Donald Trump
In September 2015, writer and statistician Nate Silver urged people to "calm down" about the possibility of Donald Trump winning the Republican presidential nomination.
Two months later, he wrote that the media should "stop freaking out about Donald Trump's polls" and that Trump's odds were "higher than 0 but (considerably) less than 20 percent."
Six months after that, after Ted Cruz had dropped out of the race but before John Kasich had done so, Silver wrote: "Donald Trump is going to win the Republican nomination."
"Other than being early skeptics of Jeb Bush, we basically got the Republican race wrong," Silver wrote.
It's easy to cringe at how, in August, for instance, Silver outlined the "six stages of doom" that he foresaw for Trump in the coming months and how, in December, he updated the post to note that "the most difficult hurdles between Donald Trump and the Republican presidential nomination are still to come."
So how did the site that prides itself on a numbers-based approach end up just another late-stage Trump bear with its tail between its legs? Silver has some ideas.
For one, some of Silver's earliest Trump doomsday analysis was, by its own admission, in line with the theories of "The Party Decides," a theory that posits candidates must be electable and believers in the party's positions. By January, Silver was rereading the book: Either the book's hypothesis that functioning parties nominate strategic candidates is wrong, he said then, or the Republican Party is not a functioning one.
But Silver isn't giving up on "The Party Decides" quite yet. So when, on Wednesday, Silver admitted that "in Trump, the Republican Party may have a candidate who fails on both counts," he unsurprisingly also argued for the "failings of the Republican Party as an institution."
Silver wrote:
To me, the most surprising part of Trump's nomination which is to say, the part I think I got wrongest is that Trump won the nomination despite having all types of deviations from conservative orthodoxy. He seemed wobbly on all parts of Reagan's three-legged stool: economic policy (he largely opposes free trade and once advocated for a wealth tax and single-payer health care), social policy (consider his constant flip-flopping over abortion), and foreign policy (he openly mocked the Bush administration's handling of the Iraq War, which is still fairly popular among Republicans).
But Republican institutional failure alone is only one of three major facets of Silver's analysis, which simultaneously addressed Trump's success and Silver's (predictive) failure.
The other two components are the incredible volume of media coverage of Trump and the tribal nature of Republican primary voters. On the latter, Silver said that Trump's appeal to "cultural grievance" worked.
"It's a point in favor of those who see politics as being governed by cultural identity," he wrote, "as opposed to carefully calibrating one's position on a left-right spectrum."
Statistical analysis is always still analysis and, as such, requires some assumptions. And though he didn't address it specifically in his post on Wednesday, Silver said in November that a statistical approach to presidential-campaign prediction is a hard problem because there's only so much data.
Perhaps Silver's most prescient prediction came in November: "Unprecedented events can occur with some regularity."