This investigation examined whether the mode of locomotion matters in how 8-, 10-, 12-, and 14-year-old children (N = 91) judge dynamic affordances in a complex perception-action task with significant safety risks. The primarily European American children in the sample came from the area of Iowa City, Iowa and were balanced for gender. The same children crossed a single lane of continuous traffic on foot and on bike (order counterbalanced) in identical immersive virtual environments. We found that although 8-year-olds chose significantly larger gaps when crossing on bike than on foot, these gaps were not large enough to compensate for their delay in entering the gap and their slowness in crossing the road. As a result, they ended up with less time to spare when exiting the roadway on bike than on foot. In contrast, 14-year-olds exhibited no difference in their gap choices on bike than on foot, nor did they exhibit a difference in their timing of entry into the gap. However, they crossed the road much more quickly on bike, resulting in significantly more time to spare when crossing on bike than on foot. The 10- and 12-year-olds' performance fit neatly between that of the 8- and 14-year-olds. We conclude that as children gained better control over the bike with age, they were better able to match their gap decisions with their crossing movements such that bicycling afforded even safer road-crossing than walking for 14-year-olds.