After this years disaster, the Republicans are reportedly considering some retuning of the primary calendar. What’s on the table seems pretty marginal, see here.
But, if you think that early primary states have the most influence, and you want to win the general, there is actually a simple formula: hold primaries in the order of closness of the most recent general election, and have only open primaries.
If, on the other hand, you want a nominee that reflects the “base” put first the states in which the margin was greatest, and do not allow any open primaries.
Thus, in 2016, under the win-maximization strategy, the first primary would have been FL, and NC, with OH and VA following. (Chart here). Interestingly, these four states were all won by Clinton, and three of the four by Trump. (Kasich won Ohio, his own state.) Interestingly, one thing this approach might do is encourage politicians who have done well in their own such “close” states to run for President — not a bad thing, I would think.
Using the loyalty strategy, the first ones would have been DC, UT, HA and WY. (feels less right somehow.) (Link here.)
It hard to argue with the proposition that “close” states will tend to test the impact of different candidates upon likely general election success, although no system is perfect. The biggest problem with this system might be if a really large state like CA ended up as a very early primary. But one of the things Trump showed is that so called “retail politics” may be a thing of the past.