Add ability to conduct election night conditional probability updating

[markjrieke]

Basically just need to write the draws somewhere.

On election night, when any competitive states are called, filter to the subset of draws that match reality. Only needs to be run interactively during election night, so should be a separate script. Likely will want to amp up the number of draws for that run specifically. Will also want to communicate uncertainty around the probability.

This can be a 1.2 level effort that gets done after #4

[zecellomaster]

I have a suggestion, if I may. I remember the last election you had issues where you ran out of simulations as races were called. I've been toying with the live election model idea that Nate Silver/FiveThirtyEight came up with in 2014. The also used a similar methodology in 2016. In both cases, they utilized something more dynamic that conditional probabilities, and I think it will be a immune to running out of simulations.

[markjrieke]

I think the live coverage model Nate describes in that link is useful for his particular type of weighting model, but would be difficult to implement here.

I've toyed with the idea of generating a pure election-day only forecast that conditions the state covariance matrix based on observed outcomes, but that comes with a couple problems of its own:

• States are called before the final vote tally is known. To be fair, the degree to which this actually matters is probably minor.
• I consider NE/ME as aggregations of their congressional districts, rather than being modeled directly. I'd need to grok a way to potentially condition on a known state voteshare while still allowing for an unknown district voteshare.

I'm only planning on sharing on twitter, so the simplest solution is to implement the first portion of what 538 did in 2020: crank the number of simulations up & filter down based on called races. In my case, I'll probably only share updates when competitive (p < 0.85) races are called.