Let's start with what the electoral map looked like in 2012 (all of these maps come courtesy of 270towin.com):
That map yielded Mitt Romney just 206 electoral votes -- 64 short of what he needed to win to be elected president.
It's easy to blame Romney for under-performance in 2012 -- as Trump has done. But that 2012 map simply codified much of what we already knew about the state of the electoral college: It heavily favors Democrats at the moment.
As Dan Balz points out in his terrific electoral college analysis, Democrats have won 18 states and the District of Columbia in each of the past six elections. (That's back to 1992.) That's 242 electoral votes total, meaning that Clinton only needs to find another 28 electoral votes to win the presidency in November. Win Florida and its 29 electoral votes plus those 18 states and D.C. and Clinton is president, for example. It's just that easy.
The reality of the map for Republicans in 2016 is that anything close to the status quo virtually guarantees a loss. If the map stays stable, the GOP nominee will find himself in a very similar situation to where Romney was toward the end of the 2012 race: needing to run the table of competitive states to barely crest 270 electoral votes.
Cruz may be an outsider to the GOP establishment, but his candidacy is not radical or risky in the way that Trump's quite clearly is. Cruz is, on virtually every issue, a traditional social, economic and foreign policy conservative. Which makes him a safer pick for Republicans but not necessarily one who can alter the electoral map in any meaningful way. Look at the map above and ask yourself which state Cruz can win that Romney didn't? Florida? Maybe. But he still needs 31 more electoral votes to get to 270. Virginia? I doubt it. But Cruz would still be 18 votes short.
Here's a semi-plausible map that would get Cruz over 270:
That map would give Cruz 275 electoral votes to 263 for Clinton. But to get there you need to give Cruz Florida, Ohio, Colorado and Virginia. President Obama won all of those states twice.
Now, consider Trump and the map. There is, without question, a scenario where Trump loses 40 states and gets completely swamped. But the possibility also exists that Trump is able to re-jigger the map that has been in place since 2000 -- a map where Republicans face an uphill fight before a single vote is cast. (Trump is the low floor, high ceiling candidate of the 2016 electoral college.)
Among the 18 states that have been in Democratic hands since the 1992 election are Michigan, Pennsylvania, Wisconsin and Minnesota. Along with Ohio and Iowa, those heartland states are likely to be the most intensely contested battlegrounds in the country if a Trump-Clinton race materializes. ... All those states have higher concentrations of white voters, including larger percentages of older, white working-class voters, than many of the states in faster-growing areas that Obama looked to in his two campaigns.
A map in which those Midwestern states Balz highlights are legitimately in play is a different -- and potentially more advantageous -- one than Republicans have been playing on for the last decade-plus. Here's what a best-case Trump scenario map could look like:
That map gets Trump to 285 electoral votes and assumes he loses all of the southwestern states -- with the exception of Arizona -- due to his rhetoric and incredibly poor numbers in the Hispanic community.
Which map -- the Cruz one or the Trump one -- seems more plausible to you? For me, the Trump map (wins in the upper Midwest + Florida) seems more likely than Cruz's patchwork of wins in all four regions of the country. Trump's message of economic populism should -- or at least could -- sell better than Cruz's rock-ribbed conservatism outside of the solidly Republican South and plains states.
Establishment Republicans seem to have convinced themselves that nominating Cruz will avert a certain electoral disaster. It might in that Cruz would run a more traditional campaign on a more traditional electoral playing field that would likely endanger fewer downballot Republicans. But the idea that Cruz is more likely to get elected president than Trump is a far iffier proposition.