Trip Kendall
Web and App Developer

Blog Post

2020 Presidential Election Odds

July 18, 2019 Politico, Words
2020 Presidential Election Odds

Even if you stick to the math, don’t pick sides, don’t talk issues, etc. the rabid blue twitter machine will call you names if you publish political data.

It’s like they *REALLY* can’t conceive of someone being interested in collecting as much data on a topic as possible, running it through some machine learning, and publishing the results.

Without picking a team.  But I digress…

2020 presidential election odds

There is an interesting discussion about whether there really is any way to determine who the next president will be this early in the campaign cycle.  Both of the previous links are to fivethirtyeight.com and discuss only polls – the second of the two historically so.

These days there is so much more data available that one does not have to rely exclusively on polls for predictions.  I’ve collected data from a number of sources ( polls, vegas odds, social media followers/mentions/trends, money, and media coverage).  I have been amassing it, in some form or fashion, since the first of the year.

pandas presidential election odds

Over the last few months I have been feeding that data into a machine learning system of my own creation to make predictions.  I completed and can recommend this machine learning bootcamp if you are are interested.

Predicted Odds For 2020 Presidential Election – 7.18.19

Which party takes the White House in 2020?  It’s very close.  The Democrats have a slight edge on Donald Trump generically, but when you drill down it gets more interesting.

1 Donald Trump 48.38
2 Joe Biden 13.5
3 Kamala Harris 10.78
4 Elizabeth Warren 7.21
5 Bernie Sanders 6.45
6 Pete Buttigieg 4.31
7 Beto O’Rourke 3.24
8 Andrew Yang 2.15
9 Tulsi Gabbard 1.5
10 Cory Booker 1.2
11 Amy Klobuchar 0.52
12 Kirsten Gillibrand 0.22

 

Weld, Vice President Pence, and to a far lesser extent Kasich and Cruz,  get small percentages in the system increasing the odds a republican gets the nod to 49.103%.

Obviously the blue vote is split among the large field of candidates.  Essentially after Trump this list is a ranking of what my current model thinks of the Democratic Candidates right now.

D = 50.477%  R = 49.103%  (with .42% an indy (Oprah?) ) Wins the Presidency
1.37 is pretty much inside the margin of error for everyone’s system.

I know the #rabidbluetwittermachine doesn’t want to hear it, but my system / model was built to be non-partisan.  It’s not my opinion, it’s just math, made honestly to the best of my abilities.

It’s analytics based sports handicapping applied to this one political game.  

At this point in time it thinks that the Democrats have a 1.374% better chance at winning in 2020 than the Republican’s do.

That’s closer than anyone on any team would like to admit.

Taggs:
2 Comments
  • Zander 8:56 pm July 18, 2019 Reply

    These are very interesting. Does your system have fluctuation?

    • TripKendall 9:01 pm July 18, 2019 Reply

      Hey Now -my results do fluctuate from posting to posting.

      First there is always new information.

      Second I am always tinkering. My hope is that by the time the campaign gets interesting my models is much better than it is today.

      Since it’s a spare time project I don’t have regular hours I commit to it. Therefore things change allot more some weeks than others 🙂

Write a comment