Four out of my five predictions were correct. My biggest (and only) mistake was Ohio, in which I underestimated the number gravitating from the Cubans to the Favorite Son, due to my observations of the vote shares in the Michigan counties bordering Ohio. Ohio is a classic case of voters uniting around a Favorite Son candidate candidate due to the winner-take-all nature of the primary there. It was the only state Kasich won. Kasich in Ohio, like Cruz in Texas, is overwhelmingly popular in his home state, winning 63.8% of the vote in his 2014 re-election campaign. Voter turnout was extremely high in the Ohio Republican primary, over 700 thousand votes higher than in Michigan, while only a few thousand more votes were cast in Ohio than in Michigan on the Democratic side. This suggests the voters that gave Kasich his double-digit win were Kasich Independents who would have felt the Bern had they not been attracted to Kasich, whether they genuinely liked him or whether they just wished to #StopTrump. Either way, this is really bad news for the Trump campaign, despite the fact he only slid less than one percentage point between his vote share in Michigan and his vote share in Ohio. The fact that a bunch of Kasich Independents are willing to be so spiteful as to mount a totally meaningless symbolic blow against the Trump by backing a losing candidate rather than giving a very real and much more significant psychological victory to the Bern against Billary suggests very bad news for the Trump campaign and very good news for Her ahead.

The other races were pretty straightforward. Little Marco won just the tip of Trump’s “something else”, giving credence to his guarantee to us that he has “no problem” there. The polls, for once, were completely correct in the case of Florida. Cruz picked up a lot of counties in North Carolina, but this should not be discouraging to the Trump campaign -no doubt there were a lot of counties in Virginia in which the Cubans’ combined vote share was greater than that of Trump. No doubt Cruz won all those counties due to Little Marco’s complete failure to get more than 15% in any county of North Carolina.

Missouri was a close match. Its plains setting suggested a strong Cruz victory, but its status as a former slave state and its open primary suggested a Trump victory with equal strength. The pre-caucus polling in Iowa also gave me an indication Trump would have won an open primary there, and, therefore, would have also won in more southerly Missouri. Ultimately, I decided in my pre-March 15 predictions to predict a Trump victory there. I had already said the Arkansas results suggested a “strong showing” for Trump in Missouri. And boy, what a strong showing there was. Missouri was an almost perfect example of what a Cruz-Trump one-on-one would look like, with both getting over 40% of the vote, but neither getting the decisive victory they had hoped for. Trump was very strong in the Missouri countryside, but Cruz was very strong in three of Missouri’s four major cities, Trump winning the famously de-industrialized St. Louis. Cruz also put up a strong showing in Missouri’s partly rural Southwest corner, with Trump doing extremely well in the Southeast.

I was surprised at Trump’s strength in Illinois, especially southern Illinois and the outskirts of Chicago. I thought it would be a much closer race than it actually was. Ultimately, Kasich acted as a spoiler candidate there, winning no counties, but successfully splitting the vote there to give Trump an 8.5 point margin of victory.

Yesterday’s results give hope, but also suggest a grave chink in the armor of Trump’s three-state strategy to win the general election. Ohio is a severe danger, due to the threat of the Kasich saboteurs. Florida is relatively safe. Virginia is less so. My suggestions for Trump are to run Spanish-language ads in Florida, placating nervous Hispanics, campaign very strongly in Virginia in person, and to hit Ohio with every bit of weaponry the Trump campaign’s got, encouraging especially strong turnout among women and in Ohio’s Southeast. Trump doesn’t need to win even Rmoney’s percentage of the Hispanic vote. He just needs Black, Cuban and Puerto Rican turnout in Florida to be low.

Advertisements