17’s A Crowd: How to Structure the GOP Debates

There’s a somewhat famous interview question, rumored to be used at Google, that’s become a bit of urban legend. There are some slight variations, but they have almost the same structure. The question goes something along the lines of:

“You have 25 horses and no stopwatch. Using a racetrack that can only run five horses at a time, what is the minimum amount of races needed to determine the top three fastest horses?

With Scott Walker joining the Presidential race, the amount of candidates in the Republican field has grown to fifteen, with two more expected to announce before the end.

Fox News, which hosts the first Republican primary debate, may very well be asking its interview subjects:

“You have 17 candidates. Using a debate stage that can only hold ten candidates at a time…”

As of now, the powers in charge have determined that the answer is one. And they’ve got some pretty counter-intuitive rules to determine the candidates selected to debate.

Is there a better way of doing this that doesn’t exclude any candidates? Let’s put aside some ideas, such as having all 17 onstage together. Let’s also assume that the current method Fox uses is flippant, arbitrary, and undemocratic.

CNN has come up with the idea of hosting two debates, one with the “top-tier” candidates and one with the “bottom-tier” candidates. This set up is almost as insulting as Fox’s current method, so let’s scratch that.

So, assuming that 10 is the maximum we can have on stage at a given time, what would make an equitable, fair, and informed debate process?

  • Candidates all facing each other: While we can’t have it so that every candidate will face each combination of candidates-there are 19,448 combinations of 10 out of a 17 person group, it is easy to ensure that every candidate will face the other.
  • Candidates at different levels of support in polling are on stage at the same time: To avoid a “bottom-tier” debate that would draw little meaningful interest, and marginalize candidates before any vote has been cast.
  • Fair for voters: Voters need to the one with the ultimate say, not the randomly selected members of a Fox News poll.
  • Make it engaging for the public.

There is an item out there can help us create such a system – ping pong balls. Widely used in sports from basketball to soccer, ping pong balls are used to determine everything from which team gets Lebron James to who plays each other in the World Cup.

Here’s how it could work:

There are currently six RNC-sanctioned debates scheduled before the Iowa caucuses in January (not including CNN’s “bottom-tier” debate). If we have a corresponding debate accompanying each currently scheduled one, so as to have everyone debate but keep a manageable amount of people on stage at a given time, that brings us to twelve debate segments total.

With 17 candidates there are a number of ways to determine who is on stage with whom and when.

You could have the candidates drawn into groups (a la the World Cup) with three groups of four and one of five. Then for each debate a draw would occur to see which groups will face each other (with no groups facing each other twice in a row). This would result in one debate segment of eight candidates and another with nine candidates.

This could provide some monotony as you’d have the same combinations facing off with each other. After three debates (six debate segments) you’d have seen every combination. At that point you could redraw the different groups, and if candidates have dropped out by then you have a chance to restructure.

Another way would be to have a lottery for each debate. This would be the most truly random method, however could provide for more repetition if candidates continue to get drawn into the same field. You could protect against this by having different “pots” after the first debate. The candidates who appear in one debate together get put in the same pot, the other candidates get put in a second pot. For the draw of the debate you alternate drawing spots from each pot until you’ve filled up the debate, and the rest will face each other in the second debate segment. That ensures at least half the debate is different than each previous debate.

All these methods of selection can implement a form of seeding. If one is insistent on using perceived front-runner status then you could seed the top candidates into one pot and separate the other candidates into other pots. You could also seed for other reasons that might serve the party or the networks. You could do it based on office (so you don’t have a debate of all governors) or by geographic region (so you don’t have a debate of only southerners) or by ideology, for instance. The World Cup, for example, seeds by both ranking and geography. For those unfamiliar with the logistics of soccer tournaments, here’s an example how such a seeding could work:

We assign each candidate a ping pong ball number and put each ball into its own pot as follows:

Pot A (frontrunners): Jeb Bush, Marco Rubio, Rand Paul, Scott Walker, Ben Carson

Pot B (Tri-state area): Donald Trump, Chris Christie, George Pataki, Rick Santorum. (Pennsylvania is close enough.)

Pot C (the South): Mike Huckabee, Rick Perry, Bobby Jindal, Lindsay Graham. (Side note: you could have a draw to see which Southerner gets bumped out of this pot, similar to what the World Cup did with European countries this year. In this hypothetical random drawing, Ted Cruz gets bumped.)

Pot D (other): John Kasich, Jim Gilmore, Ted Cruz, Carly Fiorina.

Then you pick a ping pong ball out of each pot until you have filled the debate stage.

I did a quick random drawing just to see what kind of debate field this seeding might generate. Here’s what I got, in order of their draw:

Debate 1: Marco Rubio, Donald Trump, Lindsay Graham, Jim Gilmore, Rand Paul, Christie, Bobby Jindal, Carly Fiorina

Which means Debate 2 is: Bush, Walker, Carson, Pataki, Santorum, Huckabee, Perry, Kasich, Cruz

Both of these debates would be marquee matchups for Fox. They provide a nice blend of ideology and front-runner status. There is no “bottom-tier” debate, and each one is compelling for the viewer.

Soccer draws and NBA draft lotteries garner huge ratings. Imagine the attention that a televised RNC lottery would get. Candidates could stage rallies and lottery-viewing parties (I’m imagining that they wouldn’t use the international soccer parlance of “draws”). The instant reactions, analysis, and discussions in the build-up to the debates would be a huge boon for all the networks, and the RNC as well.

Ultimately, the end product would be far more engaging to viewers, and more importantly more fair to voters. In this scenario, the Lindsay Grahams of the world get to debate foreign policy with the Rand Pauls of the world. And the John Kasichs get to have their voice heard in front of his home state (where the first debate will be held), instead of getting cut off by an arbitrary number designated by Fox News.

Of all the options, this method would provide the best blend of fairness, excitement, and edification (to the extent that Presidential debates can be edifying affairs).

There are a couple of other methods that could be considered but ultimately fall short of the criteria listed.

Tournament style: In answering Fox’s version of the Google question they could have three debates. One with nine candidates, another with the other eight. Then the top four performers in each (judged by the audience members) can face off in the third debate. This method, however, fails to meet the criteria of being fair to voters, as only those lucky enough to be seated in the debate would have a say. And to expand the vote would mean to risk turning the debate process into American Idol.

Or, in keeping with the soccer theme, Relegation style: The RNC keeps CNN’s method of a top-tier and bottom-tier debate, with one exception. The bottom three performers of the top-tier debate get “relegated” to the next bottom-tier debate, and the top three performers in the bottom-tier debate get “promoted” to the next top-tier debate- as judged by the audience or scientific polling of the debate performance. This might make watching the debate more exciting, especially to those who feel that the debates themselves offer little in the way of meaningful discussion or substance. But again, this method faces the problem of fairness to voters.

Plus, waiting for poll numbers to be released is not nearly as exciting as a lottery.

 Cover photo credit:
David Dow/NBAE/Getty Images, http://bit.ly/1AF44bZ.