sizes are smaller (shorter) than ideal for this kind of study, the results are fascinating.
While the smaller point guards were for the most part effective immediately --
Brevin Knight's best season was his rookie year at age 22, while Damon Stoudamire's came in year three at age 24 -- the larger guards generally took longer to develop.
Rod Strickland peaked at age 29,
Gary Payton peaked at age 30, and Cassell did not reach his peak until age 34.
So as a group, the taller point guards didn't hit their peak until age 30, much later than the smaller point guards (age 26, a little younger than the typical age-27 peak across all positions). Every year after age 27, the taller point guards were relatively better, often substantially so.