A persistent challenge to technology-enhanced learning is a general need for evidence-based ecological validation protocols―especially as it pertains to transfer-of-training for serious video games and virtual simulators. It is often missed that the critical objective in achieving ecological validity in virtual simulation trainers is not necessarily to maximize realism but to optimize training by targeted and deliberate stimuli, and that validation must be against operational simuland referents as well as affordance referents. That is, validation must ensure not only that the body of knowledge of real operational environments being simulated (i.e., simuland referents) is properly selected, specified, and codified in computable models but also that quality measures of human sensory detection instigated by virtual environments are consistent with those afforded by corresponding real/live simulands (i.e., affordance referents). In this first-of-its-kind exploratory pilot study, a part-task hybrid simulation-couple paradigm (i.e., virtual simulator in tandem with mechanical simulator counterpart) was implemented to garner and establish oculomotor affordance referents relative to ecologically-valid 60-mph live fastballs in the baseball batting task as well as to instigate and measure horizontal eye movements relative to corresponding virtual fastballs. Results indicate virtual fastballs in full-size 3D stereo theater afford similar overt attentional responses as live fastballs at modest speeds but not at faster speeds. Onset times of catchup saccades, succeeding smooth pursuit, while observing off-center incoming virtual fastballs, lagged those while observing live fastballs―suggesting possible accommodative-vergence impairment induced by the fixed focal plane of 3D stereo display. Results offer nominal evidence of limited ecological validity for 3D stereo virtual environments involving horizontal eye-movement tasks while observing objects moving in depth―demonstrating the relevance of affordance referents in selective ecological validation of virtual stimuli.