- Divide training set into
equal-sized portions (“folds”) - Train
models, where the -th model uses the -th fold as a validation set and the rest is the training set - Generally considered the “default version” of cross-validation
- Still subject to some variance due to selection