What is the actual cost to a political party of losing a reliable voter?

Rough question: How much does it cost a political party if a reliable vote for them bolts to the opposition?

Let’s use the election of 2008 between Obama and McCain as an model.

1.) If the reliable voter numbers for each candidate are about even, with a critical independent group up for grabs…

2.) If that independent group will probably break about even, with the critical winning margin being decided in the difference…

3.) And if I have been a reliable vote for McCain, who bolts to Obama…

Question:

Does the Republican Party then have to make up 2+ votes to undo the damage of my one vote shift? Does it also have to add some additional percentage of a vote to get to a win for McCain after having “fixed” the damage of losing my previously reliable vote?

To get my figure of 2+ votes, I’m assuming [1 vote] to neutralize my Obama vote, plus [1 vote] to get the Republican Party back to status quo ante before I bolted.

How do numbers like this actually work? If I’m wrong, by how much am I miscalculating? How much damage does a reliable vote shifting do comparted to a middle-ground independent vote lost?

The entire election if the vote tally is close enough.

It leads to a two-vote change. Now, the party might make up for that change by convincing two people who otherwise wouldn’t have voted at all, to instead vote for their guy, or they might make up for it by convincing one voter who would have voted for the other guy to switch back. And they don’t necessarily need to convince the die-hards: There’s just as much benefit in convincing someone who just barely slightly prefers the other candidate to instead just barely slightly prefer your candidate, even though that might take only a very small nudge.

Well, it obviously depends on your behaviour and the behaviour of the other vot(s) they gain. If you were a sure-fire McCain voter (in the sense that you were determined to actually vote for McCain) and decide to become a sure-fire Obama voter, then there are two possibilities for the Republicans to make up for that: Either (a) convert one sure-fire Obama voter into a sure-fire McCain voter; or (b) convert two non-voters into McCain voters.

Let’s take a numerical example:

Status quo: 1,000 sure-fire Obama voters, 1,000 sure-fire McCain voters, 1,000 non-voters.

You switch, turning the constellation to 1,001 Obama voters, 999 McCain voters and 1,000 non-voters.

Easy fix (a) will obviously restore the status quo numerically. Easy fix (b) will turn the constellation into 1,001 Obama voters, 1,001 McCain voters and 998 non-voters.

Arguably, this constellation is still worse for the Republicans than the status quo - they restored numerical balance to the Democrats, but the potential for gaining non-voters has become smaller. And the two voters who switched from non-voting to voting for McCain would be expected to have leaned towards McCain anyway, whereas the Democrats have not “used up” non-voters with a tendency for Obama.

For presidential elections, it depends what state the voter lives in. If your state leans heavily towards one candidate or party, one vote or even a thousand wouldn’t make a difference.

That argument brings us to Zeno’s old paradox with the heap of sand.