In a recent article, Eastwick, Finkel, Mochon, and Ariely (2007) reported data to indicate that selectivity might be an important factor in determining romantic desire. Using a speed-dating paradigm, they found that individuals who, on average, rated potential dates as highly desirable were likely to receive lower average ratings from their dates, as evidenced by what they termed as negative generalized correlations. However, the dyadic correlations were positive, suggesting that, across pairs, desire was somewhat reciprocated. Eastwick et al. go as far as to claim that “… daters somehow broadcast their unselectivity… ” (page 318), which we find to be a deeply dissatisfying explanation. We present an alternative and more principled approach in order to account for the disassociation between the generalized and dyadic correlations. We implemented a multi-agent model that allows an assessment of the relative contributions of selectivity and matching on ratings of attractiveness. The model suggests that the match between potential daters’ attractiveness is the most important predictor of romantic desire. We believe that Eastwick et al’s (2007) article is just another example of a dangerous pattern in social psychology research: spectacular claims are made on the flimsiest of evidence.