The great Florida recount is under way. The results are trickling in–and, while they seem to boost the contention of Gore diehards (we really got more votes), this news is unlikely to have much of an impact on politics as-is, especially since the information is coming from a variety of diffuse sources and can be debated by those who don’t desire a clear and definitive post-election accounting.

Actually, it is not a recount that is happening; call it an inspection. There are no official reviews being conducted. But in counties across the state, different news organizations inspecting various sets of ballots. These outfits are then free to classify and evaluate the ballots in manner they wish. They can report whatever results they see fit to publicize. They can state the number of ballots that had missing chads, or dimpled chads. They can add up ballots that would have or could have been counted for either Bush or Gore–or they can let the public do the math. The Miami Herald (with USA Today), the Palm Beach Post and the Orlando Sentinel have been busy surveying ballots. And a big-media consortium–the New York Times, the Washington Post, the Wall Street Journal, AP and others–is set to start its own statewide assessment. In addition, news organizations and academics have conducted computer analyses of ballot data to determine if votes for Gore or Bush were lost due to technology problems.

It will be hard to derive a single bottom-line from all this–to state that a recount would definitely have granted Gore (or Bush) a victory of X number of votes. There will be arguments over standards. I spent three days in Miami studying one-third of the 10,600 undervote ballots in that county. My review–using tight standards–projected a Gore gain of only fifteen or so votes. (See “In the Field of Chads,” January 29.) The Palm Beach Post‘s inspection–which utilized looser standards–uncovered more votes for both Bush and Gore, with Bush picking up six votes. And there will be disagreement over what votes should be considered when tallying the true outcome. For instance, as I found during that review–and as Anthony Salvanto, a faculty fellow at University of California at Irvine discovered by scrutinizing ballot data–in Miami-Dade, Gore outscored Bush by 300 among ballots where voters, for inexplicable reasons, had punched out the chad one spot below the chad for either Bush or Gore. Should such votes be counted?

Even as questions like that go unanswered, new numbers have started emerging. The Palm Beach Post examined one group of 4,513 undervotes in its home county. These are ballots that the county’s canvassing board looked at in November and decided that no vote had been cast, with Democratic and Republican observers disagreeing. The newspaper concluded that Gore beat Bush 2500 to 1818 among these voters. The 682 difference here surpasses Bush’s 537-vote margin of victory. (This would be in addition to the 174-vote gain Gore acquired during the county’s hand recount–a gain which was not accepted by Secretary of State Katherine Harris, for it was submitted several hours after the deadline.)

Sounds like a big boost for Gore. But…not so fast. The ballots reviewed were only those undervotes disputed by the observers during the post-election hand recount. Since the Democrats were arguing for the counting of dimpled ballots and the Republicans were not, Republicans did not object as often when the canvassing board ruled a ballot a no-vote. Consequently, this group of ballots, the newspaper acknowledged, “carried a heavy Democratic tilt.” The Palm Beach Post is continuing to examine another 4,600 undervotes in the county plus 19,235 overvotes–ballots that contain more than one vote for a presidential candidate.

The Orlando Sentinel, the South Florida Sentinel and the Chicago Tribune looked at 15,596 undervotes and overvotes in 15 different counties. All but one were mostly Republican, and each one used paper ballots and optical-scan readers. These papers unearthed 1700 ballots “on which a voter’s choice for president could be easily determined.” Among these, Gore had a 366-vote lead. The Gore campaign had not pushed for recounts in any of these counties–which represent 4.6 percent of the total ballots cast in the state. The Orlando Sentinel reported that its review “found hundreds [of ballots] that were thrown out even though it was clear which candidates those voters wanted.” More than half of these ballots were rejected because the voter selected Bush and Gore and also wrote in the candidate’s name. Some voters used pen rather than pencil or made marks outside the designated oval. Should any candidate be credited ballots with mistakes? These figures don’t answer that debate. In any event, the Orlando Sentinel said, “It’s impossible to know how important the lost votes identified in the review may have been without examining the much larger number of uncounted ballots around the state.”

The Miami Herald review of the undervote throughout the state should be out soon. But in the meantime, the paper has reported that its analysis of the overvote in Miami-Dade, Broward, and Palm Beach counties showed Gore “lost far more votes than” Bush. According to the Herald, in Palm Beach county–home of the now-famous “butterfly ballot”–of the 500,000 votes cast, 5,264 ballots contained votes for both Gore and Buchanan. If Gore had picked up 11 percent of these, he would have won. But no official recount would have included such ballots. The Herald has also reported that at least 3,000 illegal ballots were cast throughout the state–by felons, by residents who were not properly registered, by people who voted twice. Obviously, there is no way to ascertain who benefited from these ballots. Nor can there be an exact count of citizens who went to the polls and were turned away by mistake.

A Washington Post analysis of the computerized records for 2.7 million votes of the eight largest counties in Florida concluded that “democratic voters were significantly more likely to have invalidated their ballots than Republican voters.” Gore, the paper found, “was by far most likely to be selected on invalid overvoted ballots, with his name punched as one of the choices on 46,000 of them. Bush, by comparison, was punched on 17,000.” Whether or not these ballots deserve counting, the trend of intent is clear. Trends, though, don’t trump individual ballots.

And this is all far from over. In the coming weeks, the major-media consortium will review the 180,000 undervotes and overvotes throughout the state. (The Florida state Supreme Court order mandating statewide hand recounts–which was overturned by the US Supreme Court–applied only to the undervotes.) This big-foot effort will produce no comprehensive, agreed-upon, end-of-story numbers. The University of Chicago’s National Opinion Research Center, which was hired by the consortium to examine the uncounted ballots, will compile an inventory of the ballots (so many dimples, so many hanging chads, so many outside-the-line ovals) and then leave it to each consortium member to analyze these findings. So hold on–even when the most extensive review is finished, there will be a host of competing findings.

In a better-run contest–with better machines, better pollworkers, and better voters (who assiduously followed instructions)–Gore would have won. Or so it seems at this point. Perhaps if Gore had fought for a state recount–instead of hand counts in only four counties–the election in Florida might have approached that better-run contest. But the system failed; Its error rate (which all systems have) surpassed the margin of victory. Gore may end up with various numbers on his side. But the argument–fuzzy math, some will shout–will never be settled. And the declared winner will not be moved or removed.