Here’s one thing that was remarkable about Wednesday night’s Republican presidential debate in Boulder, Colorado: It was just the third debate. By this point in the 2008 cycle, there had already been 12; at the same point in 2011, there had already been eight. Thus, in the 2016 cycle the public has had less chance to see the Republican candidates interact than in other recent cycles. All of this is by design. Following the 2012 cycle, the Republican National Committee took control of the debate process, vowing to cut the number of debates. In their view, the extensive debates that cycle hurt eventual nominee Mitt Romney. They attributed the the rise of other candidates to the debates, when in fact the opposite could just as easily have been true. With that many debates, alternatives to Romney did have the opportunity to rise, but Romney also had more chances to knock them down. The same thing might be happening in 2016. Following each debate, Donald Trump’s poll numbers have dropped, and other candidates have improved their standing. The limited number of debates this cycle hasn’t offered the rest of the field — which is much larger than the 2012 field — sufficient opportunities to knock down Trump’s numbers. Jeb Bush’s polling numbers, rather than rising and falling along with the appearance of alternate candidates as Romney’s did in 2012, have slid steadily since Trump entered the race. The fact that fewer debates have been held is not the only flaw in the Republican National Committee’s plan, however. Another mistake was to determine eligibility for debates based on polls — which the pollsters themselves say is a mistake. By basing debate eligibility on polls, the RNC is essentially creating a national primary election that is completely fictional. In the United States, we don’t have a truly national presidential election: we have 50 separate state elections for president. During the general election, it’s easier to gloss over this fact, as all the states now at least vote on the same day. In that election, national polls have a greater significance, but they are still not the be-all end-all: swing states matter the most. Primary elections, though, are a different matter. States hold their elections on vastly different days, from the Iowa caucuses in early February to the District of Columbia primary on June 14. Many of the later primaries will be largely irrelevant, yet in the national polls there are as many respondents from those states (and perhaps more) than there are in the hugely important early primary states. Right now, poll respondents in states with irrelevant primaries are having an effect on who appears on the presidential debate stage. Regardless of whether that might make sense in an ideal world, it doesn’t reflect political reality in the slightest. Using polls as the basis for debate eligibility also introduces the observer’s paradox — where observing a situation produces different results — into politics. At each debate, candidates know that they are trying to get a slot at the next debate, and this affects their behavior. Even though Trump has been hurt by previous debates, the candidates who have attacked him most forcefully — such as Jeb Bush and Rand Paul — haven’t gained in the polls. Counterintuitively, this discourages other candidates from challenging him directly, rather than encouraging it. If anything, this election cycle shows that it is difficult for the national parties to try and structure the race in favor of a particular kind of candidate. We saw this in the 2008 Democratic primaries, when the establishment was convinced that superdelegates would create a firewall for Hillary Clinton against any challenger. That proved to be false then, just as the establishment’s hope that fewer debates would help their chosen candidate has proven to be false today. Hopefully, the lesson both parties will take from this is to refrain from further meddling in the primary process.