Skip to content

Now that 7pm has come, what do we know?

(followup to this post)

On TV they said that Trump won Kentucky and Indiana (no surprise), Clinton won Vermont (really no surprise), but South Carolina, Georgia, and Virginia were too close to call.

I’ll run Pierre-Antoine Kremp’s program conditioning on this information, coding states that are “too close to call” as being somewhere between 45% and 55% of the two-party vote for each candidate:

> update_prob(trump_states = c("KY","IN"), clinton_states = c("VT"), clinton_scores_list=list("SC"=c(45,55), "GA"=c(45,55), "VA"=c(45,55)))
Pr(Clinton wins the electoral college) = 95%
[nsim = 65433; se = 0.1%]

Just a rough guess, still; obv this all depends on the polls-based model which was giving Clinton a 90% chance of winning before any votes were counted.

One Comment

  1. Ara says:

    Cool. Following along at home with the same R model!

Leave a Reply