A hard truth, one that only a man with zero chance of winning Iowa or New Hampshire would dare utter.

I like the idea of mixing up a stale primary schedule, though. It’s strange that the same two states enjoy outsized power over the nomination process cycle after cycle just because tradition requires that they vote first. If they were hugely populous or strategically important in every election, it’d be easy to justify. But really, they continue to vote first because that’s Just The Way Things Are.

We could change it! There are at least three alternatives:

1. Reorder the states to reflect the demographic diversity of the Democratic coalition. This is less of a concern for the GOP primaries since the right is overwhelmingly white, but the left is not — and yet it’s two nearly all-white states that get first crack at choosing the nominee every four years. That’s Castro’s complaint. He understands only too well that if Texas or Arizona or another state with a large Latino population were voting first this year instead of Iowa that his chances of becoming the nominee would be much higher than the <1% he’s at now. The demographic imbalance in the Democratic primary is somewhat eased by having South Carolina as the third state to vote, since the primary electorate there is majority black, but that doesn’t help Castro much. And it’s usually the case that if you lose Iowa and New Hampshire, a win in South Carolina won’t overcome those losses and ultimately send you on to the nomination. That’s how important the first two are.

2. Reorder the states to let the battlegrounds go first. We all understand by now that our “national” election is really just 50 state elections and that only six or seven of those will be decisive given the certainty of the outcome in the rest. The obvious thing to do is to let all the swing states go early. If the election will come down to who’s strongest in those particular states then, er, it makes sense to find out in advance which candidate is strongest in those states. Democrats could tweak the rules so that independents in each state are eligible to vote in the Dem primary without having to register with the party, too; the results will thus give them a clue about which candidates are most likely to attract swing voters against the Republican nominee. And it’ll get indies in those states invested in the Democratic presidential process early, which may produce more votes for the party in November. Have Florida go first, then Pennsylvania, then Michigan.

3. Forget ordering the states at all and hold a single national primary. All 50 states vote together on Super Tuesday. That’s the only way to prevent the outcome in one from influencing the outcome in others. The potential problem with this scheme, though, is that a national primary will favor better known, better funded candidates, which means exciting outsiders who’d stand a real chance at building momentum towards the nomination over a series of primaries would be squelched in a single evening. There’s a case to be made that each of the last two presidents wouldn’t have been their party’s nominee in a “national primary” system since each wasn’t taken completely seriously at the start. Trump might have won anyway, although he surely benefited from the opportunity to string together a sequence of wins in New Hampshire, South Carolina, and so on that proved to doubters that he had staying power. Obama is more questionable, though, since it was his win in Iowa in 2008 that convinced skeptics that a black candidate really could contend seriously for the nomination. He leveraged that belief into a mammoth win in South Carolina a few weeks later. If Iowa and South Carolina (and everyone else) had voted on the same day, black voters may have gone to the polls doubting Obama’s chances and ended up sticking with Hillary instead.

An obvious possibility given all of those options is to combine elements of each. You might not want a national primary — but what about letting several states vote first as a group on the same day, taking demographic diversity and battleground status into account in choosing which states should have the honor? Iowa’s and New Hampshire’s bruised egos could be soothed by including each in the first group; to those we might add Michigan, Arizona, and South Carolina. Two weeks later another group of five would vote: Nevada, Wisconsin, Texas, Georgia, and Maine, let’s say. Different candidates would focus on different states in each group, raising the risk of a protracted stalemate, but I think enough buzz would build for the winners after the first and especially second rounds that the field would narrow quickly. Better-funded candidates would obviously enjoy an advantage by being able to compete in multiple states, but the small groupings would minimize that and at least avert the sort of major imbalance you’d see in a national primary. I think it could work.