Nate Silver of FiveThirtyEight, America’s premier source (argue your point!) for elections and polling analysis, has unveiled the site’s forecast in the 2016 presidential election.
A FiveThirtyEight polls-only model rates Clinton’s chance of winning the presidency at 80.6% and Trump’s at 19.3%. A second model that incorporates economic statistics and historical data portrays a slightly tighter race, 73.6%-26.4%. A “now-cast” – who would win the election today? – has the race at 86%-14% for Clinton.
Why credit what Nate Silver thinks? Because in 2008 he predicted 49 of 50 states correctly (missing Obama’s squeaker Indiana win) and all the Senate races correctly, and in 2012 he predicted 50 of 50 states correctly. Silver’s 2012 call was especially sweet because he predicted a blowout 332-206 electoral college victory for Obama when a lot of other loud people said the president was in a close reelection fight with Mitt Romney. They were wrong, Silver was right.
Why discount what Nate Silver thinks? Because in August he gave Trump a 2% chance of winning the Republican nomination, in March FiveThirtyEight’s model gave Clinton a 99% chance of winning the Michigan primary (Bernie Sanders won by a half-point) and in the 2014 midterms he failed to anticipate a Republican wave that created multiple upsets of FiveThirtyEight’s modeling (“the polls did have a strong bias this year,” Silver wrote afterward).
Caveat lector. Here’s what Nate Silver thinks this year:
The model rates the election in each state. For a sense of Clinton’s advantage at this point, consider the 12 tightest state races in FiveThirtyEight’s model. Trump could win them all – and Clinton would still win the presidency, so long as she won the 13th tossup state, Florida.
The model has Pennsylvania at +7 for the Democrat with Clinton having an 85.9% chance of winning. The model has Arizona, which should be a Republican shoe-in, basically tied with a slight edge for Clinton. The model has North Carolina, Iowa, Colorado, Ohio, New Hampshire and Florida all leaning toward Clinton. The model really stinks for Trump.
There’s an entertaining section that awards probabilities for “crazy and not-so-crazy scenarios”. For example the model thinks that Trump has a 90.6% chance of winning at least one state Obama won in 2012; keep your eye on Colorado, Iowa and New Hampshire. The model puts the probability of a “Clinton landslide” at 29.2%.
Here’s how the basic FiveThirtyEight map looks (but click through for their real map, which captures varying strengths, state-to-state, of redness and blueness):
It’s the kind of map that would deliver on David Plouffe’s prediction of 350+ electoral votes for Clinton. (Not rare: Obama beat John McCain in 2008 365-173; in 1996 Bill Clinton beat Bob Dole 379-159; in 1992 Clinton beat George HW Bush 370-168.)