A method includes logging past event sequences in a knowledge base, receiving a real-time event sequence, comparing the real-time event sequence to the past event sequences to determine a predicted event sequence for the real-time event sequence, and providing the predicted event sequence, a suggested course of action with predicted outcome, and a recommendation of a user who previously encountered the event sequence. In the above manner, use of subjectively characterized and/or raw historic data to match real-time event sequences provides users with both subjective advice from those who had previously faced similar conditions as well as raw statistical predictions as to what is expected to come next. This allows the users to take a course of action that results in the most desirable outcome.