# Prob Biden Wins WI in 2020

#### hlsmith

##### Less is more. Stay pure. Stay poor.
I did this quick for @Dason!

Code:
################################################
# Monte Carlo Simulation
B = 1630337
T = 1609640
Tot_counted = B + T
Counted_Per = 0.97
State_Tot = Tot / Counted_Per
Outstanding = State_Tot - Tot
N = 100000
set.seed(2020)
y_rbeta <- rbeta(N, shape1 = 10, shape2 = 10)
hist(y_rbeta)

B_pred = (y_rbeta * Outstanding) + B

Target = State_Tot * 0.50
Target

hist(B_pred, main="Simulated Prob Biden Wins WI",
xlab='Simulated Biden Votes'); abline(v=Target, col='red', lwd=2)

prob = (sum(B_pred > Target)) / N
prob
###############################################

Simulated values do not take into account which counties are outstanding and their past tendencies!
Prob 82% given remaining 3% of votes follow beta dist centered at 50%

Last edited:

#### noetsi

##### Loves R
How can you make a prediction with no known distribution?

#### hlsmith

##### Less is more. Stay pure. Stay poor.
Percentage of votes received is a percentage between 0-100. So the distribution is known. I just place the center at 50% and allow the values to land within 5-95% with the focus in the middle for 100k sim samples.

No magic, just stats. If desired you can do this for every county then weight the counties and sum across states based on the percentage of remaining votes.

#### noetsi

##### Loves R
The real population distribution is not known.

#### hlsmith

##### Less is more. Stay pure. Stay poor.
The population wouldnt have a distribution in the the sense that it would be a set value (e.g., 50.5% votes for X)

Last edited: