my election prediction

I have a little election prediction spreadsheet. It takes the poll averages for swing states as reported by RealClearPolitics, generates a random number for each with a 4% standard deviation, and runs 1,000 trials in about 10 seconds. Go to Nate Silver or other online sites for a much more professional and sophisticated approach. I do this just for fun and to help me understand how the system works. So without further ado, here is how I think Tuesday night might unfold. The poll closing times are the earliest closing times in a given state according to ballotpedia, so you would expect some numbers to start tricking in at that point. I’m writing Sunday around noon, just in case there is some big development between now and Tuesday.

Based on RealClearPolitics, if both candidates win the states they lead in right now, Hillary would win with 298 electoral votes to 240 for Trump. Nate Silver predicts 290-247, and puts the odds at 65-35. Betfair puts it at 323-215 and the odds at 80-20. My spreadsheet comes up with an electoral college average of 295-243 and odds of 85-15.

Here is one way the evening could unfold to get in the ballpark:

First, I assume Clinton and Trump have both won all the states considered relatively safe by RealClearPolitics. This means Hillary starts off with 218 and Trump with 165. Seems unfair, doesn’t it? But these are the demographics, and why the breathless media coverage of swing states is a bit misleading. If Trump is leading half the swing states on a given day, that doesn’t mean the race is anywhere near tied.

7:00 p.m. EST

  • Results start to trickle in from Florida, Georgia, Virginia and New Hampshire.
  • The night starts off with a bang for Clinton with wins in Florida and Virginia.
  • Trump gets Georgia and New Hampshire.
  • Clinton leads 247-185.

7:30 p.m. EST

  • North Carolina and Ohio
  • I’ll throw both to Trump.
  • 247-218. Getting slightly interesting.

8:00 p.m. EST

  • Pennsylvania and Michigan
  • I don’t think Trump has a realistic shot at either. They go to Clinton.
  • 283-218. It’s over!

9:00 p.m. EST

  • They split Arizona (Trum) and Colorado (Clinton).
  • 292-229

10:00 p.m. EST

  • Iowa and Nevada
  • I’ll throw both to Trump. I’ll also throw him New Mexico to look like slightly less of a loser.
  • 292-246

This is what I expect to happen. Of course, the votes get counted slowly, and we can pretend there is some suspense as they are counting votes in states that are not expected to be close. Still, I think we might all be in bed at 10 p.m. on the east coast knowing who the next President. And this is what I want to happen. Although I would enjoy some suspense on some level, rationally I know it is better not to live in interesting times.

For Trump to win, a lot of unlikely things have to fall into place, but here is a plausible scenario: Trump starts the night with a huge bang, winning Florida, Georgia, North Carolina, New Hampshire, and Ohio. Clinton gets Virginia. Trump would be up 252-218. By the way, I gave Clinton New Mexico to start off this time. Clinton wins Pennsylvania and Michigan, going up 254-252. They split Arizona and Colorado (263-263!). Trump gets Iowa and goes up 269-263. It comes down to Nevada. Right now it looks like Nevada is reasonably solid for Clinton, so it comes down to a 269-269 tie. The House of Representatives casts the deciding vote, picking Trump for President. The National Guard is deployed in some states to ensure order.

So Florida is a big deal, obviously. We knew that.

Leave a Reply

Your email address will not be published. Required fields are marked *