Residuals

Subtract what you already expected, then plot only what is left over. Instead of asking how big the number is, you ask how far it sits above or below the prediction, so ordinary variation flattens out and the surprises stand up.

Subtract what you already expected, then plot only what is left over. Instead of asking how big the number is, you ask how far it sits above or below the prediction, so ordinary variation flattens out and the surprises stand up.

Example

Your weight chart still wanders because weekends always run high. Build the expected value for each day from your own weekday pattern, then plot actual minus expected. Now the 2.1-pound spike after the Thursday work trip pops out, while normal Saturday drift sits flat at zero.