Sunday, September 18, 2016

Clinton vs. Trump.
18 September 2016,
51 days from the election

The long and short of it was that Hillary Clinton fell down, both literally and figuratively. Trump had his best week ever and is now in striking distance. The big question for the last seven weeks is if this is a trend or if Hillary can get back on track. Here are the numbers.

Solid Trump (more than 95% Confidence of Victory [CoV]): ND WY AL ID TN UT MT SD OK KY WV MS LA MO AR AK SC KS IN TX  
Total: 159

Leaning Trump (between 50% and 95% CoV for Trump): ME-2 IA NE GA OH AZ FL


Toss-up: NC
Leaning Clinton (between 50% and 95% for Clinton): NV CO MI WI RI NJ NH


Solid Clinton (more than 95% CoV for Clinton): ME PA VA MN DE NM OR WA IL ME-1 CT DC MD MA NY VT HI CA
Total: 215

The current count 



Almost all the movement this week was in Trump's favor, gaining the median lead in both Ohio and Florida and bringing North Carolina into a 50-50 tie. Unlike 2004, where winning Ohio and Florida was just enough for George W. Bush to beat Kerry, the demographics have changed and Trump will need to get other wins in state that currently do not favor him. 

Here are the odds of victory using the 15 battleground states according to my algorithm.

 Trump's non-battleground count of electors: 122

The battleground states ranked by pqn
(Clinton%, Trump%, electors)

1. FL: 39%, 61%, 29
2. NC: 50%, 50%, 15
3. OH: 23%, 77%, 18
4. AZ: 36%, 64%, 11

5. MI: 85%, 15%, 16
6. CO: 71%, 29%, 9
7. GA: 10%, 90%, 16
8. NV: 66%, 34%, 6
9. TX: 3%, 97%, 38 
10. WI: 87%, 13%, 10
11. PA: 96%, 4%, 20
12. NJ: 94%, 6%, 14
13. VA: 96%, 4%, 13
14. IA: 9%, 91%, 6
15. NE: 10%, 90%, 5

Current probability of victory if the election were held today:

Clinton 89.7%
Trump 10.3%

Instead of using the algorithm and looking at paths to 269 electoral votes, my choice of 11 states for battlegrounds going from most likely Trump to least likely would OH, AZ, FL, NC, NV, CO, MI, WI, NJ, PA and MN. This only brings his numbers up to 12.1%, not much of a difference and this is as generous a list as I can create in good conscience.

How can his odds be so bad when he only needs to get North Carolina in his leaning category and pull away one of several states from Clinton? The reason for this seeming contradiction is that he could lose his new won advantages and he has so many states in the leaning column he still has to defend.

Why would I think his gains of this week aren't certain? I looked at the numbers I compiled in 2012. After a disastrously low energy first debate, Obama's numbers against Romney sank in a similar way. After about a week and a half, they returned to their previous levels. A bad debate difference is not the same as a health scare, but if they are both temporary, I expect the effects will be temporary as well.

Earlier in the campaign, I wrote the race looks like an anaconda swallowing a pig: long, slow, ugly and inevitable. Right now it's doesn't look as inevitable, but I will wait a week to see if his current trend continues. With Hillary back on the trail, Trump defending his retreat from birtherism and his return to Twitter attacks, I have am not willing to call this race close yet.

Back next Saturday with information on the Senate.


  1. Not to be pedantic, but your totals for Trump in the text are 285, and in the chart it says 265. Typoo or math error, your choice...

    I was alarmed in the first reading...

    1. All I see is 245 in the chart and 159+86 in the text. What paragraph are you looking at?

    2. OK, OK, gotta clean my screen more often...

  2. Professor I hope your probability of Clinton winning remains favorable. This thing is making me nervous. Hopefully the bad few weeks will have lit a fire under the Clinton campaign's collective seat. Nate Silver's 538 still has this as basically a coin flip. Do you think we may see a repeat of 2004 where we'll be up 'till 2 a.m. waiting for one state to decide this?

    1. Hi, Marc, thanks for stopping by. My numbers don't agree with Nate Silver's, which is one of the reasons I'm publishing them. Things can definitely change and Clinton's health scare proved that, but I'm guessing it's more like 2012 than 2004, for reasons stated in the post.

      In 2004, Bush needed both Ohio and Florida and got them. They were both close to coin flips. Right now, Trump would have to win much more difficult races to get to 269, as the House will likely stay Republican and he will win any tie.

      The next few weeks will tell us if this will be relatively easy or a nail-biter and the first debate should be a big decider one way or the other.