Food-hoarding patterns range between larder hoarding (a few large caches) and scatter hoarding (many small caches), and are, in essence, the outcome of a hoard size–number trade-off in pilferage risk. Animals that scatter hoard are believed to do so, despite higher costs, to reduce loss of cached food to competitors against which they cannot defend their food reserves (henceforth: superior competitors). We tested the underlying assumption that the cost of having more caches under scatter hoarding, thus increasing the likelihood of cache encounter by superior competitors, is outweighed by the benefit of having small caches that are less likely to be detected upon encounter by superior competitors. We carried out a controlled experiment in which we distributed a fixed number of acorns over a fixed number of patches within a fixed area, varying cache size and cache depth, thus mimicking alternative hoarding patterns. We then recorded cache pilferage by a fixed number of wild boar, a well-known pilferer of acorn caches. The time wild boar needed to pilfer the first cache was shortest for scatter hoarding, but the time needed to pilfer all caches was slightly longer for scatter hoarding than for larder hoarding. Overall, however, the rate of pilferage did not differ between scatter hoarding and larder hoarding, and was not affected by cache depth. We conclude that the effects of alternative hoarding patterns on reducing cache pilferage by wild boar were smaller than expected, and that superior competitors may thus not be important drivers of scatter hoarding. Instead, other factors, such as conspecific pilferage or the risk of cross-contamination of food items in large caches, which can also cause catastrophic loss of food reserves, may be more important drivers of scatter hoarding.