I read the article and checked his math. Natapof is mostly right, about the math. (In case you wonder about my credentials, I have a Ph.D. in physics, also from MIT). But he's wrong about his conclusion!
His math predicts that the electoral-college system increases the probability that a single vote will be the deciding factor, relative to a popular-vote system. He's right (though, technically, he underestimates the effect by a factor of about 50).
As an analogy, he uses a world-series. If we decide the winner by "popularity", we add up all the runs from the seven games. If we go by winners of most games, that's like the electoral college system. Clearly, the latter method (the one in use) increases the probability that a single run will be the deciding factor for the whole series.
But what Natapoff neglects is that this current method (most games instead of most runs) method also maximizes the probability that the decision
will be made by a single error!
Likewise, the electoral-college system maximizes the probability that a single relatively small "voting irregularity" will be the deciding factor for the whole country.
We need only look at the current situation to see the effects. Some confusing ballots in one county of one state could be the deciding factor for who the next president will be. And even if the ballots are fair, the present system allows accusations of a small irregularity to cloud the whole issue.
What Natapoff has proven mathematically is why the electoral college must go.
Slashdot Mirror
I read the article and checked his math. Natapof is mostly right, about the math. (In case you wonder about my credentials, I have a Ph.D. in physics, also from MIT). But he's wrong about his conclusion!
His math predicts that the electoral-college system increases the probability that a single vote will be the deciding factor, relative to a popular-vote system. He's right (though, technically, he underestimates the effect by a factor of about 50).
As an analogy, he uses a world-series. If we decide the winner by "popularity", we add up all the runs from the seven games. If we go by winners of most games, that's like the electoral college system. Clearly, the latter method (the one in use) increases the probability that a single run will be the deciding factor for the whole series.
But what Natapoff neglects is that this current method (most games instead of most runs) method also maximizes the probability that the decision will be made by a single error!
Likewise, the electoral-college system maximizes the probability that a single relatively small "voting irregularity" will be the deciding factor for the whole country.
We need only look at the current situation to see the effects. Some confusing ballots in one county of one state could be the deciding factor for who the next president will be. And even if the ballots are fair, the present system allows accusations of a small irregularity to cloud the whole issue.
What Natapoff has proven mathematically is why the electoral college must go.