Political science’s common understanding is that elections are normally genuine contests -- that is, that there are losers. The prototypical statement is the district-level application of “Duverger’s Law,” which Cox has extended to SNTV election rules in multi-member districts of M seats (including standard plurality-rule single-member districts, at M=1). This conventional wisdom holds that Duvergerian strategic processes tend to whittle the field of parties down to M+1, but then grind to a halt once M+1 remain, preserving competition. But no endogenous mechanism can explain why these processes should halt at all. As I have argued elsewhere, Duvergerian processes, and strategic party entry in particular, tend to drive the number of competitors down toward M, not M+1. That is, they tend to undermine competition, even in “competitive” party systems. Contested races – including those with M+1 parties, not to mention those with more than M+1 – represent “non-Duvergerian” equilibria. They are sustained only by “friction” exogenous to Duvergerian processes: either impediments to short-term instrumental strategic decision-making, such as uncertainty, or abandonment of strategic decision-making outright. But all this tells us is that in equilibrium, elections can be either uncontested or contested (with varying degrees of closeness). The theory says nothing about how often either type should occur. This is an important finding in itself: it undercuts the idea that uncompetitive races are exceptional or out-of-equilibrium cases. But it is also unsatisfying. If we care to think about what constitutes a “normal” proportion of uncontested districts or a “normal” proportion of closely-contested districts, or about what institutional features might shape these proportions, theory provides no benchmarks. I attempt to wrest some amount of predictive traction from this theoretical indeterminacy through a set of simple simulations. In each, I construct a hypothetical set of districts, and randomly assign values on various parameters of interest, including district magnitude, the number of potential entrants, potential entrants’ relative strengths, how keenly voters can perceive these strengths amid uncertainty, and how strategically voters and parties tend to behave. In practice, of course, more than pure randomness may be at work, but these simulated districts yield generic patterns of heuristic value, and allow for clear illustrations of comparative statics. The results show that we have strong reason to expect frequent uncompetitive elections, even under “reasonable” assumptions about uncertainty and district sizes, and especially in SMDs. I then show that the simple causal mechanisms driving the simulation also are likely driving parties’ electoral strateiges in practice. Real-world distributions of uncompetitive and competitive elections, and how these vary across different sets of institutional features, strikingly resemble the simulated results. I examine a large, richly-varied, and novel set of Japanese Lower House, gubernatorial, and prefectural assembly elections, as well as U.S. Congressional elections. [ABSTRACT FROM AUTHOR]