After last week’s surprise win by Sanders, many pundits are very cautious about making predictions this week. I share the caution.  You can read good takes by Harry Enten at Fivethrityeight.com regarding expectations for both the Democrat and Republican races.

For the Democrats, Paul Krugman linked to a model from Emory political scientist Alan Abramowitz which looks at racial composition and a dummy variable for whether the state  in question is part of the South.  This simple model explains over 90 percent of the variance in Clinton’s vote share.  It predicts FL and NC to be Clinton’s best states; with MO as perhaps her weakest.  An interesting question is whether Dem voters cross over to vote for Kasich in Ohio’s open primary.  Finally, the proportional way that Democrats award delegates makes it hard for Bernie to catch Clinton unless he has some big wins along the way.

Read Harry Enten’s take at Fivethirtyeight for a full rundown on the Republicans.  I think the divided field continues to give Trump the ability to win big delegate counts with plurality victories.  But will voters be coalescing around single anti-Trump candidates in specific states or remain divided?  With no special knowledge or secret sauce in my possession, my hunch is that Trump will win FL and Rubio will drop out of the race. Kasich will probably win Ohio and stay in, but I’m not sure he has much potential past his own home state.   If Cruz can win a couple of states, that might give him some momentum to become the single conservative alternative to Trump.  But it may be too late.  Who knows?