Using the Take Back Half approach to effectively “take back” the error terms for theoretically zero integral nonlinearity.

How do we take how we feel during a run, and make it something measurable, repeatable and something we can use in training?