An error rate of 68.05% means that 34.025% of the ballots were switched.
That’s not my understanding. There was a 68% error rate in all votes cast. Those ballots could be flagged for error for various reasons. What matters is that of those “erronous” ballots cast, someone had to adjudacate them and assign them to Biden or Trump. That is why the error rate is so high. So that poll workers could manipulate the total.
Section J line 3 Error rates page 18
explains in a very concise way exactly how this was done.
Lines 3,4, 5 and 6 should be read and understood by all Freepers to see exactly how the election was stolen.
Thank you for pointing out the ratio-weighted algorithm. I was confirming the PRECISE accuracy of the forensics report; I should not have just written “switched” as this word is imprecise. I apologize if I (unintentionally) misled anyone.
I presume you want to get into the arithmetic weeds. As can be seen at J2 in the first bullet point, the REPORT total error rate is limited by Michigan’s enjoined mandate of federal guidance at 1/125,000 = .0008%, which AMOUNTS TO 1/250,000 ballots being incorrectly reported by a line of tabulating machines that has been proposed for purchase and use. Those evaluating the equipment are only interested in the machine’s report total, not why the machine reports wrongly.
The initial numbers reported out of Antrim County showed our President at 4,509 votes with 98% of precincts reporting. The final numbers reported by Antrim after correction gave 9,783 votes to President Trump. This is a difference of 5,274 votes. 100% of the precincts revealed an even larger difference of 5,333 votes.
15,972 votes were cast in Antrim for President including third party and write-in. The forensics report discounts the write-ins and prefers 15,949 Presidential votes for consideration. The tabulators account for 15,676 votes; 296 votes were recorded that the tabulators did not read. We are concerned with the 15,676 votes handled by the tabulators.
10,667 ballots suffered alteration within the machines before bogus whole integer numbers were spit out. Some machines performed ratios leading to a greater than 80% error rate while others were closer to 55%. The aggregate for the entire lot of cheating machines as recorded by the server was 68.04699585...%. 10,667 altered ballots out of 15,676 ballots handled by the tabulators is whence the ~68% figure derives.
5,333 were weighted light for Trump while 5,334 were weighted heavy for Biden, with different weightings on different tabulators. These various weightings affecting 10,667 ballots (5,333 + 5,334) HAVE THE SAME EFFECT as switching 5,333.5 ballots, which number (5333.5/10,667) is ~34.025%.
The shrewd evil of internal fractional apportionment is that it usually escapes audit. The number of ballots equals the server-reported total. Even more devious is coding the harmed ballots as in need of adjudication and then approving them to the nearest whole number. More devious yet is buying time to manufacture ballots for defeating a hand count.
Again, I apologize for MY imprecision. The precision of the forensics report is untouchable. There were many posts belittling the forensics report, and even belittling Ramsland, although Ramsland is just the CEO; his technical people did a great job. Freegards.