Sure thing. First, put your data into a time‑series table: weight, reps, rest, session date. 1. Run a simple linear regression of weight vs. session number to see the trend slope. 2. Overlay a rolling average (say 5 sessions) – that’ll smooth noise and show the real trend. 3. Calculate the residuals; if the residual variance starts dropping, you’re moving past a plateau. 4. For a probabilistic hit, fit a Bayesian updating model: prior on your PR probability, update with each session’s result. 5. Finally, run a logistic regression with predictors like previous session weight, rest days, and cumulative volume; the predicted probability > 0.8 is your “statistically likely” PR window. Keep iterating as you collect more data—every new lift is a new data point to refine the model. Good luck!