Sign In
0 Items (
Search
Comments (59)
It now looks like Deep State criminals have tweaked voting machine software to switch Trump votes – possibly millions of them — to Biden to create the illusion of a Democrat victory. You don’t have to believe me on this. If you still retain a smidgen of high-school algebra, the data patterns in voting in the swing state of Michigan speak for themselves.
A team of three engineers and software analysts have demonstrated that a minimum of 69,000 votes there were switched from Trump to Biden. Biden is supposed to have won Michigan and its 16 Electoral College votes by about 146,000 votes in an electorate of 5.4 million.
If the vote switching occurred, the two candidates would be virtually neck and neck, Trump probably ahead. The same voting machines and their software were used through all the swing states. Another ‘if’: if similar fraud is accepted by courts to have prevailed in those states, the US is headed for a giant constitutional crisis.
The Michigan auditing team was Dr. Shiva Ayyadurai, an MIT trained data scientist and Fulbright Scholar and now Republican Senate candidate; Bennie Smith, a Democrat and software engineer, data analyst and an election commissioner; and Phil Evans, engineer and data analyst. They analysed voting patterns in Michigan’s four biggest counties and concluded the patterns must have resulted from an inserted algorithm that switched salami-slices of votes from the Trump tally to the Biden column. The greater the voting support for Trump and Republicans, the bigger was the slicing-off for Biden. This can be shown by straight trend lines in the graphs, which would normally display a wide scatter.
The prima facie and weird result would be that – absent fraud – the strongest Republican precincts demonstrated the weakest support for Trump. Even if Republicans in general didn’t like Trump compared with the party itself, there is no reason why the strongest Republican districts would be his worst detractors.
The trend line has a peculiar shape of flat followed by a straight slope of about 25 degrees. It prevailed across three counties — Oakland, Macomb and Kent — involving many hundreds of voting precincts each, and it prevailed in the same county when early votes were compared with those on election day. But the clincher for the algorithm’s presence is that in Michigan’s heavily Democrat Wayne County, the conspirators didn’t bother to use it. The resulting pattern among precincts was a limited scatter showing rising support for Trump personally as districts rated as more loyal to Republicanism. This data pattern is nothing like the pattern in the three fiercely-contested counties.
The fraud was easy to implement because, astoundingly, the voting machine software since 2001 has included a “weighted race” feature enabling votes to be skewed to one candidate or another, depending on the software settings. Why this feature was included has never been explained. Moreover, the machines take an image of each vote and these images are what generates the vote tally. Since the election, and contrary to federal law, Michigan officials have been busily deleting the images. This prevents any meaningful audit, other than using the paper votes which is barely practicable. Hence the Ayyadurai team is instead using statistical detective work to highlight the frauds.
In the video below Dr Ayyadurai sets it all out in layperson’s terms. Start watching at 13.30 minutes and you can stop at 50 minutes.
The “fix” can be explained but is better visualised in the charts below. In essence, voters (as in Australia) can vote for individuals, i.e. Donald Trump or Joe Biden, or for their party (Republican and Democrat) in which case their votes are allocated to the individuals. They can’t vote for both. Each precinct’s Republican vote, for Trump or for Party, can be separated for analysis.
A natural result would be for Trump and Party to get about the same vote percentage, regardless of how strongly or weakly Republican a precinct happens to be, whether 30 per cent or 80 per cent. Trump actually has so much charisma among supporters in precincts that he often gets a higher vote than his party.
The fraud analysis turns not on total Trump votes but on the margin favoring him compared to his party. If an algorithm is inserted to switch (individual) Trump votes to Biden, that margin of Trump above Party shrinks. As more Trump votes are switched, the margin fades to zero (50-50) and with even more votes switched, we suddenly find Republican voters are favoring their party above Mr Trump, i.e. his margin goes negative. And when vote switching becomes full-on, it might look as though Trump is being supported by only, say, 25 per cent of the Republican voters in that precinct. How likely is that?
The odd “shape” of the algorithm’s impact on the vote tally cries “fraud” from the rooftops. Where a precinct is heavily Democrat, the conspirators didn’t bother to switch Trump votes so the trend starts at flat. As precincts on the chart are mapped progressively more Republican, the trend abruptly goes on its downward slope, with Republican voters seeming to increasingly favor Party over Trump. That’s thanks to the secret switching. Indeed the inserted algorithm can be reverse engineered to its actual metrics.
This flat-and-then-downward slope occurs in the same precinct or county on different electoral periods, and in different precincts on the same day. To get this strange linear shape even once would be an outlier; to get it naturally over and over, time after time, is beyond reason.
Now here’s the graphed data and trends.
Chart One: The natural (uncorrupted) voting picture: Each blue spot is a voting precinct. Its position on the Y (long) axis represents that precinct’s overall support for Trump, e.g. 20% or 60%. On the X (tall) axis the zero line is Republican voters favoring Trump equally with the Republican Party. If the blue dot is above that zero line, it indicates how much extra support Trump got over his Party. If the blue dot is below the zero line, it indicates that voters rate his Party above Trump, and by how much. Note that the precinct dots just scatter above or below the zero line without trend. No tampering is in this hypothetical example.
Chart 2: A hypothetical test case – how Chart 1 would look if Trump for some reason was unpopular in his party. He would be voted less than his party, but this wouldn’t vary much whether the party was strong or weak in a given precinct.
Chart 3: A hypothetical “Red Flag” case for fraud: the natural random pattern has been transferred into a pattern where strongest Republican precincts strangely show the greatest rejection of Trump.
Chart 4: Actual evidence of vote-switching fraud – This chart shows the natural trend of Trump vs his Party being distorted into a weird flat-plus-30deg-slope, on two time periods at the same Oakland County. (The flat initial part is where the algorithm isn’t active). From these dots it can be calculated that the algorithm gave Biden a total 60,000 lead asd a result of 30,000 switched votes.
Chart 5: Actual fraud again — the identical pattern at Macomb County, with the conspirators switching 16,500 Trump votes to Biden.
Chart 6: The clincher – Democrat-heavy Wayne County, where conspirators didn’t bother to apply their algorithm. The story here is that even though Republican support was low, Republican voters heavily backed Trump compared with his party (dots clustering above the line in the 0-20% region of Republican support). And with no conspiratorial tampering, Trump support over party actually rose in precincts with greater overall Republican support, contrary to the “fix-is-in” graphs 4 and 5.
This is all just the first fruits of the election fraud investigations. USA politics have indeed moved into uncharted territory.
Tony Thomas’s new book, Come To Think Of It – essays to tickle the brain, is available here as a book ($34.95) or an e-book ($14.95)
This brings back memories of how it was done in the good old days.
In a two party for SRC president at a sandstone uni, the ballots were all paper.
When the incorruptible returning officer was not looking, whole bundles of votes were thrown down a spiral staircase to be captured at the bottom before being counted.
The fraudsters then threw out the votes for the non preferred candidate and then returned the ballots for the ‘count’. Such was the blatancy that this subterfuge was carried on during two recounts.
The gentle returning officer began suspecting a problem when on each recount, the total vote dropped and the ‘preferred’ candidate’s proportion of votes rose.
Great work Tony. This looks unimpeachable because it negates the suggestion that late (mail-in) votes would naturally favour Biden.
This analysis works because Michigan has party ticket and individual voting. It would be interesting to know what is used in the other swing states and what methodology could be used to identify the switch (if any) in those states.
It’s that Red Queen at work again!
It’s certainly ambitious. They needed a lot of votes, so they couldn’t settle for a “Trump is unpopular with Republicans” as shown in Chart 2 above. That may have been the original plan, which then had to be hastily hacked to get more votes for Biden. Doing it this way is dangerous, because it’s so clear in analysis. But if you’re going to steal a lot of votes, what better place than where the voting is heavily Republican? Trump retains his majorities in Republican areas, so anomalies are not immediately obvious. If it weren’t for the straight-ticket voting, this analysis could not have been done.
Arizona is the most clearly anomalous state. It may be a time-bomb waiting to go off, and this analysis may not be possible there.
People like Tucker Carlson, Sean Hannity and about 70 million others are asking, “How do fix this machinery?” The USA has had major political scandals in the past, and the republic has recovered from them. The challenge of this one seems to be much deeper. Little pieces of the machinery of governance and of news reporting and of opinion moulding have been deliberately smashed over an extended period, and the result has been the sort of dysfunction now visible in the election. The repeated warnings have been ignored or suppressed so thoroughly that it takes a constitutional crisis like this to show how daring these “enemies of the people” have become. That process of ignoring and suppressing may yet be successful in this instance. If they get away with it, they will rightly ask themselves, “What can’t we get away with?”
The USA is not ready for democracy. They should call in Xi or Vlad to arbitrate.
In my opinion, Trump won by a landslide.
I would have thought there would be more likely an upward trend in the normal case for Trump? If so this would make the recorded downward trend even more anomalous – dark red flags!!
This is how the vote machines work. Watch it to the end. One computer to verify identity, another to input the vote and print out a form, the third tallies the votes.
Has anyone checked whether the Q code that the machine reads matches the displayed vote typed out in English?
https://www.youtube.com/watch?v=zEWYhRYshK0
I need to inject a note of caution here. I have been looking at a site that claims to debunk Dr Ayyadurai:
https://kabir-naim.medium.com/dr-shiva-ayyadurai-the-danger-of-data-charlatans-4f675ffe793c
I am still working through the maths, which once upon a time I would have done with ease, but time and a few glasses of red have dulled my wits somewhat. So I invite more qualified readers than me to have a look at the site.
But the immediate red flag for me is that the author Naim Kabir claims that neither Macomb nor Oakland use the Dominion software. He provides a link but I have not been able to access the site to confirm this.
I have confirmed that neither Macomb nor Oakland use Dominion software.
I’ve read Naim Kabir’s piece and he is certainly confident in his rebuttal. My maths is baby maths so like Peter I’ll await any expert advice.
Hi Tony,
I went to bed thinking I’d get back to this in the morning but, of course, I couldn’t go to sleep. I think I’ve worked out where Naim Kabir is coming from and I believe he’s got it arse about face. He claims that the line is based on the formula Y = -X +%split-Trump votes. But Ayyadurai is not graphing a formula, he is plotting a series of data points and trying to work out an equation that would fit. That equation would have to involve a constant – an estimation of the probability that Trump votes and Party votes would not deviate beyond a certain range. His theory is that the deviation is likely to be constant (+/-) regardless of the nature of the precinct. That seems reasonable to me but statistics was always my weakest subject. It would help if we saw the same analysis from the Democrat perspective. I need to think more about this but I’m sure we have a QOL reader who is better qualified than me to check it. But the fact that Macomb and Oakland don’t use the software bothers me.
Peter OBrien
>”But the fact that Macomb and Oakland don’t use the software bothers me”
Yes, that is an issue. What was used ?
But so is this:
” … Since the election, and contrary to federal law, Michigan officials have been busily deleting the images” That should deter any analysis or recount.
And as noted in an earlier thread, Democrat winning margins in highly marginal areas are just sufficient (>0.5%) to avoid a scrutinised, automatic recount. Perhaps coincidence, but a spooky one.
ianl,
Oakland uses Hart InterCivic and Macomb uses Election Systems Software. As you say, the claim that images are being deleted is also of concern – if it is true. The key to this would be to see the same analysis from the Democrat perspective and I can’t understand why this was not done. If they showed a horizontal line, that would give weight to the good Doctor’s theory but he may have to look elsewhere for the mechanism.
A problem seems to lie in plotting the difference of two percentages.
https://youtu.be/aokNwKx7gM8
The difficulty for the statistically illiterate, such as me, here is that the quantities being graphed are somewhat arcane. The X-axis (horizontal) shows the percentage of total “party” votes that were cast for the Republicans, while the Y (vertical axis) shows the percentage of Presidential-candidate-only votes for Trump, MINUS the corresponding X axis value—the “party” votes.
The Y-axis units mean that any constant percentage of Trump President-only votes will result in a straight but descending (right to left) line on the graph. You can see this easily by mapping a few points on the graph at constant Trump support of say, 100% or 20%. Mr Kabir is at pains to point out the inevitability of this, perhaps disingenuously, as Dr Ayyadurai assumes correlation between Republican “party” and Trump presidential voting percentages in each precinct in (and only in) which case the line should be flat. The “stolen vote” numbers rely on this assumption.
It is when the descending line crosses the X-axis that Dr Ayyadurai’s point might be better illustrated: below the X-axis, in negative territory, every data point (precinct) shows Mr Trump’s percentage of President-only votes is less than the Republican percentage of Party-only ones. The bulk of Trump’s data points are below the X-axis; Mr Biden’s, I suppose, above it. Not actually proof of anything, but suspicious. Very.
Sidney Powell: ‘…and I’m very concerned it not only involved Dominion and its Smartmatic software but that the software was essentially used by other election machines also.’
So it looks like the contagion spread beyond just Dominion and its subsidiaries. Rather than discrediting Dr. Shiva’s analysis, he has simply demonstrated this fact with the data. She also claimed that, ‘We have sworn witness testimony about why the software was designed. It was designed to rig elections…. We have so much evidence I feel like it’s coming in through a fire hose.’
The story is growing. We know that it will be kept out of the MSM and social media so it will be interesting to see how this plays out if the courts start getting serious.
Obviously, the Hart InterCivic and Election Systems Software contained the same or similar vote-shifting software as the Dominion machines.
How else to explain the data?
Earlier I asked would the same analysis from a Democrat perspective show the same result. This bloke claims to have done that and shown that it is the same:
https://www.youtube.com/watch?v=aokNwKx7gM8
He might turn out to be the BS artist, but, for the moment, I’m not pinning my faith on Dr Ayyadurai, without some independent verification of his work. If he is proven incorrect, that does not, of course mean that fraud didn’t happen.
Kabir’s analysis is wrong, almost laughably so. The formula itself, ‘y = -x + c’, is (partially) correct. It is also correct that y = -x + c results in a negatively sloping line at 45 degrees, shifted up or down the y axis by the value of ‘c’, presuming that ‘c’ is a constant value. (Year 8 maths).
.
It is in his handling of the value ‘c’ that he digresses into la-la land.
.
He uses a pretty straightforward statistical distribution to come up with variable values for ‘c’, but essentially he treats it as a constant that varies only slightly around a mean. In the example he normalises the distribution around an average of +30% of individual Trump votes. The formula effectively becomes
y = -x + 30%, which he plots in his downward sloping graph. If x = 0%, then y = 30%. If x = 80%, then y = -50%. But you can pick any mean, and the line will always slope negatively at 45 degrees. Simples.
.
In other words, Kabir claims that the vote count for individual votes is completely disconnected from the vote count for the GOP (he states that himself, and also inaccurately claims that Shiva is doing the same). So to Kabir it doesn’t matter whether you have an 80% or a 10% vote for the GOP in a precinct, that has nothing to do with the trend for the individual votes. (Imagine presenting that argument to the lefties!)
.
What he doesn’t get is that ‘c’ is a variable that changes with the value of ‘x’ in the real world. In essence (and this is Shiva’s premise), the formula should be ‘y = 0’, because ‘c = x’, or ‘y = -x + x’.
(Or more accurately, c = x multiplied by a distribution function variable, which is much the same thing:
‘y = -x + (f)x’, where ‘f’ is a statistical function).
.
The irony is that the data is showing exactly what Kabir describes – a distribution that is entirely disconnected from reality, in that ‘c’ has become a fixed variable rather than a value that closely follows ‘x’.
.
Kabir’s analysis is statistics for morons.
.
This guy markets himself as a data scientist at Even.com. That is scary, especially for customers of Even.com. I think we understand now why we get such ridiculous models for things like viral mortality rates.
I’ve realised that I missed a small detail in the above. The complete formula is
y = -x + f(x) + c
There is a constant c that can shift the graph up or down the y axis. Shiva also notes this constant. For example there could be an additional proportion of Trump-only supporters who don’t like the GOP in general and won’t vote for them. So if you have 80% GOP voters you might have 85% Trump voters. If this forms a general trend, then c = 5%. So you end up with y = -x + x +5%, or y = 5%.
Do I win a prize Tony?
Does it really need to be this complicated? Before everyone disappears forever down the statistical rabbit-hole, I think the following paragraph is the most operative of the article:
“Moreover, the machines take an image of each vote and these images are what generates the vote tally. Since the election, and contrary to federal law, Michigan officials have been busily deleting the images. This prevents any meaningful audit, other than using the paper votes which is barely practicable. Hence the Ayyadurai team is instead using statistical detective work to highlight the frauds”.
In short, this “statistical detective work” ain’t gonna get much traction. If it gets to court, it’ll be red vs blue techno-nerds at ten paces, and there will be enough grey area in between for the benefit of the doubt to remain with the official result (which is effectively 9/10 of the law).
A more fruitful line of inquiry would be to conduct an investigation into the alleged deletion of images that is “contrary to federal law”. Those making such serious allegations would presumably have good prima facie evidence and could therefore force the matter to be heard in the appropriate court. The court would then have no option but to order a properly scrutinised recount of the original physical ballots, or a subset of the ballots from the “suspect” counties. This would not be “barely practicable”. It would in fact be an absolute necessity, to be conducted with rigour and dispatch. No lies, damned lies and stats required. Just good old-fashioned vote counting.
Ed Dowling
Naim Kabir has made a foolish fundamental error. He has assumed that the X-axis represents ” the % of straight-ticket Republican voters”. It doesn’t. It represents the TOTAL of Republican voters, ie the straight-ticket Republican voters PLUS the Individual-Candidate Republican voters. His criticism is totally unfounded.
“Deep state”? How childish. It’s it just a bogeyman, a lazy slogan that relieves both the user and the reader of the need for critical thought, and presents a convenient target for a two-minutes’ hate. So, these people reckon that the ‘deep state’ has perpetrated some sort of IT fix to engineer a vast electoral fraud. It’s an interesting theory, and by all means expose and investigate it. But you simply adopt it and present it as irrefutable fact in a supposedly intelligent and critical magazine, though you have no qualifications in the field at all. That’s how conspiracy theorists behave. Conspiracy theories always founder on the sheer numbers of people in on the terrible secret, perfectly kept – ordinary Americans with families and mortgages who are so morally abandoned that not one has breathed a word of it over, apparently, decades. 9/11 is a classic example, and this one, I’m afraid, has the same smell.
tblack, some of us do have an understanding of statistical methods and also bother to look further into these claims. We are all awaiting an official investigation – that is the point – but on the face of the evidence available to the public, what can be seen is not what we would expect to see. Presented as fact? Is that what the sharing of information is? You are putting the cart before the horse. The whole point of examining this data is to determine whether or not it is a conspiracy theory or whether it is a horrible problem that needs to be fixed. So far the best argument against the analysis, as I have shown, is childish nonsense.
Also, the term ‘Deep State’ is a commonly used descriptor for widespread corruption within the US elite, particularly in Washington. You may not believe it exists, but it still requires a name for those who are willing to explore the possibility that not all is as it should be.
The Wikipedia tells us that Doc Shiva is consistently unreliable – see https://en.wikipedia.org/wiki/Shiva_Ayyadurai The Wiki often tells the truth but their agenda is more important.
Bennie Smith reads as clean – see https://www.netrootsnation.org/profile/benniejsmithgmail-com/
Phil Evans is n Australian with a decent background – if this link is the right one – https://www.linkedin.com/in/philip-evans-407291122/?originalSubdomain=au
Dr Shiva et al tell us that the people manipulating votes are stealing them from Don but not from the Republicans. This implies stupidity. If they robbed them both equally this pattern would not be seen.
https://medium.com/@jennycohn1/the-corruption-of-americas-computerized-elections-3d5b77124ebe
written by Jennifer Cohn [ [ Cohen misspelt ] @jennycohn1 covers the criminals involved
I can tell you that I don’t know a single Republican who didn’t plan to vote for Trump again–and know many Republicans who voted for Trump this time who didn’t in 2016 (including my beloved bride, who went from not voting for him to being more full-on MAGA than me). Unscientific sample, of course, but given that I openly identify as a reluctant last-minute Trump voter in 2016 and have been honest in my criticisms of his failings, I believe that my Republican friends would be honest with me–they would know that I wouldn’t give them grief. This doesn’t pass the smell test.
john2 what’s complicated about it? It’s just statistics. It’s hardly even that – it’s natural progressions. There is a natural data set that isn’t natural. Voting precincts that trend Republican are not trending towards Trump for their president, but rather towards Biden. That is not natural. And it is occurring in a linear fashion that has the appearance of design. It smells.
This kind of analysis can be a very powerful tool for determining what is going on and for providing sufficient evidence to scrutinise the software and hardware used. Only after that happens can the public get to the heart of the matter.
Tony didn’t just “share information”. He stated as absolute fact that members of the “Deep State” – a silly, meaningless term – had conspired to commit the criminal act of rigging the election.
Who are the members of the ‘deep state’? No, I don’t mean the identity of individual members – let’s think about groups. Financiers? Captains of industry? The military? ‘Political insiders’ so beloved of lazy journalists? Bill Gates? George Soros? The jews, perhaps? Anyway, somebody must have originated the ‘conspiracy’, as Tony expressly called it. Did these people sit around a table, or have a Teams meeting, organising and mapping out the strategy? Next, how did they enlist the executives of the conspiracy? How was the plan then coordinated and communicated to all the footsoldiers and by whom? Presumably the footsoldiers were the dark hordes of Sanders-loving dems – odd allies of the big-money types in the Deep State. But in any event, as with all conspiracy theories it unravels absurdly at this point. The sheer numbers of morally depraved people involved – yet not a word of it has leaked out, from top to bottom. From the bigwigs in the ‘deep state’ who devised, planned and directed the whole conspiracy; to their underlings who carried out their orders in monolithic silence; to the footsoldiers who executed it. Not to mention, of course, the IT people who worked on, tested and maintained the software. No whispered attack of conscience among the hundreds of people who would have been needed to give effect to such a wicked conspiracy.
These sorts of things only happen in Hollywood movies. It’s astounding to see Quadrant, with its history of intellectual rigour, of exposing smelly orthodoxies and generally skewering nonsense, should be rabbiting on about the ‘deep state’.
Re Peter O’Brien 16 November:
The video you provided a link to does not, in my opinion, refute Dr Ayyadurai’s analysis. In fact it supports it. At about 50 seconds in, scatter plots of both Trump and Biden voting are presented, purporting to show similar voting patterns.
The first thing to notice is that different portions of the Y-axis are shown. Mr Biden’s data points are centred roughly around +10%, and Mr Trump’s around -10%. The second is that one graph is in fact the complement of the other; to compare the two patterns we should reverse the X-axis of one of them. If Republicans receive, for example, 75% of party-only votes, Democrats get 25% (excluding minor parties, if any) and vice versa. When this reversal is performed it is clear that when one pattern trends up, the other trends down—but with Mr Trump’s pattern lower on the graph.
This is all consistent with Dr Ayyadurai’s hypothesis that some president-only votes for Mr Trump were allocated to Mr Biden, in proportion to the Republican party-only vote in each precinct. There may perhaps be some sociological explanation for the apparent voting pattern, but it is hard to imagine what it could be.
Reply to tblack
I endorse your rejection of conspiracy theories for they are ill disguised ideology; but they must be distinguished from demonstratively real conspiracies common throughout history.
You are simply wrong to dismiss the Deep State as “a silly, meaningless term”. American society has moved on far beyond the industrial military complex (which has been studied in the closest detail), with intelligence now playing a dominant role.
The National Security Agency has secretly broken into the main communications links that connect Yahoo and Google data centres around the world, according to documents made public by Edward Snowden. Without entering into the controversy over the elections, if you place the NSA at the heart of the Deep State, your long list of objections framed in terms of personal becomes largely irrelevant.
Tony, Peter O’Brien and pgang.
I share your reservations about the use of statistics – red vs blue nerds at ten paces.
Here is a Maths bloke from Aus via UK (who is known to me as left-leaning from past viewing) who has tested the analysis by Dr Shiva. Have a look at this and see if it rings true: https://www.youtube.com/watch?v=aokNwKx7gM8
I note that statistical analysis can only point to areas requiring further investigation and thus it is a tool for researching circumstantial evidence.
It isn’t a smoking gun tool but, put together with other factors such as breaching the law by destroying the scanned images can lead to at least a recount.
Tony Thomas – 14th November 2020
His attitude doesn’t impress me. He seems to disparage the analysis as a con, rather than plain erroneous. I hear the sound of a grinding axe in the background. What is Kabir’s view of the election and the challenges to its validity?
Matt Parker does not question the motives or disparage the makers of the error although he notes that we are all biased and so mistakes can be overlooked by failing to apply your theory to proving your opponent’s case (or disproving your own case).
Did you notice the problem in that video talldad? He makes a very good point about the direct correlation method – that’s how I would have treated the data. Then look at the two graphs he creates closely.
….they don’t correlate. Note the shift in the axes between the Biden and Trump data.
I’m guessing that’s why Shiva et al chose the presentation method that they did. I have run out of time to look at this more closely, but note that his formula at the end corresponds to mine above. I am presuming that factor ‘m’ in y = mx + b plays a role in the correlation skew.
Without looking closer I am still of a mind that a linear transformation has taken place dependent on the ratio of GOP votes.
Have a read of this. It gives me some hope anyway:
https://www.breitbart.com/politics/2020/11/16/sidney-powell-lin-wood-promise-theres-ton-voter-fraud/
I might as well pull out the old “argument from authority” and note that I’m a professional mathematician / data scientist. And I might as well throw in another logical fallacy, which is that my sceptical analysis below is more credible for the fact that I’m a big Trump fan, concerned about some apparently genuine reports of electoral irregularities, and concerned also at the irresponsible, almost universal characterisation by the media of Biden as the president-elect while Trump is still disputing the legality of Biden’s reputed win.
Dr Ayyadurai posits the theory that across the precincts of a county, the Trump vote rate (t) should roughly equal the Republican vote rate (r) plus a constant (c), for all r. i.e. t(r) = r + c. Rearranging, we get t(r) – r = c, where the form on the left hand side is what he plots as a function of r, so that he expects a constant “y-value”. He then demonstrates, apparently correctly, that the resulting pattern does not show a constant trend.
However, Dr Ayyadurai gives no compelling reason for why t(r) – r should be constant for different r. All the downward trend means is that Trump’s fraction of the candidate votes, relative to the Republican fraction of the party votes, decreases as we move to more strongly party-Republican precincts. This is all but guaranteed for the simple reason that towards the right-hand side of the plot, the Republican fraction approaches 1 by design, and the Trump fraction is a different quantity and need not approach 1, but may approach something less, say, 0.7. This limiting value could be taken as a measure of Trump’s popularity among candidate-voters in precincts where the party-voters are all Republican. If we plotted r(t) – t as a function of t, we’d see the similar effect that as t approaches 1, r approaches something less than 1, say 0.8. This value would then represent the Republican party’s popularity among party-voters in precincts where the candidate-voters all chose Trump.
The biggest red flags in Ayyadurai’s video are that he only produced a single plot representing a “normal” county (a sample of size 1 does not establish what is normal) — which also featured impossible points representing negative Trump vote fractions — and he didn’t present the Biden/Democrat counterparts to his Trump/Republican plots. This fellow did: https://www.youtube.com/watch?v=aokNwKx7gM8. Assuming that he got his own logic right, it shows that the same downward trend appears in a Biden/Democrat plot. I think it’s plausible in view of my reasoning in the preceding paragraph.
In short, the downward trend is introduced by design, and features not only in the corresponding Biden/Democrat plots, but also in plots that reverse the roles of the party and candidate votes. This myth is busted.
90,000 ballots in three states are the balance of victory.
PS As r passes from 0 to 1, t passes from something larger than 0 to something smaller than 1, because those precincts which yield an extremal r will not generally yield an extremal t. This is what explains the basic downward trend of Ayyadurai’s plots, from positive at r = 0 to negative at r = 1, if there are such precincts.
Stuart, thank you for your analysis. What we have needed from the start, as I pointed out 4 days ago. And your point re the same logic in relation to the Democrat ticket reproducing the same result, I also queried 3 days ago.
That’s not to say fraud did not occur and on a massive scale, It is up to the courts to decide that. And here I’m not so sanguine. There is already push back from hostile courts which leaves the Supreme Court as the final hope. Will the Trump lawyers be able to prove enough fraud, beyond reasonable doubt, to overturn the trend? I suspect that the Supreme Court, knowing they are secure for life and thus fully entrenched denizens of the swamp, will look for any and all scintilla of doubt in order to avoid making the highly fraught decision to deny Biden the Presidency, on the spurious basis that the country must be united and they can keep their powder dry for more important issues than giving Donald Trump another (well deserved) term in office. As far as they are concerned, having appointed three of them, his work is done.
Stuart I’m with you. I got fed up with trying to decipher what was going on so I downloaded the data for Kent county myself. There is nothing to be alarmed about in the slope of the graph. When the Trump plus GOP votes are looked at together, Trump’s share of the combined vote decreases as the GOP share increases. The same thing happens for Biden, and Jorgensen. So you are correct: t(r) – r is not a constant.
Biden’s data resulted in a very similar graph. The only difference is that Biden’s graph is shifted up the y axis by around 8%, which simply reflects his voting lead over Trump.
Trump’s share of the total Republican vote is consistently around 60.7%. Biden is a little less consistent but his average is higher, at 65% of the total Democrat vote.
There is no smoking gun in the Kent data from the perspective of Trump’s share of the Republican vote from what I can see. However that does not rule out the possibility that Biden’s tallies were artificially inflated. Only a look at the time sequence and the software could determine that.
However I will also add that the slope of the graph for Kent is very different to that presented for Oakland, which has a much steeper gradient. I can’t actually find the Oakland data to compare the two.
It would seem that for Kent at least, the question of why the GOP received more votes than the Democrats, yet Sleepy Joe smashed Dynamic Don as a candidate, remains a mystery.
Re Stuart J Burrows, 17 Nov.
With respect, Mr Burrows, in my opinion—and I’m happy to be corrected—the myth has not been busted, but rather bolstered by the Biden/Democrat graph shown in the video to which you referred. At about 50 seconds in, scatter plots of both Trump and Biden voting are presented, purporting to show similar voting patterns.
The first thing to notice is that different portions of the Y-axis are shown. Mr Biden’s data points are centred roughly around +10%, and Mr Trump’s around -10%. The second is that the Biden graph’s X-axis of 0 to 100% Democrat party votes is equivalent to 100% to 0 Republican; to compare the two patterns we should reverse the X-axis of one of them. If we were to compare the two plots ON THE SAME AXES, they converge on one side and diverge towards the other, in a “V on its side” shape. On a Republican 0 to 100% X-axis they diverge to the right, and on a Democrat one, to the left. That is to say, as Republican Party vote percentage increases (and therefore Democrat percentage decreases) Mr Trump’s presidential-only vote percentage decreases while Mr Biden’s increases.
This is consistent with Dr Ayyadurai’s hypothesis that some president-only votes for Mr Trump were allocated to Mr Biden, in proportion to the Republican party-only vote in each precinct. There may perhaps be some sociological explanation for the apparent voting pattern, but it is hard to imagine what it could be.
Peter O’Brien my rational side agrees with you. The courts might poke around the edges but they won’t flip states. The Georgia recount will most likely be a fizzer because it’s probably just counting the same thing fraudulent votes again. Still, stranger things have happened.
But it’s probably a good time to start learning Chinese.
Why would a court ‘flip states’ unless there were some evidence presented to it that would permit it to do so?
And if the evidence was presented, why would it not do so?
In the fevered imagination of some here, the courts, especially the Supreme Court, it’s the double headed penny. Somehow, the Courts ought to overturn the result even if no evidence is presented in a particular proceeding that would permit such a result. On the other hand, even if the evidence is presented, the members of the Supreme Court (even the Trump appointments) would simply junk their their judicial oaths and disregard it.
That kind of thing happens on Netflix, not in the real world.
By the way Peter O’Brien, these are civil claims, and the standard of proof is the preponderance of the evidence, not beyond reasonable doubt.
Good grief. How did Quadrant become the happy hunting ground of the conspiracy theorist? Krygier and McAuley would be turning in their graves.
tblack, I take your point re ‘preponderance of evidence’ vs ‘reasonable doubt’ but my point was that US courts are political and they will find legal reasoning to support their preferred judgement. However, I think it unlikely that the conservative judges on this Supreme Court will exert themselves unduly to find in favour of Trump. I would not say the same of the Democrat inclined justices if the position were reversed. And if you think there was no conspiracy to deny Trump a second term you are living in an alternate universe.
A pleasure, Peter O’Brien. Welcome aboard, pgang.
jvernau7 — I think I understand you. Define for each precinct:
r fraction of party voters who voted Republican
t fraction of candidate voters who voted Trump
d fraction of party voters who voted Democrat
b fraction of candidate voters who voted Biden
(Note that here these are point values, not variables.) Then plot on the x-y plane:
points (r, t−r) for each precinct: red scatter
points (r, b−d) for each precinct: green scatter
(I’ll reserve the colour blue for the original Biden/Democrat scatter, (d, b−d).)
Assuming that these are the only two parties and candidates, we have b = 1−t and d = 1−r, so:
red scatter = points (r, t−r)
green scatter = points (r, r−t)
i.e. The y-values of the green scatter are the negative of the y-values of the red scatter. And this is what we see (after reflecting the blue Biden/Democrat scatter about x = 0.5 to get the green scatter).
Now define
trump(x) the average Trump vote fraction for precincts of Republican vote fraction x
biden(x) the average Biden vote fraction for precincts of Democrat vote fraction x
y = red(x) the curve through the red scatter
y = blue(x) the curve through the blue scatter
By definition of the red and blue scatters, à la Ayyadurai, we have:
red(x) = trump(x) − x
blue(x) = biden(x) − x
so
blue(x) − red(x) = biden(x) − trump(x)
We can see from the scatter plots at the 50s mark of the response video, that blue(x) − red(x) is about 15% for all x.
Therefore biden(x) − trump(x) is about 15% for all x. So the higher vertical position of the Biden scatter relative to the Trump scatter has the following meaning.
When Biden is on p%-Democrat turf, his vote exceeds the vote that Trump is getting on p%-Republican turf, by about 15% — for any percentage p.
For example, if the county had just four precincts, they might look like this:
%Rep | %Trump | % Dem | % Biden
————————————–
40 | [52] | [60] | 48 ⟵ Trump more popular than Rs, Biden less popular than Ds
[60] | 33 | 40 | [67] ⟵ But 67−52 = 48−33 = 15
[70] | [61] | 30 | 39 ⟵ Trump less popular than Rs, Biden more popular than Ds
30 | 24 | [70] | [76] ⟵ 76−61 = 39−24 = 15
Every possible winning pair appears once.
Further to my last regarding US courts (from Breitbart):
“The Pennsylvania Supreme Court rejected the Trump campaign’s complaint that observers were not given adequate access to watch the counting process, holding Tuesday that the law did not specify that observers had to be close enough to see details.
Observers complained that they were kept so far away in Philadelphia that they could not see the markings on absentee ballot envelopes.
The campaign lost at the trial court, won on appeal at the commonwealth court, but lost again at the state Supreme Court.
The decision was a 5-2 split for Pennsylvania’s highest court, reflecting the partisan divide among the judges. In September, a majority on the court ordered vote-by-mail ballots to be accepted up to three days after Election Day, even without a postmark. A 5-2 majority also granted a Democratic demand for the Green Party to be excluded from the ballot.
The U.S. Supreme Court allowed the vote-by-mail ruling to stand, 4-4, after Chief Justice John Roberts sided with the liberal minority.”
I rest my case.
Stuart, that is what I saw in the Kent data, with the difference being around 8%. Should there be such consistency? I would think so given that ballots are a measure of average sentiment, but I guess you would have to compare it with other patterns. Maybe you could argue that a percentage of Trump ballots were flipped to Biden. If I was looking for fraud I’d be looking for a consistent pattern within the real time data.
.
Of more interest is another story from Gateway Pundit in which they show real time data from Michigan and Georgia. Apparently once Biden had snuck into the lead, nearly all the ballots added afterwards were incrementally split 50.102 Biden to 49.898 Trump (or variations thereof depending on the county). The exact same fraction for all of the ballot dumps.
If they are reporting accurately then there is absolutely no question that the voting was controlled by the software.
pgang,
● green(x) = −red(x)
● blue(x) = green(1−x)
Therefore
● blue(x) − red(x) = green(1−x) + green(x)
If green(x) is linear, with green(x) = m x + b, then
● blue(x) − red(x) = m + 2 b = 2 green(0.5) = 2 blue(0.5) = −2 red(0.5)
Now green(x) is linear if and only if all of blue(x), red(x), trump(x) and biden(x) are also linear. Therefore if the candidate votes vary linearly with the party votes, the y-difference between the Biden/Democrat and Trump/Republican Ayyadurai plots will be constant across different party vote fractions. Bear in mind that we’ve assumed that there are only two parties and two candidates.
Giuliani on the spate of lawsuits now underway – https://youtube.com/watch?v=t9Wxa0DNdIQ
Mr Burrows,
Thanks for your helpful reply. I think I’m beginning to see the light, although the squinting has brought on something of a headache. My mathematics are rusty after decades of neglect.
My reversing of the X-axis was prompted by the fact that data points (precinct results) for each candidate are not independent of each other; as you succinctly put it:
“red scatter = points (r, t−r)
green scatter = points (r, r−t)”
… so each data point on one candidate’s scatter plot implies one on the other’s. If we accept Dr Ayyadurai’s implied assumption that we should expect a candidate’s percentage vote to be always proportional to his party’s party-only percentage, both plots would be oriented horizontally. As you have pointed out, the downward trend doesn’t show necessary vote-shifting. Rather, it disproves the Doctor’s assumption.
In regard to your statement that “When Biden is on p%-Democrat turf, his vote exceeds the vote that Trump is getting on p%-Republican turf, by about 15% — for any percentage p.” which explains the vertically-displaced (but parallel) scatters, this doesn’t look particularly suspicious either. I suppose it could well be that more Republicans are “anyone-but-Trumpers” than Democrats are against Mr Biden, and that they represent a fixed percentage of Republicans.
I believe that Jvernau7 is absolutely right when he says that a comparison of the Trump and Biden graphs strongly supports the view that fraudulent vote manipulation has taken place. The graphs show that in this County more than 80% of Trump votes were made via the straight-ticket Party option, whereas in Biden’s case 80% of the votes were via the Candidate option. That distorted pattern is exactly the kind of result that would be expected from an algorithmic manipulation.
Hi jvernau7,
It’s great to have your engagement on these questions! One really has to be careful about making conclusions from a single view of a complicated dataset.
I just thought of another interesting angle. Dr Ayyadurai’s assumption is that
● trump(x) − x = c
where trump(x) is the average Trump vote fraction in precincts of Republican vote fraction x, and c is a constant. Now substitute in two values for x, 0 and 1:
● trump(0) − 0 = c
● trump(1) − 1 = c
Substituting the first, c = trump(0), into the second to eliminate c gives
● trump(1) − trump(0) = 1
Now both the domain and range of the “trump” function are the unit interval. Therefore
● trump(0) = 0
● trump(1) = 1
But we saw above that c = trump(0). Therefore c = 0, and Ayyadurai’s assumption implies that
● trump(x) = x
In other words, he is effectively assuming that the average Trump vote fraction is always equal to the Republican vote fraction, which of course need not be the case. And the non-zero horizontal trends he countenances as being normal are inconsistent with the assumption.
The “never-Trump” attitude certainly might help to explain the vertical shift of the Biden/Democrat plot.
The data and the graphs clearly show that the voting pattern in each of the 3 Michigan Counties was as follows:
% of Votes using Party method % of Votes using Candidate method
Votes for Trump ~80 ~20
Votes for Biden ~20 ~80
This very large discrepancy between the voting pattern for the two candidates can only have resulted from an illegal action during the vote counting process. The reversed nature of voting ratios is consistent with the illegal transfer of a large number of Trump “candidate” votes to Biden’s “candidate” vote total.
Peter, the decision of the Pennsylvania Supreme Court. The two dissenters, including the Chief Justice, held that the point was entirely moot in any event. The Chief Justice said in his judgment: “Short of demonstrated fraud, the notion that presumptively valid ballots cast by the Pennsylvania electorate would be disregarded based on isolated procedural irregularities that have been redressed — thus disenfranchising potentially thousands of voters — is misguided.”
Here is the link, if you are interested:
http://www.pacourts.us/assets/files/setting-7742/file-10581.pdf?cb=1c606b
It now seems that the numbers involved this supposed conspiracy must be in the many thousands. As I have said before, it’s fatuous to think that not a whisper of this has leaked out, and that not a single footsoldier who was approached rejected the idea and blew the whistle, except, apparently, some postal worker who backed off when questioned in any event. Conspiracy theories involving such vast, well-oiled machinery, so many people, and such slick execution in total, monolithic silence only happen in movies. 9/11 all over again. Take the blinkers off. The idea is childish – move on.
tblack you are ignoring what is going on quite frankly. Your quote answers to the question in itself – it assumes the absence of fraud, and is therefore somewhat moot to the discussion.
There are many more people coming forward with affidavits than a single postal worker. It’s not a conspiracy theory if there is testimony from hundreds of witnesses, plus the expert opinion of analysts who are all in agreement (not that we are there yet). Particularly if the only cases against the theory are, ‘there’s nothing to see here’, or, ‘it’s just too incredible to be true’.
Dr Shiva’s case has been shown to be flawed. It doesn’t matter, as it is only a minor piece in a puzzle of growing complexity. Perhaps nothing sinister is going on, or can be proved, but that seems to be increasingly unlikely. If a tenth of what Sydney Powell is claiming can be evidenced, we are looking at the most devastating political scandal in the history of Western democracy – one with the potential to spark events that we don’t want to have to think about. You don’t just ignore what someone of Powell’s status and credentials has to say. None of this may result in a Trump presidency – that still remains highly unlikely – but what is important is that the USA gets to the bottom of it and provides a satisfactory conclusion one way or the other.
You move on if you like, the rest of us will continue to watch actual developments rather than ignoring them.
Pgang, I agree with everything you say, except when you state that Dr Shiva’s case has been shown to be flawed. If you are referring to the rebuttal by Tendor (https://youtu.be/aokNwKx7gM8), I have checked his statement and he has clearly misunderstood the significance of his own data analysis. The fact that the Trump and Biden graphs have exactly the same shape is totally irrelevant. What is important is that the data clearly shows that the Party/Candidate vote ratio is 80%/20% in Trump’s case and 20%/80% in Biden’s. If it can be demonstrated to a Court that the difference between the 2 types of vote has never for any Candidate in any election exceeded, say, 20 percentage points, then the incredible 60 percentage points difference that was recorded in these 3 counties must surely indicate unmistakable evidence of fraud.
Pgang,
but there isn’t ‘testimony from ‘hundreds of witnesses’. That’s the problem.
Stuart J:
‘Ayyadurai’s assumption implies that
● trump(x) = x’
.
Hey Stuart, you were just catching up with what I said ages ago!
.
‘What he doesn’t get is that ‘c’ is a variable that changes with the value of ‘x’ in the real world. In essence (and this is Shiva’s premise), the formula should be ‘y = 0’, because ‘c = x’, or ‘y = -x + x’.’
.
That is precisely his assumption, which I think you have shown is incorrect, and which I now disagree with also. It ignores the fact that there is a closed universe of Trump supporters, who may either vote Trump or GOP. If more of them vote for GOP, then there are fewer remaining to vote for Trump. The high GOP ratio probably reflects people who don’t like Trump very much, and vice versa.