The present research on risky decision making extends the exploration of the singleplay/multiple-play distinction. Mathematics students are asked to respond to two choice problems, each having three decision tasks: choosing between multiple-play gambles, choosing between single-play gambles and matching paired possible outcomes. It is shown that the generalized model of Expected Utility (EU) theory appears to be redundant for the multiple-play situation, given that the simpler Expected Value (EV) theory correctly predicts the observed behaviour, and in the single-play situation it appears to be predictively inadequate. The observed choices in single-play situations could be better accounted for by the equate-to-differentiate approach revealed by the matching data. The overall results suggest that the long-run and short-run perspectives are in fact so utterly different that the short-run one is perhaps not based on any kind of expectation rule.