One of, if not the, biggest problem with Presidential debates is that they really aren't. Candidates frequently, at the Presidential level, engage in debate camps beforehand. I link to a story about Barack Obama setting up his 2008 camp, but it could have been anyone. We could even link to a West Wing episode about it. They are peppered with whatever questions that they deem the moderator as most likely to ask. They rehearse answers, including rebuttals to what the opponent is likely to say, and how to steer an answer from an uncomfortable topic to a more comfortable one. The result is largely 90 minutes of kabuki theater, with the bulk of the drama centering around who messed up their script and who got off the best zingers.
Not far behind is another major, yet underrated, problem: all sides involved are fighting the battles of the previous term. Questions from a moderator will usually, nearly always, revolve around major topics that have been fought over during the previous 2-4 years. Sometimes, these topics will carry over to the next term as well.
But sometimes they won't.
Monday morning quarterbacking the previous term- 'well, I'D have done things this way, and if you vote for me, I will'- would be fine if issues carried over from one term to the next. And as we just stated, sometimes they do, issues such as the economy and any ongoing wars. Time, however, marches on. Sometimes the issue turns out to have already been settled, and will not come up to a large degree in the next term. Sometimes whether the issue is settled depends on who ultimately wins the election. And the next term is always filled with events which have not yet happened, and will push issues to the forefront that were never covered in the debate and which nobody focused on.
Case in point: how many of you over the course of the previous election figured that Egypt, Tunisia and collective bargaining were issues that merited serious examination before you cast your vote? Whoever raised their hand is a liar.
Obviously, trying to anticipate these things is nigh-impossible to do reliably. Attempts were made to mix up the topics over the course of the 2008 Presidential primaries by allowing people to submit questions online, but the online entries were either fighting the previous term as well, or asking about obscure niche topics that would never in a million years become a hot-button issue.
Both of these issues- the debate-camp problem, and the previous-term problem- can be addressed, maybe not solved entirely, but improved upon, with a debate format I call 'September Surprise', after a month debates will commonly take place in. (Yes, I realize we're a year and a half away from there. Like that's going to stop me.)
September Surprise's premise is to attempt to simulate, as best as possible, the next term, a period filled with unforeseen events that can cover a much wider range of topics than one is currently considering. The first step, of course, is for a moderator to try to figure out what might be one of those topics. This, you do by compiling a list of topics. Obviously, the current hot-button issues would have their place- Iraq, gas prices, partisanship, budget deficits, Medicaid, and so forth. But also in the list should be a wide range of topics that, while they may not be big now, one could reasonably consider to be an issue that might flare up under the proper circumstances or is at least is of sufficient importance. These would be things like tobacco, health food, imported goods, homelessness, bridge infrastructure, any country that stands a fair chance at popping up in the news regularly. I've tried this myself, and I got into the 150's before I got bored and called it quits. I wasn't done. I just got bored.
You can't just toss any topic into the list. Some things will always be minor issues or fringe issues. The gold standard, for example. That is just not going to be a hot-button topic. School uniforms, Swaziland, and college football playoffs may have their place, but they're not going to be big huge things. They just aren't.
Step 2 is to construct one question pertaining to each topic on the list. (Much as one might like to see what happens, you can't just go 'Candidates, you now must discuss corporate personhood. Go.') As the debate goes along, pick topics one at a time, right in front of everybody (use a random number generator, a bingo hopper, slips of paper out of a hat, however you want to do it), and read the corresponding question. If Afghanistan gets picked, fine and dandy. If gun massacres come up instead, or airport security or Taiwan or patent law or multilingualism, well... surprise.
No debate camp in the world is going to be able to completely prepare for that. However, since there are too many topics to anticipate, focus would likely be shifted to trying to steer things toward their preferred topics. But that defeats the entire purpose of September Surprise. A candidate who is unwilling or unable to stop fighting the previous battle is a candidate less prepared to handle the next battle. The moderator must therefore be vigilant about keeping things on topic. If a candidate jackknifes a question about the tax-exempt status of churches into a question about the Bush tax cuts, it's the moderator's responsibility to jackknife it back to the churches, interrupting the candidate if absolutely necessary.
After all, some of these things might become important later.