Let's see if I've got this straight...

[Rex_Goodheart]

USCV says that the WPE's were higher in heavy Bush precincts, which one would expect in cases of fraud.

But Febble says "no, no, no... we have to apply a 'bias index' to adjust those higher WPE'S to get rid of a computational 'artifact'."

Well, gee... Let's go even one better... Let's just make an even BIGGER bias index, and show that it was BUSH who got cheated in the election, who actually would have won by 6 million votes, if it weren't for Democratic corruption.

[autorank]

1. Frankly Freaky Function Facilitating Fascist Cheerleaders

I've got the boys in the back room working on it. I'll let you know when I'm hiring a PR firm. It's going to be hot, at least 6,000,000 * votes were simply not included in the tally.

It was a MAN DATE!

[Kip Humphrey]

2. Some naive points and questions

Our investigations into fraud suggests 3 forms:

Voter suppression: estimated at 1+ million and indeterminable by reported vote or exit polls

evoting machine vote switching exploits: restricted to precincts with evoting machines. By design of the algorithms, would be hard to detect because in most instances, the vote switching is subject to chance whereby the effect of machine exploits are ameliorated by the number of voters who spotted and corrected their votes manipulated by the exploit. Furthermore, the suspected employed algorithms spread the vote switching from Kerry to: a) bush, b) no vote, c) 3rd party candidates, while maintaining the number of votes switched to below the thresholds of detection and using a variable factor for switching which seems to average out to 4% but will vary from tabulator to tabulator (since evoting machines are configured through GEMS running on the tabulators)

Tabulator vote switching: would not be detectable at precinct level, but would show up if the data were clustered by tabulator (county-level or state level or a combination of both); would not have been employed across the board but only in certain states on certain tabulators.


NAIVE QUESTIONS:
How can any algorithm applied uniformly to exit poll data do other than act to reduce (or, conversely, exaggerate) the small spikes suspected fraud would produce?

If vote manipulation is small overall (say, 4% average), is not uniformly applied, and is variable by design, wouldn't it simply appear as random noise either at the precinct or state level?

If there was a general swing to Kerry, wouldn't those tabulator clusters left alone (no vote switching) show a swing to Kerry in the reported vote? For those tabulators where votes were switched away from Kerry, wouldn't those tabulators show a swing to bush and slightly higher reported votes for 3rd parties and slightly higher undervoting?

If properly applied, wouldn't an amalgamation of manipulated and non-manipulated tabulator results appear to be random either at the precinct or state levels?

Let's assume the exit poll raw data were correct (assuming Mitofsky's nuclear defense that they don't conduct exit polls well is false). If the vote manipulation algorithm employed on select tabulators mirrored the actual slope caused by a swing to Kerry at the tabulator level with a slight edge given bush, wouldn't the exit poll scatter plot appear to show a balanced distribution of swings to Kerry (actual) and swings to bush (actual + manipulated)?

I have to go to work so can't continue my naive questions.