That's not how epicycles were used. Ptolemy set up the standard geocentric model with epicycles (and deferents and equants), and these were used more or less unchanged for over a thousand years. When, in the late medieval period, some Arab astronomers began tinkering with modifications, these were motivated by a desire for better agreement with the Aristotelian ideal of uniform circular motion (i.e., to somehow replace Ptolemy's non-uniform equant device with some combination of uniform circular motion), not any desire to match the measurements better.