Wargaming the Electoral College
You've probably seen the above map already, or ones very much like it, showing Hillary Clinton's prohibitive lead in the Electoral College. In my previous Wargaming column from four weeks ago, Clinton's worst-case scenario gave her a squeaker 273-265. And that, I must reiterate, was her worst-case outcome as of just four weeks ago.
But then Donald Trump went to Mexico in what might prove to have been a game-changer. It is already at the very least a direction-changer -- if the massive new WaPo/SurveyMonkey poll is to be believed.
Let's look at the maps based on that poll of nearly 75,000 voters.
First, a caveat. SurveyMonkey's poll is only of Registered Voters, and this close to Election Day you really want to start whittling your results down to likely voters. Here's what SurveyMonkey said about it:
This Washington Post-SurveyMonkey 50-state poll was conducted online Aug. 9 - Sept. 1 among 74,886 registered voters. The sample was drawn among the respondents who completed user-generated polls using SurveyMonkey’s platform during this period, and results are weighted to match demographic characteristics of registered voters in each state. No margins of sampling error are calculated, as this statistic is only applicable to randomly sampled surveys.
That aside, the sheer size of the sample plus SM's weighting method ought to be more than enough to lend credence to their numbers.
When I began the 2016 Wargaming series in July, I showed you Trump's "northern route" to 270, taking the fight through PA, OH, MI, WI, and MN. By August that route had appeared to have closed. But if we look at SurveyMonkey's results from the four-way race (GOP, Dem, Libertarian, Green) we see that that route may have re-opened.
In the map above, I've colored states where either Trump or Clinton has at least a four-point lead over the other. The results look much like they did two months ago, with Clinton comfortably above 200, and Trump uncomfortably well below 150.
But let's look a little deeper.
To get a feel for the leaners, I've taken the states where Clinton or Trump has a three-point lead, and colored them in a lighter shade of blue or red -- and shockingly, Texas and Colorado are still unshaded.
Gary Johnson seems to be sucking the life out of Clinton in Colorado, and out of Trump in Texas. He's also hurting Trump in Arizona and North Carolina, and putting the squeeze on Clinton in Wisconsin and Michigan.
Unsurprisingly, Nevada and PA are going blue in this map, and Florida remains as delightfully crazy and un-callable as ever.
But what about those elusive Likely Voters -- the voters who actually, you know, vote? It's a well-established political fact that pool of Likely Voters lean more Republican than the broader field of Registered Voters, typically by two to four points.
So what if I were to put my thumb of SurveyMonkey's scale, and spot Trump three points across the board?
Clinton has the same 218 Electoral College votes we saw on the initial SurveyMonkey map, but Trump has shot up to a slim lead of 239, if we include the new leaners. And that "northern route" I first mentioned two months ago is almost wide open, with Clinton retaining her lock only on Minnesota. Ohio, which no Republican has ever been elected without winning, now leans Trump. And MI, PA, and WI are up for grabs.
The only holdout out west is NV, and Florida -- as ever! -- remains as delightfully crazy and un-callable as ever.
Earlier today the Washington Post reported that "despite long odds," Trump was making his biggest media buy to date in Virginia. And indeed those odds are long, as VA hasn't budged from blue in weeks, and not even the surpassingly strong third-party runs of Johnson and Jill Stein are making a dent in Clinton's lead there.
Trump instead ought to look west to the Rocky Mountains, and north, north, north to the Rust Belt states which Democrats have taken for granted -- and run into the ground -- for decades.
CORRECTION: I'd misread the dates of SurveyMonkey's weeklong poll, which concluded the shortly after Trump returned from Mexico. So rather than the results showing an improvement for Trump, they instead reflect the drip, drip, drip of Clinton's FBI revelations and other (ahem) trust issues.
That ought to mean Trump still has plenty of headroom on the growth side.
Thanks to several readers for pointing out my mistake -- corrections are always welcome!
(All maps created with 270toWin's iPad app, which is highly recommended. Perfect for fiddling with on the sofa while you watch the news. If you prefer to do your wargaming in a browser tab, the website is great, too.)