The oddball duration effect describes how a rare stimulus amongst a string of standard stimuli is perceived to have a longer duration than the standards, even if they are of the same objective duration. Several theories have been proposed to explain this phenomenon. In order to adjudicate between opposing explanations, we have borrowed three extensively studied paradigms from the variable foreperiod literature: the sequential foreperiod, temporal cueing and a skewed foreperiod distribution. This approach allowed us to examine the effects of positional expectation on perceived oddball duration, while avoiding confounds from first-order positioning of the oddball in a sequence of standards. Through these three experiments, we demonstrate a clear role of positional expectation in the lengthening of the perceived duration of an oddball. We show that this expectation effect is separable from other drivers of the oddball duration illusion.