He wrote, "I agree with a quote by John Allen Paulos, a professor of mathematics at Temple University, when he wrote that, 'the margin of error in this election is far greater than the margin of victory, no matter who wins.' Further judicial process will not change this self-evident fact and will only result in confusion and disorder."

I was pleased at being mentioned, but also a little upset that my words were used to support a position with which I disagree.

Emails I received didn't help. A number of people wondered how I could support Bush. One asked if I were related to Justice Wells, and some were clearly delusional. With all due respect to my correspondents and to the esteemed judge, I believe that the statistical tie in the Florida election supports a conclusion very different from the one he drew.

For one thing, the tie seems to lend greater weight to the fact that Al Gore received a third of a million more votes nationally than did Bush. People are reluctant to bring this up because the Electoral College makes the popular vote irrelevant, but this year's dead heat can be seen as giving Gore's national plurality the status of a moral tie-breaker.

Another aspect of the moral case for Gore is that since the election, it's become clear that more Floridians intended to vote for him than for Bush. Twenty-thousand overvotes in Duvall County alone were discounted (rightfully) and more than 90% would have gone for Gore. The butterfly ballot in Palm Beach cost Gore thousands of more votes as did the chads that were reluctant to separate.

Most of these problems were in Democratic Counties and help explain why the Voter News Service called Florida so early for Gore; they were rash enough to believe that people actually voted the way they thought they had.

What about the legal votes? I'm not a lawyer, and I never play one anywhere, but I know that intentions don't count, only real votes. So what is real? This is where the chad hits (or doesn't hit) the pan.

The vote is so close and the totals hinge so precariously on the various counties' decisions regarding proper chad behavior that the outcome will be, to a great extent, arbitrary - essentially a coin flip with marginally different probabilities of heads and tails. Since most reported undervote problems occurred in Democratic areas, it would seem that the more liberal the chad standard adopted in them, the bigger the net gain for Gore.

Yet Bush will certainly enjoy an indeterminate net gain in Republican counties. In no case, I fear, will the final Florida outcome be indisputable nor the result statistically significant. There may, in fact, be no fact of the matter.

Nevertheless, short of a coin flip or a concession by Bush or Gore, I think the U.S. Supreme Court should allow the manual recount to go forward despite its problems. Even if such a manual recount is more or less equivalent to a coin flip, counting the votes is our tradition and doing so will bring closure to the process and legitimacy to the winner.

A sports analogy suggested to me by a reader of my New York Times OpEd, "Measuring Bacteria With a Yardstick," is apt. When the referees in a football game measure for a very close first down, they often seem to cavalierly plop one end of the ten yard chain down and then very, very carefully check the other end to see if the tip of the ball is beyond it.

This practice gives one the illusion of definitiveness, but the margin of error in the initial placement disguises the fact that the process, important though it is, is sometimes arbitrary. Such is the situation with regard to the manual recount.

Assuming the game is not called for time, I'm hoping that Gore's nose makes it over the line.