Effective communication of risks requires a theoretical understanding of the influences of cognitive and task constraints. We present progress toward the identification of performance parameters in an experiment that explores the dynamic interplay of four key risk communication variables (i.e., cognitive strategies, numeracy, complexity, and format). Specifically, we conducted a protocol analysis to trace the types of cognitive strategies used when comparing the helpfulness of two treatments. Variability in cognitive strategies was also examined as related to (1) format (expressed either as absolute or relative risk reductions in either a frequency or single-event probability format); (2) task difficulty; and (3) numerical skill. Results indicated that highly numerate people often effectively used more complex strategies. However, the performance advantage of highly numerate people only existed when comparing two relative risk reductions (which requires a complex strategy), but not when comparing two absolute risk reductions (which requires a simple strategy). A frequency format was also found to produce additional benefits in very difficult tasks (i.e., when comparing absolute and relative risk reductions). Generally, although strategies and accuracy are influenced by many factors, risk communication tended to be most transparent when presented in terms of absolute risk reductions.