I'm having fun today! Using state odds from PredictWise I set up and ran an election simulation...

[Godhumor]

Just a really basic one based off state by state win probabilities from the PredictWise site. Anyway, I set it up to run 10,000 times and then ran that multiple times. Off those runs, here is what I found:

Hillary wins approximately 94% of the time (Seriously)
Average EV total is 309 to 312
Most common result is between 297 and 326 (Depending on states she wins in each iteration, obviously)

Raw data pulled from http://predictwise.com/politics/2016-president-winner

[Farmgirl1961]

1. Awesome!

But can anyone tell me what gives over at 538? Hillary's numbers keep dipping.

[leftynyc]

2. Look again

They're up since yesterday.

[Helen Borg]

5. 56.8% / 43.2% nt

[leftynyc]

6. Since that shows her UP from

yesterday, which is all I said, what's your point?

[Cary]

8. Nothing gives, don't worry

The polls have been nothing if not steadily in favor of a Clinton victory. People naturally are dropping off, temporarily, after the primaries and before the election kicks into full gear.

Turn off the talking heads.

[Imperialism Inc.]

3. The problem with that though is that states are correlated.

If Clinton under performs in one state she is likely to do so in other states. For example, losing Wisconsin but winning Iowa, which will rarely happen in practice, should have happened in about 10% of your simulations.

[Godhumor]

4. Which is why I said it was a basic sim

Whipped up during lunch. Certainly didn't bother to sit down and figure out which states move together and by what percentage. Even still, the states she should win every time, regardless of her basically at or above 270.

[Sunny05]

7. Wow! Thank you!

And yeah, that does sound like fun! 😊

[Lucky Luciano]

9. How did you set up the state by state pair wise correlations? nt