- Jul 19, 2017
- 915
- 1,431
- 44
- Country
- United States
- Faith
- Non-Denom
- Marital Status
- Married
2020 Election Forecast
I am a statistical modeler by profession and I've always enjoyed Fivethirtyeight's election model. They do a great job of explaining the basics and answering questions about how the models made, how polls factor in to the models, what the polls mean, what the uncertainty in the model means etc. They also have regular articles and blog posts talking about some of the more qualitative aspects of the election that are harder to formally incorporate into a model. All around it's a good experience (with the exception of the visualizations they used this year for the model... those are not great).
Right now their model has Biden winning 85% of simulations and Trump winning 15%. Based on national polling averages which have Biden at +10.1% over Trump, it's shocking that the simulations come out that good for Trump since 10.1% is not even remotely close. But, in our country the deck is stacked against democrats. Democrats need to be up 4-6% before we even have a chance of winning the national election because the electoral college over represents empty republican states, but that's where we are.
Think about that for a second. In the last 7 presidential elections (back to 1992) the democratic candidate has won the popular vote 6 times. Democrats lost to Trump in 2016 and Bush in 2000 despite winning the majority of voters. That is not a just system.
Well, no matter. The majority no longer have to suffer the tyranny of the minority. There is much greater democratic support this year than typical. Trump poked the true silent majority and now it is our turn.
I am a statistical modeler by profession and I've always enjoyed Fivethirtyeight's election model. They do a great job of explaining the basics and answering questions about how the models made, how polls factor in to the models, what the polls mean, what the uncertainty in the model means etc. They also have regular articles and blog posts talking about some of the more qualitative aspects of the election that are harder to formally incorporate into a model. All around it's a good experience (with the exception of the visualizations they used this year for the model... those are not great).
Right now their model has Biden winning 85% of simulations and Trump winning 15%. Based on national polling averages which have Biden at +10.1% over Trump, it's shocking that the simulations come out that good for Trump since 10.1% is not even remotely close. But, in our country the deck is stacked against democrats. Democrats need to be up 4-6% before we even have a chance of winning the national election because the electoral college over represents empty republican states, but that's where we are.
Think about that for a second. In the last 7 presidential elections (back to 1992) the democratic candidate has won the popular vote 6 times. Democrats lost to Trump in 2016 and Bush in 2000 despite winning the majority of voters. That is not a just system.
Well, no matter. The majority no longer have to suffer the tyranny of the minority. There is much greater democratic support this year than typical. Trump poked the true silent majority and now it is our turn.