Show swing state swaying power

[qwertie]

First off, the site is about the presidential election, isn't it? But it doesn't say so on the site.

This electoral map shows the theoretical difference in vote values in different states. It doesn't reflect the practical difference between voting power in different states. This map, for instance, claims votes in California have 1.17x the value than votes in Florida, but that's not true in practice; in practice, individuals in California have almost no voting power because the result is a foregone conclusion.

You could add a series of additional numbers - one for each recent presidential election - showing how much votes are worth in terms of their power to sway the election results. For example, let's say the margin of victory in 2012 was 1% of registered voters in Florida but 10% of registered voters in California. Then a vote for one of the top two parties (but not for a third party, of course) was roughly 10 times more powerful in Florida. You'd combine that by the original datum (1.17) and say that in 2012, a vote in Florida was worth 8.54 times as much as in California. If my math is right.

To implement this feature you'd also have to take into account the two states that don't use winner-take-all.

[qwertie]

Hmm, I think my math is wrong.

For an individual voter, they are interested in the probability that their vote could change the election result, or at least, that their vote could have any effect whatsoever on the electoral college result. Obviously:

• this value is microscopic, so it makes more sense to think about a bloc of, say, 10,000 voters instead of one
• this value varies greatly from state to state, making it relevant to votes.equalcitizens.us.
• in a typical election, this value is zero for third parties
• for the D and R nominee, it is unknown and unknowable prior to an election, but can be estimated. A reasonable proxy is the margin of victory of the winning candidate after each election.

However, the relationship between the margin of victory and the value of a vote is non-linear:

• in case of a small margin, since the margin is unknowable in advance, a margin of victory of 10 votes almost the same as a margin of 1000 votes in terms of how much one should be motivated to vote, so it would be a bit misleading to say votes are 100 times more powerful in the former case.
• at the large end of the scale, the probability of your vote making a difference is very nearly zero. So if a candidate is up 30 points in California and 3 points in Florida, the Floridian's voter matters more than 10 times as much.

I suspect it would be best to show the margin of victory separately from the original number.

[qwertie]

Note that the probability of your vote mattering is independent of the size of your state and the number of electoral college votes your state gets: the margin of victory is all that matters.

However, perhaps it's enlightening to define the "power" of voters, which combines the margin of victory with the payoff for winning:

Voting power = Electoral College votes / margin of victory (as a number of votes)

In this formulation, "voting power" measures the maximum possible impact of getting more people off the couch on election day to vote for your candidate. Voting power is correlated with the number currently shown on the site (the relative per-capita combined "value" of all voters in a state) but quite different. It's normally a tiny number, but it doesn't matter if we focus on ratios. Probably presidential campaigns have developed a related (and more complicated) formula to help them decide how to allocate ad dollars and GOTV efforts, and even how to decide the finer points of their platforms...

Random fact: reformers who support the National Popular Vote initiative should be aware that it has made a big mistake by failing to define the concept of "popular vote".