[kcoul]

Exactly - I have done a number of research studies focused on how musicians do or do not develop robust memory models of the pieces they learn, based on how they practice, and the data consistently pointed to a correlation between robustness and practice approach.

Anyone who has taught piano for example knows the most common mistake of novice students is to omit sharp or flat notes (the black notes) when learning notated music which begins to include them. In this case, the student will practice the wrong notes for long enough before their next lesson that the wrong notes become engrained in their memory model, resurfacing under duress such as in recital. There are similarities for other instruments (strings: intonation, woodwinds: embouchure, percussion: rhythmic accuracy, etc.)

The most talented students seem to gravitate consistently towards robust memory models for their respective instruments technique as a way of freeing themselves up as quickly as possible for the more enjoyable aspects of perfecting a piece of music: refining expression.

Perhaps one day there will be tools which can assist those less naturally predisposed to developing robust memory models, before it's too late for their brain, the way the most talented students do.