lukestein’s Twitter Archive—№ 15,898

@cameron_pfiffer Unbelievably, my model has R squared of 1 out-of-sample!* *n=2 and DGP known to be linear ex ante