Cynthia writes:

> If I recall correctly, calling the states was based on media "exit
>  polls", not on actual election results. If someone asked me something
>  like that as I exited, I'd give them the wrong answer ON PURPOSE and
>  giggle to myself as I did. Who I vote for is my business, not theirs.
>  The actual results ought to be reported, not exit poll results.

Actually, "calling" the state is based on more than that. It's also based on
continuous polling prior to the election, giving especially high regard to
its recent political history. And I suspect that the exit polling
organizations expect a certain degree of disingenuousness in their responses
and fully account for that.

"Calling" a state is generally not all that difficult nowadays. If the state
under discussion were, say, Kansas, and Kansas hasn't voted for a Democrat
since Kennedy, and most contests don't even have a Democratic candidate, then
it's relatively simple to "call" Kansas before the polls close. Heck, in
these circumstances, it's relatively easy to "call" Kansas several months
before the election :-), and in fact, that's why neither party bothered to
visit the state or spend any money there this election cycle. The state was a
foregone conclusion long before NBC or CBS announced the winner.

The same, but converse, statement was true of California. Republicans have
recently had very little chance there. Dole was mightily criticized by the
California Republican party eight years ago for never bothering to show up,
but he rightly wrote it off as a lost cause. Because of the criticism that
Dole received, Bush actually gave California more than lip service this time
around, but he never really had a chance, and he and the Republican Party
knew that to be true before, after, and during the campaign.

Time and money are the great problems of every campaign, and both sides work
very hard to optimize their expenditures of both. Indeed, they've got this
optimization process down to such a science now -- and because the
first-order, second-order and third-order attributes of the process are now
so well understood -- and because there are no great pressing issues at this
moment in time -- the two parties have managed to work themselves into this
extraordinarily improbable tie.

All of this reminds me very much of the history of baseball. The same sort of
optimization process has occurred over the last 120 years in baseball. The
best players now may or may not be any better than the best players then, but
the opportunities for a really great player to distinguish himself have
dropped off radically as time has gone on, as have the statistical variances
experienced from game-to-game have decreased year-by-year.

Baseball is no longer merely a game of first-order processes (the explicit
rules written in a rule book) but is now a game of implicit second- and
third-order processes (how to align the field against the statistics of a
particular player, when two men are on base and a power hitter is two away
from his turn at bat). It's this kind of accurate knowledge, gained over a
century's worth of trial-and-error, that exists on both sides that prevents
"blowouts" from occurring regularly now. A tie (or near-tie) has become the
statistically expected result, with the games becoming longer and longer on
average each year.

This is the fundamental lesson from all coevolutionary events, natural or
artificial -- and I would expect it to become more and more common in
American elections too, simply due to our increased capacity to do "data
mining" and adjust our behaviors accordingly to the most opportune
circumstances.

Wirt Atmar