The state or quality of being perplexed.
Something that perplexes.
(information theory) A measure of how well a probability distribution or model predicts a sample.