Why every statistician should know about cross-validation (robjhyndman.com)

It's sad to hear that some statisticians don't think they need cross validation. It's not a vapid trend: there are sound statistical reasons to prefer it over other resampling methods like the bootstrap in many applications.