Data Showing Systematic Vote Theft in Michigan
by John Cleer - ABCU|8
This is almost a month old but still relevant as it relates to some of Sidney Powell's claims. Sundance mentioned it when it came out and the comments there might be worth a glance. [1]
MIT Ph.D. Shiva Ayyadurai, who goes by "Dr. Shiva", tracked the vote by precinct in each of Michigan's four largest counties: three are "battleground" counties and the fourth is liberal Wayne County/Detroit, which serves as a control group of sorts: Ayyadurai's study found a statistically impossible vote distribution repeat itself perfectly in the three swing counties, and not show itself at all in Wayne County.
He attributed this distribution to a "weighted vote" algorithm that siphoned off votes from Trump and gave them to Biden in direct proportion to how Republican a precinct was (and how many Republican votes it had to transfer).
https://www.youtube.com/watch?v=Ztu5Y5obWPk
Not everyone has 70 minutes to watch this so I'll summarize. I saw it around the time it came out and found the subsequent "debunkings" unconvincing and probably fraudulent. I also saw a number of armchair commenters on leftist sites who clearly did not understand the relationship that was being graphed, taking strong positions against Ayyadurai's data.
Ayyadurai ran for Senate in Massachusetts as an Independent (not as a Republican, as I've seen reported) and as I recall he claims the election was rigged against him using similar methods to the ones he describes here. He may have pioneered the term "election fraud" in this video, arguing that voter fraud wrongly pins the blame on the voters who are impersonated. Here's the gist of his study:
Ayyadurai made a separate graph for each of the counties by plotting its precincts. The graph is simple once you figure it out but you'll have to think about it for a minute (I did): The X-axis quantifies how "Red" a precinct voted in 2020, by computing the percentage of its "Straight Ticket" votes that went Straight Republican. The Y-axis quantifies Trump's performance relative to the other Republicans on the ticket (senators, city council, etc.)
So on the left you have the blue precincts plotted and as you go right the plotted precincts become more and more conservative; on top you have the precincts where Trump overperformed (received more votes than) downballot Republicans, and on bottom they overperformed him.
In all three swing counties, Trump overperformed the other Republicans on the ticket (by around 7%), but only in precincts where the straight ticket vote was less than 20% Republican.
For precincts that were over 20% Republican (by straight ticket vote), Trump's performance relative to downballot Republicans declined linearly, in direct proportion to how Republican the straight ticket was: the more Republican the precinct voted, the worse Trump did against the other Republicans.
This relationship did not present itself in Wayne County, where almost every precinct is all the way left and consequently very few votes would be transferred: in Wayne County, Trump continued to overperform downballot Republicans in precincts with a high percentage of Straight Tickets going Republican.
Thus, in Macomb, Kent and Oakland Counties, you can see the precincts plotted stay level as you move from 0-20% Republican, and from there decline sharply and move further down as you move further Right: like a hockey stick. In Wayne County, where most all of the precincts are on the Left, the average of the precincts moves in a straight line to the right, without breaking down at 20%.
This is less eye-grabbing to a layman than a statistician but what it shows unambiguously is the data and therefore the votes being rigged: even if somehow the Republicans in red precincts favored Trump less than they do in blue precincts, as implausible as this would be, it would never add up to the precincts breaking downward at exactly 20% and maintaining a perfect line to the end.
Ayyadurai attributes this to a "weighted vote" being done by algorithm and names Diebold voting systems in 2001 as the first to use such a "weighted vote" system, by which one candidate's votes would be multiplied by a different number than his opponent–say Candidate A's votes would be multiplied by 1.25 and candidate B's by .75; if they then received the same number of votes, Candidate B would have a quarter of his votes transferred to his opponent.
He cites a local law that was made to do this legally and posits that it's still being done, possibly all over the country.
Powell has referred to these weighted votes, which she sometimes called fractional votes, happening in Georgia as summarized here.
Here is a source of hers, Russ Ramsland, saying that fractional voting was used to rig Michigan.
https://www.youtube.com/watch?v=Lg4_Qd_rYVc
The 7 minute clip includes a 2016 excerpt from Bloomberg Technology's Digital Defense webcast, in which the narrator describes Diebold's vote tabulation system, "Gems":
Gems is a vote tabulation system that was created by a company once known as Diebold. A lot of people think Diebold is dead, it's not, it's just been absorbed into other vote technology companies. An interesting thing about the gems tabulator is not just that it's used in many of the elections that many of you may participate in, but it has a 'bug' or a 'feature' as its creators would call it.
"Bug" of course is a synonym for "glitch".
They then quote Bennie Smith, who later contributed to Ayyadurai's study, saying "having access to decimals [fractional voting] allows you to perform and finance what they call an allocation," such as 'make sure Smith gets 27.64%' (on screen example) …"you have to be in line with what polling data says, if it's a close race you have to make it close… it allows for decimal precision."
If I understand him right, Smith also suggests that a precise program like this could be, and was, used to sell votes.
To conclude: My goal here was to present this tabulation problem/"feature" in a way that makes sense to laymen and gives insight into why Dominion's system warrants skepticism and investigation. Powell is litigating this in multiple states, and Trump's team has forensically examined 22 Dominion machines from Antrim County, MI. By the timeline in that article, the results should be announced on or around the evening of December 9th.
Sources:
[1] https://theconservativetreehouse.com/2020/11/12/mit-phd-and-statistician-outlines-algorithmic-fingerprint-within-vote-data/
[2] https://www.newsmax.com/newsmax-tv/sidney-powell-campaign-lawyer-dominion/2020/11/21/id/998181/
[3] https://www.wnd.com/2020/12/trump-team-begins-examination-dominion-voting-machines/