R

Ridge, lasso and elastic net returned identical test error, to the last digit. The reason is a mistake hiding in a workflow you have copied.

Your model says 1.60. That cannot be a probability. Four things that go wrong reading logistic output, and none of them throw an error.

I removed 53 outliers. R-squared rose from 0.818 to 0.856 and the model got worse. Every number that improved ignored the rows I deleted.

Your slide says $368. A colleague who ran the same data says $422. Four ways a regression coefficient means less than it appears to.