Example outcomes resulting from repetitive input patterns of length 3 and 729.

Example of outcomes resulting from repetitive input patterns of length 3 and 729 under two task conditions: Pattern matching (when the animat evolved in an environment where it had to predict the next element of a sequence composed of a repetitive string of length 3 or 729), and Random sequence (when the animat, after evolving under a repetitive string of length 3 or 729, had to predict the next element of a completely random sequence). For sequences composed by short strings (3 digit long), the animat predicts all the elements correctly (pattern matching), but does probability matching when faced with the prediction of the next element in a shuffled random sequence. When the input sequence is composed of a very long repetitive string (729-digit long), the animat is not able to learn it, adopting a perseveration strategy, making many (expected) mistakes; but when the same animat has to predict the next element of a randomly shuffled sequence, it perseverates as well, achieving a better performance in comparison with the animats that had been able to learn a short-patterned sequence (3-digit long).