It has been argued that there will be more efficient deterrence if no restrictions are placed on the ability of the rich to buy better legal representation than the poor when accused of a crime: such restrictions lead to over-deterrence of the wealthy. We show that the conclusion does not hold...

We study model selection strategies based on penalized empirical loss minimization. We point out a tight relationship between error estimation and data-based complexity penalization: any good error estimate may be converted into a data-based penalty function and the performance of the estimate...