From yesterday's post:
Probabilities rely upon the standard deviation (σ) of a given sample. The σ will be higher when polls are more spread out in their margins, and lower when they all are close together. This is what we have going on in NC. The σ in my model there is actually under 1%. The current D +3.5% margin in NC is over 3σ (actually it is roughly 3.9σ) away from being tied. This is why the NC chances for the GOP are <1% (0.22%) today while they were 13% yesterday. This is the major change which leads to the GOP chances dropping overall.
The confusion seems to be: why is the probability of the control of the Senate not calculated based upon my current balance of power scoreboard? This is actually a more simple answer than I appeared to make it seem yesterday. In short, the probability and the balance of power are calculated from the same set of numbers and not from each other. The GOP 52 - Dem 48 spread is calculated based solely upon a weighted mean of the data at hand. The σ for each race is then calculated. The mean (x̄), and the σ of each race then are plugged into a formula the same for all races that are not Alaska (historically polls here are unreliable more so than other states) and Kansas (this is a special situation). Though, for the purposes we are concerned with here: lets just consider them the same. This gives us the P(A). Event "A" is a GOP victory.
The overall control probability is then arrived at through these P(A)values. These values are used to form a 50,000 trial Monte Carlo simulation. I take the number of P(A) occurences and divide that by 50,000 to arrive at P(B) with "B" being a GOP Senate majority. This is the number I use for my "chance of GOP control" which today is 50%. This histogram (bellow) shows the frequency each Senate result occurred in my series of simulations.
*As a reminder: The numbers at the top of the Histogram chart are the average seats won (in those Monte Carlo simulations) by each party and NOT the probability for control for each party.
New Mexico Governor
New York Governor