Someone help me understand this Star Tribune results vs. Minn State re: Franken

According to the Minnesota website it seems Coleman is still ahead:
http://ww2.startribune.com/news/metro/elections/returns/2008/recount/msenco.html

But the Secretary of State website has Franken ahead:
http://electionresults.sos.state.mn.us/20081104/SenateRecount.asp

What's what?

I don't even want to think about this anymore until this all goes to court.

:shrug:

Secretary of State count votes actually recounted (to be compared with votes from Nov 4 to see whether Franken made progress)

StarTribune started with Nov 4 votes and modified them by adding the vote change for each candidate in each county.

Both show that, at this stage of the road, Coleman has improved his numbers by about 100 votes.

Franken numbers, where he says he is winning, is using the challenged ballots and what the people who counted the votes thought they were showing. This would have to be confirmed in front of a different court.

When an election is this close (something like less a few thousandths of one percent) it is impossible to get a totally accurate count (recount, or re-recount) of voter intent. In essence, this is a statistical tie, and it probably should go to the next step in Minnesota law when a tie occurs: choose by lot. They should toss a coin. Much as I despise Norm Coleman, the people were split dead even on this one. Winning a recount by, say, 22 votes, does not give me any confidence in the system. It's a tie.

Remove all the other candidates but the top 2 vote getters. I think a system like this would give more chance to 3rd party candidates knowing that if there wasn't a definitive decision that it would go to recount.

I don't see how a coin toss determines the will of the people more than even a VERY slim margin.

As slim as the margin may be, it was still arrived at through, much like a court case, the adversarial process. We don't determine court cases by the flip of the coin. Both sides has hundreds of lawyers and volunteers monitoring this re-count.

If one side comes out on top, so be it. If the other side has a LEGITIMATE case to be made they can litigate it before either the canvasing board OR the courts. THAT'S how the truth is determined.

This isn't a game.

Many states have this provision for tie races, and coin-toss outcomes are not uncommon (though I personally have never heard of one at the senatorial level).

The problem is that there will never be an accurate count. Both revotes and prolonged litigation are expensive. (And if you recall Bush v. Gore, litigation doesn't always work out that well, does it?) I would opt for a runoff vote over litigation any day. Dropping the third-party candidate and revoting would be optimal. But unlike Georgia, MN doesn't seem to have this as a legal option. It does have a law requiring chosing by lots in the case of a tie. And given the natural flaws in the system of adjuding voter intent, this can only be called (at this point, at least) a statistical tie.

State law calls for the drawing of lots in a tie, not a "statistical tie."

There's a big difference.

If anything, it should go to the senate if there is any hanky panky - THAT'S the law. Especially if Coleman wins by less than the 46 votes that "disappeared" but were counted on 11/4 and match the log-in sheets.

In other words, all votes stay the same until they change. Since this is not a reVOTE and the results are unlikely to change drastically, this is the proper way to view the recount. The vote difference tells you the net change that needs to occur for the ultimate result to flip.

The SoS is tallying up votes from scratch as they are recounted. This yields misleading results since we already know the general breakdown of the remaining votes and the vote difference DOES NOT tell you the required net change for a Franken/Coleman win. What it tells you quite simply is the number of votes each candidate garnered if you ONLY include the recounted precincts. That means that this ~2000 vote lead can easily evaporate once areas that went for Coleman are factored in.

The SoS numbers are useless, IMO. I mean, if they recounted all of Franken's strongholds first, would we be expected to celebrate the resulting 200,000 vote "lead?"