Exceptions to generalisations are easy and obvious to spot when they clearly violate the general rule, ie a human with one arm. This is also known as a positive exception, where the general rule obviously needs to be enhanced by an additional rule. However, consider a different type of exception: imagine the generalisation that all pub-goers drink both dark and light beers. How easy would it be to notice that a particular person only drank light beer without explicitly being told? On each occasion, the observed drink fits in with the generalisation that has been made, and the rule is never obviously violated. This is known as a negative exception, where there is an event allowed by the general rule that is missing, ie never actually occurs. This research investigates how people learn such negative exceptions to general rules. Here, computer-based games will be used to investigate the general rules and exceptions to be learned. The games will track what sort of general rules people form, and when people learn exceptions to those rules for common exceptions vs. rare exceptions.