The process(es) by which hunter-gatherers learned novel food resources as they dispersed across previously unknown landscapes, such as Late Pleistocene North America, are not known. A key trade-off in the process must have involved the decision, once high-ranked resources in a patch declined, to stay in place and learn about untried food resources or move to another patch to exploit familiar resources. Modelling the process suggests that learning a new resource will be more likely when the handling time for that resource by a novice forager is not too much higher than that for a practiced forager, and when patch residence times are long enough for within-patch resource depression to occur. The decision to move or stay would likely have varied depending on the richness and structure of the resources in different patches and habitats. The model further indicates it is not large mammal hunting per se that favors the most frequent residential moves by central place foragers, but rather the pursuit of the rate-maximizing suite of resources, whatever the composition of that suite may be. These findings imply, as previous empirical work has suggested, that the dispersal process likely played out in different ways across Late Pleistocene North America.