This is very compelling. Dr.Shiva provides an MIT PhD analysis of MMichigan voting patterns; and the data reveals some irrefutable evidence of what has taken place behind U.S. voting systems. The video can be shortened to start at 18:00 BUT STAY WITH IT.
By looking at two specific metrics: (1) Republican straight party vote percentages [x-axis]; and (2) the percentage of people who voted for Trump without voting for a straight Republican ticket [y-axis]; mathematicians and data scientists (Phil Evans) are able to find the exact moment when natural voting patterns are manipulated by algorithmic triggers.
.
What am I missing? (I am not a statistician and my last stat’s class was in 1967.) If I understand the y axis, it is the % of votes for Trump by the ticket splitters minus the % of the straight party line votes. Take the extreme example, an extremely red precinct votes 80% straight party. That leaves only 20% of the voters left, 80% of the remaining voters would have to vote for Trump or else you would have a negative y axis value. So you either have 96% of the precinct voting for Trump or you have a negative value. Do the math, 80% plus 80% of 20% equals 96%.
You would expect the value will be negative, only Saddam got 96% of the vote. If this is right there is no magic here at all, it is the expected outcome. McIntyre this morning plotted the 2016 election in one of the counties involve and got the same “slope.”
Maybe I am missing something?
Yes the biggest thing of all…. is… DONT get lost in the Weeds
Don’t get lost in the weeds.. I could write a program that had 4 or 5 subroutines to vary the numbers at any data input.. we don’t need to exactly figure out each line of computer code….
It’s already been approximately done …..there is significant proof that the election software was not accurate…. this leads to a mandated manual recount.by hand…yes yes yes…. actual results ….thats how the first 6000 vote error was found manual recount.. they did not have the computer code or program.. they were not in the weeds
The manual recount shows what the computer program should have tallied against what was reported on the paper printout… do that in 5 counties in one state it’s a state problem..
Do recounts in 2 states and have 50 % of the precincts with errors that enough to throw the national election and possible house and senate majorities… that leads to the need to do recounts in all 50 states…
Match the local paper print out with the numbers in the servers.. and differences here indicate a second program manipulating outcomes… I don’t think these crooks only had only one routine.. and they could modify the end result field as wanted….
In the end.. hand counting will prove who legally won the election….. there are lots of ways to cheat.. President Trump needs to have manual votes counted in each state… not only the states he won.. only that way can the American people have faith in the 2020 election
MAGA
The recount must be statewide, not just selected counties. Bush v. Gore.
Electrical gadgets like computers can get people hoisted by their own petard. I am surprised they did not have on sight shredding machines for Republican ballots. I guess the thought was it was too much to recount and were counting on public laziness and being tuned out to get away with computer assisted fraud.
An audit that is based upon MI federal laws passed by the Legislature. Photo ID, signature verification etc.
The problem begins in your second sentence and continues in the third.
Using Oakland county as an example, in a precinct with 30 registered republicans out of 100 voters, 30 voted for Trump. In a precinct with 60 registered republicans out of 100 voters, 48 voted for Trump. There is no need to calculate a percentage based on remaining voters.
As the number of registered republicans goes up, the percentage of those not voting for Trump goes up and it does so in a straight line beginning around 20% registered republicans. This is what the graph indicates. I am not speculating on the accuracy of the data or the credibility of the presenter.
Neither Shiva nor MacIntyre calculated percentage of republicans not voting Trump. They counted Trump-only versus straight-ticket R’s, both of which voted for Trump.
I posted a comment explaining why (IMO) the Shiva graph is something one would expect with or without fraud, i.e., correlation between mixed Trump votes and all-R Trump votes becomes increasingly negative as the latter percentage increases. That and another comment have disappeared and it looks like they were deleted, although logically much the same as several other people’s arguments here. What gives, admins?
Fraud occurred and flipped the election. But Shiva’s presentation isn’t the smoking gun proof of that, unless there is something that MacIntyre and many others are missing.
If I had access to the data sets, I’d be looking at Trump’s performance in within-state bellwether counties. He won most of the national ones and it helps build the case if he also won the counties that predict victory within each swing states.
A more conclusive approach would have examined the Biden data as a comparison to see if a different trend was seen — or at least past election data. Without that, its difficult to ascertain if the Trump trend was an odd phenomenon or the expectation.
I suspect Shiva’s analysis may be at least partly flawed due to self-correlation in the way he’s set his variables up
Can all you guys in this discussion please have a look downthread where I’ve posted a big black screen capture of his variables
… I’d appreciate some critiquing by firmer minds, I’m 9 years into retirement and my analytical capability has gone somewhat soft
thx.
The Shiva analysis is flawed – I’ve looked at it. Take an example precinct with 5 Republican (R) and 5 Democrat (D). If 4R vote straight party (SP) and 1D – you get an X axis of 80%R on his chart. Then for the non-(SP) 1R votes Trump and 4D vote Biden – you get -60% on the Y axis of his chart. However think about it, this would be a 50/50 precinct with a 50/50 vote for Trump, this guy would label this as a strong 80%R precinct with a huge differential in (-60%) the NSP votes. Yet all the R voted Trump and all the D voted Biden.
The assumption is that the non-(SP) pool of voters should have a similar makeup as the (SP) voters in that precinct. That linear graph says that the more the (SP) pool of voters leans republicans the more the non-(SP) pool of voters trends away from republican. If on average 80% of Republicans vote (SP) and 20% of Democrats vote (SP) like in your example it still wouldn’t explain the linear trend, it would just show Trump underperforming.
yes, exactly and this is the flawed assumption that he makes near the beginning of the video. Two flaws that I see at the moment 1) these are not weighted by votes – so all the little blue dots are just precincts but not the vote totals 2) That the percentage of SP votes for R should be similar relative to the NSP votes for Trump (big assumption). I used the example above to illustrate how his assumptions would be flawed in a 50/50 precinct with 50% registered R and D where all R votes Trump and all D Biden. it would plot a blue point in the lower right quadrant and he would call that an 80% x axis SP with a -60%. I think his analysis is done in good faith, but it doesn’t show what he thinks it shows. what it is showing is at the extreme’s of SP voting, the remaining NSP (non straight party) will have a larger differential to the candidate. That example 50/50 precinct, but this time 4D SP vote and 1R SP. The NSP then has 4R Trump and 1D Biden. Chart X axis 20% (Trump SP) y axis +60%. You could almost say the linear down Left quadrant high to right quadrant low is a function of how he chose to do the analysis.
I think you have nailed it. I watched this and was thinking to myself, that something isn’t right. I settled on the fact that he was making some broad assumptions about the makeup of voting pools and not accounting for how they can shift depending on how partisan it is.
Your post made it click. I wanted to see him do Bidens/Dem votes and I wanted to see the Trump/R vote in a red state that uses different software like OK. If that showed different results then….ok.
To be clear. I think there is clear fraud here that is much easier to show. The problem is that none of it is actual proof. Its like opening the fridge at work to get your sandwich but its gone. You know someone stole it but there isn’t any proof of who stole it, or that you even brought the sandwich to work.
It the same with the poll workers affidavits. Sure you can show they were excluded and prevented from doing their job but how does that prove fraud actually took place.
Everything I have seen falls into this category. Unfortunately the way ballots were processed in these same suspect counties will prevent anyone from proving anything.
How would your analysis change under the assumption that 91-94% of R’s voted for Trump?
Ok – so first I don’t see that in the actual data from Precincts – but let me take a very high actual case in Michigan Kent County. Grattan Township Precinct 1 – 70% Trump total vote. Straight Ticket 525 Republican 155 Democrat. Non straight Party (NSP) 330 Trump, 218 Biden. Shiva’s X axis for this precinct would show 77% (525/(525+155)) the NSP 60% (330/(330+218)) difference from 77% = -16% Y Axis. Note this should be an actual blue dot in the data he showed on Kent County. Now what you are asking is how many registered republicans are there in this precinct and how many are registered democrats – I would be curious if you can get me this data then I could look at. But still this wouldn’t tell you that all registered republicans voted or all registered democrats voted (turnout of registered voters was 77.28% in this township). We would need at least figures on how many registered R’s voters and how many registered D’s in this precinct. I’m not sure that is published.
Just one thing to add – my assumption in the hypothetical was that 100% of Republicans voted for Trump. That was the point of the example, even in that scenario – you still see the result is 80%, -60% for which Shiva is claiming is indicator of stolen votes.
Yes, that´s a big assumption, as voting “against” is usual were minorities live everyday rounded by “enemies” But tendency line croosing x-axis is not.
In first graph (Oakland I think) declining starts at 20% in x-axis and cuts at 40% in same axis (20% diference). In Macomb declines at 40% and cuts x-axis at 80% (40% double diference). It does not sense from a statisitical point of view. The more aggregated the figure the bigger the sharping.
Whether right or wrong, you’ve posted this same thing all over both pages of comments here – which not only violates CTH posting guidelines, but is also highly suspect.. What is your investment here?
My investment? I’m a two time Trump voter that lives in Switzerland and I vote in Kent County. My concern is that Shiva’s analysis is flawed and I see all the comments and see concern in buying into it where there will be not just disappointment but a loss of credibility. More analysis needs to be done. Everyone saying this proves X – when I’m not sure it does. He could be right in his conclusion without being right in the analysis. I also have an MBA from the University of Michigan and have spent 16 years of my career investigating fraud – white collar fraud, accounting irregularities, etc. I’m a CPA and listen I suspect serious fraud in this election to, the most major of which was allowing massive mail in voting which is forbidden and highly restricted in most countries. Most countries that allow absentee is for citizens living out of country only. But yes, I believe that Shiva’s analysis is flawed and is not proving what he thinks it is proving and it could be dangerous to push this. By the way I have followed this community since Travon Martin, and highly respect the investigative work, and I mean real work, that is done here.
Have you considered emailing Sundance directly? He thought this research was worthy of a special article.
I hadn’t – good point. I’m so fired up about this election fraud – and worried certain things are going to be used to discredit this site, gateway pundit etc.
I stopped at…”University of Michigan”…
As in you think everyone that went to Michigan is a liberal or like you know who Steve Deace is?
I no longer use the word liberal.
It does not fit.
Marxist, fascist…they fit.
You just answered your own question.
ah ok, you don’t know who Steve Deace is and you were insulting me. To quote one of my favorite comedians “you can fix a lot of things, but you can’t fix stupid”.
That is true. Mirrors work 100% of the time.
John R
I don’t see anything in your curriculum vitae which would make you a more qualified statistician than Dr. Shiva. I find your concern regarding election fraud to be very suspect. If you are so hot and bothered by election fraud why are you aiding and abetting it by attacking Dr.Shiva’s presentation with minutiae. You obviously have an agenda, but it seems to me to be the opposite of what you contend.
Its not minutiae – I’m not more qualified, but his analysis is not correct, so I guess you heard it from me first is all I can say. The flawed assumption is that the percentage straight party republicans should hold in the non-straight party votes for Trump. That is a big assumption even if there were only Rs and Ds. with independents it makes this assumption even more of a, I would say clear mistake. I think his analysis seems to have been done in good faith, the math is correct, I have run sample data from Michigan. The problem isn’t the math. It’s that the linear line down left to right is what you would expect to see because of how he defined the x and y axis. I’m more than qualified to find a flawed assumption.
John R said:
“… the linear line down left to right is what you would expect to see because of how he defined the x and y axis”
I’m getting at the same point about x and y variables downthread where I’ve captured his definitions in a screen cap (look for the big black box near end of the thread)
John R, would appreciate your critique and comment on what I presented
thx
@In the Land of Poz. Yes, I agree with you. I have been looking for someone that get’s it. Shiva did something interesting – but it doesn’t say what he thinks it does and it is going to damage credibility. I saw Shiva on Gateway pundit, and did some analysis. I was hoping this wasn’t going to show up on the treehouse. Don’t get me wrong – I still think there is fraud here, and even I am still concerned about Kent County Michigan, but not due to the Shiva analysis.
I wonder how an individual voter, could file a class action suit, to be sure their vote for Trump, did not get flipped to the senile Plugs. That’s as infuriating, as all get’s out.
I’m asking for a friend, who lives’ in Kent county MI.
Here is something else strange about Kent County. where I voted as well. Decision Desk still shows Kent County for biden 186,753 to 165,318 (54% Biden). I pull up: https://www.accesskent.com/Departments/Elections/Results/
Follow me here and click on the link for yourself. Under the heading August 4, 2020 – State Primary Election – results summary – if you click the link it’s not the Primary – the document is Nov 3 2020 unoffical results. It shows on straight party republican 90,649 and then Trump votes on 2nd page non straight party 142,105 versus Biden at SP 66,835 and NSP 125,966. Ok that is 425,55 votes compared to today’s Decision Desk number in Kent County of 352,071. Decision Desk shows Biden 54% but this result shows Trump 237,754 versus Biden 192,801 – Meaning Kent County went Trump 55%. It shows Decision Desk is wrong or outdated, they have the county as blue. You can check this all out.
What is strange is that this report is posted under the wrong label in the Kent County website – they linked to the wrong report, but the report looks like 59 pages of real results for the Nov 3 election.
Is this an early indicator that the vote totals reported by the Media are wrong. I would go with the official Kent county website showing with more votes counted Trump won 55% in Kent County instead of losing with 46%.
Very strange!!!
note this change in Kent County alone represent 61,388 net votes that should got to Trump over what is reported in Media (Decision Desk).
also strange – I’m looking in more detail at the report now – I’ve downloaded the PDF. It shows date of 11/4/2020 4:15am with 56.7% of Precincts reporting. So it must be what I said above is not correct – you can’t add the SP votes to Trump they already did that on page 2. But it does show that at 4:15am on the 4th Trump’s lead in Kent County was 53% / 47% 142,105 v 125,966. The votes that came in after this (if decision desk is accurate) after 4am swung 72.4% Biden – 23,213 for Trump and 60,787 were added to totals for Biden.
Can u check which areas reported results after 4 a.m. 11/04?
Biden’s 72.4% share of the votes reported after 4 a.m. 11/04 is suspicious. Why? It exceeds his share of the total vote (i.e., from the start of the vote count to the end) in Kent County’s 2 most Democrat areas. The shares are about 60% in East Grand Rapids and 68% in Grand Rapids.
Note, I just checked – the data at 4:14 am is not detailed at the county level. So it isn’t possible with this to tell where the votes came in at such an improbable percentage.
@DM – yes this can be checked when they finalize the votes by comparing to this report I downloaded that comes from 4am Nov 4. More people than just me should download this report – so we can look to see what changed.
Ben, I did hear about voters suing. I think that this has merit. This is infuriating because no evidence yet exists for substantial proof. I like the stolen sandwich analogy.
The data Shiva used are available via election reporting dot com.
@DK, There’s two buckets of voters with each totaling to 100%. Those who voted straight ticket and those who chose each candidate. They make an assumption that the % in each of those categories should be closely related and that may not be true. Let’s say you’re in a district of Independents of 1000 people and only 10 vote straight ticket, 8 Republican & 2 democrat and the other 990 vote however. His chart would put that in the Repub at 80% and expect to see 80% vote for Trump at the individual level which would be unlikely. That’s an extreme case, Anyway, the pattern should have been more chaotic. If we think of it as those who vote straight ticket are hardcore party voters whereas those who vote for individuals are more independent minded, then there should be a pattern, but there is one with a distinct negative slope and that may indicate that something is “guiding” it.
Yes, you misunderstood.
Y is all the Trump votes (including the R straight ticket voters) minus the straight R ticket votes.
What remains are the democrats that voted Trump, the Independents that voted Trump and the Republicans that voted Trump, but did not vote all Republican down the ticket.
This analysis is flawed – I’ve looked at it. Take an example precinct with 5 Republican (R) and 5 Democrat (D). If 4R vote straight party (SP) and 1D – you get an X axis of 80%R on his chart. Then for the non-(SP) 1R votes Trump and 4D vote Biden – you get -60% on the Y axis of his chart. However think about it, this would be a 50/50 precinct with a 50/50 vote for Trump, this guy would label this as a strong 80%R precinct with a huge differential in (-60%) the NSP votes. Yet all the R voted Trump and all the D voted Biden.
Your example is a one-off analogy. That’s why Shiva and Team’s analysis plots all the data in an effort to simply show a trend that appears to have been affected by an algorithm as can be seen from the clear linear trends as opposed to random plots.
No it’s not – I have looked at the Michigan counties and precinct data – other posts go through in detail. I used the example to illustrate the flaw in his assumptions to make it easier to understand. Its fine if you don’t understand it.
I understand analysis and statistics.
The data is publicly available. Analyze it and prove that no algorithms were used in the Dominion counting systems, so you can prove your arguments.
Your examples are weak relative to the OBJECTIVES that Shiva and Team are attempting to demonstrate whether you want to believe it or not.
The entire explanation is quite intriguing for non-statisticians. on Dr. Shiva’s twitter page @va_shiva he explains the graphs. At about the 13 minute mark is the synopsis.
Dominion has a number of issues, one being votes can be weighted; in other words a vote for A=1.5 and a vote for B=.05.
The main phenomenon Dr. Shiva & the others found is that as you get into “Red” districts the % difference between those who voted a straight party ticket yet voted Biden INCREASES as you get into more Republican districts and it creates an upside down hockey stick on the graph.
I’m not a statistician either but the entire video is extremely compelling. Another video on YouTube by CDMedia breaks down the fraud in the 2018 election KY Governors race.
I thought exactly the same thing. I want to see the results for Biden/Democrat straight ticket. Is this slope a natural function of the population dataset using these percentages? Or is it in fact a sign of manipulation?
You are close, but watch the video carefully.
The x axis is the percent of straight ticket voters who vote R, not the the percent of voters who vote straight R. If you have a precinct with 10,000 voters and only 10 straight ticket voters, and if 8 of those 10 vote straight R, Shiva would put 80% on the X axis.
The y axis is the percentage of non-straight ticket voters who vote for Trump minus the percent of Republican straight ticket voters.
So, if the 9,990 non-straight ticket voters voted 50% for Trump and 50% for Biden, Shiva would assign 50-80 = negative 30 for the y axis, and would conclude that 30% of the votes were switched to Biden.
<— Mechanical Engineer.
Yes, you are missing something. He is saying they have an option to vote straight party with one mark of the ballot. The order of the precincts is most Democrat (left side) to most Republican (right side). Voters in the most Democrat precincts voted for (D) except for President – voting Donald Trump. While at the same time voters in the most Republican precincts voted straight (R) ticket except for President – voting Joe Biden.
The fact this is not a linear distribution, combined with the fact there is a repeatable change in function PROVES there is fraud. This is not a naturally occurring distribution. As they said – prove them wrong (can’t happen).
Yeah, you’re missing what is happening.
The dots represent only the incremental deficit or surplus of Trump votes (%) cast by the ticket-splitters, compared to the party-line voters in that precinct.
The expectation is that the percentage of Trump voters among ticket-splitters (of either party) in a precinct will be roughly in line (within, say, a margin of 5-10% either positive or negative) with the number of party-line Republicans in that precinct.
It beggars rational expectation for Trump to do significantly worse among voters in a heavily Republican precinct than was expressed by party-line votes in that precinct, especially such that exactly zero of them show a positive surplus for Trump (the old “every glitch seems to favor the Democrat” problem). And it absolutely defies logic that the deficit increases as the Republican vote in that precinct increases.
For those who were reading Steve McIntyre threads…
Noticed on POTUS daily thread that Solomon went on Lou to say ‘voting computer programs’ are not looking to be as suspicious as they thought. AKA McIntyre correct.
Bartiromo’s comments about Cuba… will probably be the tie in. Money poured in to precincts to shove ballots into the scanners. I like her less than I used to. Liz Mac seems to be the honest one out there these days.
I don’t care. PA should have sequestered ballots they didn’t. They stopped counting and they shouldn’t have.
Legislatures do your jobs!!!!
The problem with McIntyre’s analysis is that his overview graphs are way too general, we are only dealing with the transfer of about 2% of Trump’s votes overall and we do not know where the formula kicks in the counties that McIntyre reviews. Dr Shiva was looking at counties where is started at 20-30 percent R. The formula, we will find is adjustable. But looking at his AL Jefferson County graph, one can see that at higher %R the slope of Mix R decreases significantly compared to the slope of Str R , i.e. The Mix R votes being transferred to Biden. It appears to start at 25% R but it is very hard to see where it starts. Dr Shiva is way ahead of everyone on this. Thanks Sundance, great and needed post of Dr Shiva.
I disagree, I believe the Shiva analysis is flawed – I’ve looked at it. Take an example precinct with 5 Republican (R) and 5 Democrat (D). If 4R vote straight party (SP) and 1D – you get an X axis of 80%R on his chart. Then for the non-(SP) 1R votes Trump and 4D vote Biden – you get -60% on the Y axis of his chart. However think about it, this would be a 50/50 precinct with a 50/50 vote for Trump, this guy would label this as a strong 80%R precinct with a huge differential in (-60%) the NSP votes. Yet all the R voted Trump and all the D voted Biden. Don’t put your confidence in this analysis, it can be easily discredited as a conspiracy theory.
So you’ve told us — for the 1000th time. Unbelievable.
But John R assures us that he is concerned about election fraud.
Your analysis is faulty. You have postulated an extremely unlikely real-world voting pattern, then complain that Shiva’s analysis flags it as unlikely. Real-world voting patterns do not distribute in any way resembling the synthetic dataset you have constructed.You can’t complain when a technique designed to highlight improbability flags an improbable dataset.
Tell you what – take another state – that you trust the results from that doesn’t use the software, run the same analysis as Shiva. My prediction is you get a very similar result.
Or a simpler way to explain it. I ran the exact same analysis with Kent Count data – get the same scatter plot as Shiva. I then analyzed in the same way for Biden. Guess what. Shiva’s analysis formula shows the exact same downward slopping line for Biden v Trump. Biden X axis is % democrat straight ticket with Y Axis % Biden split ticket less %Dem straight ticket. Which means he would conclude that there is an algorithm that was shifting votes for Biden. Try it for yourself.
sorry – shifting votes for Trump on the Biden analysis.
Is it possible to use a grass roots effort in every state to get this issue on the ballot. I know, I know the irony of trying to change how we count the vote by getting it on the ballot seems counterproductive. I’m sick and tired of the protest mentality, call your congressman, or let’s get this out to the news to try to resolve issues. I’m a little late in the game on this one; maybe others know about grass root organizations really trying to tackle this on a state level.
Loralie…I shall agree with you.
It is a bit late for any of that.
So glad to see this show up on here. (Long time lurker, first-time post-er). Been following this angle for a few days now (38 years it I/T / programming / big-data / analytics). There is NO WAY in a hot-dark-place this wasn’t a manipulated election. Period. End.
Be careful though – this analysis is not showing what people think it is. I agree with you there was a lot going on in this election to get to the bottom of, but if we look at these kind of kook analysis (I know he has a PHD from MIT supposedly) but his analysis does not say what he thinks it does. At least he is making erroneous assumptions.
I’m tired of your repetitive posts as if you know more than Shiva AND THE OTHER 2 GENTLEMEN that are experts in election fraud manipulation.
By the way, all the data is publicly available and posted. Stop with the commentary and provide some data to show that no algorithm was used to affect the counts in those Michigan precincts.
In other words, Trump won by a landslide and we are in the middle of a coup. That is the upshot of this impressive video.
A lot of things are now clear. 1. Why the rush to confirm Biden as the president elect is to distract from stealing this crucial election. The non-issue Covid is another weapon of mass distraction. 2. The nauseating articles about him picking his positions. It lends legitimacy. 3. That legislatively unsupported and completely unconstitutional title of ‘The Office of the Presidental Elect’ is bandied around to, again, give the impression of legitimacy and inevitability.
It is all a fraud. And, it is a straight-up coup.
Since leftists have lost the momentum and are no longer convincing, what will they do next? Will they allow a terrorist to detonate an atomic bomb in flyover country? Do you thing I am joking? Brandon Van Grack, who unjustly prosecuted Gen Flynn, worked in an office dealing with weapons of mass destruction. Do think you that it is beyond the realm of possibility that someone in the deep state would allow them to proceed with their nuclear terror just like allowing those two muslims in Garland, Texas, to attempt to shoot up the cartoon contest?
I think not. Plus, I think that the Chinese government would reward them with a big fat bonus in their next payola.
Compelling. President Trump’s first terminations after November 3rd were with the Nuclear Energy Department.
Huh…
It seems that a very good control would be to run the same analysis in another state (preferably not a swing state) that has party line voting. You could look at the data by county and see what patterns and trend lines form. My guess is that more counties would resemble the Wayne County graphic and that there would be considerably more randomness in the deviations.
I have already seen this in NYS, it is so obvious.
Upstate, in red counties within a blue state where they did NOT need to change the President’s popularity and vote, he was running ahead in northern counties.
Now, I have no idea if he was winning down state or not. But if cheating against PT was going on, it would not be needed in our county.
So, his numbers were UP (50.3%), but then down ballot for our Republican R House member is down 3% and under PT’s. Fallacy! Not True! She was up against Malony until the wee hours of the morning…then down about 1000 votes.
There was a libertarian and a whopping 4000 undervotes. What happened that she was suddenly behind after 2 AM?
No one I know would not vote straight ticket for all R candidates.
Yes, I didn’t vote SP on my vote in Kent County. The Shiva analysis is flawed – I’ve looked at it. Take an example precinct with 5 Republican (R) and 5 Democrat (D). If 4R vote straight party (SP) and 1D – you get an X axis of 80%R on his chart. Then for the non-(SP) 1R votes Trump and 4D vote Biden – you get -60% on the Y axis of his chart. However think about it, this would be a 50/50 precinct with a 50/50 vote for Trump, this guy would label this as a strong 80%R precinct with a huge differential in (-60%) the NSP votes. Yet all the R voted Trump and all the D voted Biden.
Why didn’t you vote straight?
It’s out of principle. I like to mark each one because my dad has always said when Michigan moved to straight party option it was a trick from democrats in Detroit to get people to the polls and make it easy for them (just one selection to make). It’s not too much to fill out each one and actually know who you are voting for. In fact, I voted for every Republican, and then did research on the non-partisan section to make sure I was voting for good judges etc.
Note – also, I live in Switzerland now and vote absentee. It is banned in Europe to mail in vote, exceptions are for absentee of those citizens living outside of the country. The mail in vote is the fraud. It set’s up the fraud and was part of the plan. Dems never let a good Covid crisis go to waste. I also don’t vote straight party, not just out of principle, but because I grew up in Kent County. You know more about local people, so there is always a chance that I might vote for a democrat in a local race.
You said this already.
That’s an understatement!
I think this guy has OCD.
I agree in part, first we will find that the reading device formula is adjustable by the people manning the readers. In the example we were looking at, the formula kicks in at 20-30 percent R which I expect means about 2% of the votes being transferred. Honest poll workers would have set it at 100%, i.e. no transfer of votes. Less than 20%R could cause problems with numbers being too low. These people knew what they were doing.
here is what THAT line of assumptions will present problems…let me be clear here
a. the assumptions is that each state used identical and reliable voting systems and regimes AND rules. we know this is not true. thus comparing data output for pattern anomaly and reasonableness is unuseful.
b. what “we don’t know” (yet) is the degree of automation irregularities or if it was deployed globally or only selectively. again presenting a problem where developing comparative baselines will be unuseful.
c. in theory even if the systems and all code is uniform, the reality is each state performs the entire process quite differently AND with disparate human interventions…the last mile people human factor.
4. in a robust audit… sampling for quality thus becomes comparing truly random real physical ballots to automation output. AND looking at actual ballot images versus actual physical ballot cards…they should match. AND comparing of course other hard data that can be verified: registered legal voters versus the received and processed ballots. who many ballots were rejected due to false matching? what was the scale of this mismatch ?(something we have heard nothing about but we would assume occured…and not simply because of a fraud reason..?
5. when this sampling is completed does not compel the argument if a full recount should or should not be done. this is not a normal audit and should not be used to guide a full recount. it creates the legal argument that ANY amount of irregularities located in random tests are the trigger to do so. the question and the answer should be obvious and simple
how much irregularity is acceptable?
how much rodent feces is acceptable in your hotdog facility for human consumption to be regarded “safe”?
the answer:. no irregularity…zero rodent feces.
this is where science …the practical real world application by policy isn’t legitimate and should not be followed.
voting is different… it’s the very WILL of nation.
we should use these science tools yes.. indeed…but not be persuaded that normally acceptable allowances for anomaly and irregular outliers are acceptable.
if I find any rodent feces…any at all…does this mean that is all the feces there really is..?
that’s the crux.
in areas as important as this…the assumptions driving action must be to regard ANY irregularity that cannot be reliably reproduced and attributed to a specific and well understood factor to be THE THRESHOLD LIMIT that drives a complete end to end verification of the system.
full recount…and predictably when that cannot be reasonably demonstrated …we are back to square one. full do over and vote again…this time with ALL verification and cert and authentication firewalls engineered into the system.
the design of our election systems are ridiculously incomplete.
Thank you Frank, a control would be an excellent idea.
Dr. Shiva “assumes the proof.” He assumes that the non-straight-ticket voting should result in numbers that can be predicted by the straight-line vote. He does this early in the video without explaining or justifying or proving the assumption, which is critical to the conclusion he reaches.
To illustrate my criticism, apply Dr. Shiva’s assumption and ensuing analysis to an extreme hypothetical, where a precinct has some straight ticket voters, all Republican and none Democrat; and the rest of the precinct’s voters all split their ticket. Dr. Shiva assumes that because the straight ticket voting shows it is a strongly Republican-inclined precinct, that therefore the non-straight ticket voters should also vote in the same proportions, which in this extreme example would mean if the non-straight ticket voters vote for anything less than 100% Trump there is a vote “switch” from the expected vote. Based on that assumption, Dr. Shiva created a formula and made a graph. The x axis is the percent of straight ticket voters who vote Republican, and the y axis is the percentage of non-straight ticket voters who vote for Trump minus the percent of Republican straight ticket voters. According to Dr. Shiva, if the non-straight ticket voters (who include all of the Independent voters and Democrats who split their ticket) vote only 95% for Trump and 5% Biden, there has been a 5% vote “switch” from the expected 100% Trump numbers toward Biden.
Common sense dictates that Trump would gain substantially fewer than 100% of the non-straight ticket votes because the non-straight ticket voter group includes all of the Independents and registered Democrats who split their tickets.
Another problem with Dr. Shiva’s analysis is that it would require us to believe that heavily Democrat Wayne County, where massive cheating probably occurred, “switched” vote from Biden to Trump, since Trump’s precinct numbers mainly fall above the line on the Wayne County chart.
We should not be giving this analysis credibility because all it does is send us looking in the wrong direction.
However, there are plenty of questions raised across the board about the count. Georgia is on the right track in hand counting the entire state to find out where any problems actually are, and to correct them. .
Dr. Shiva’s analysis clearly shows algorithmic election fraud. It allows one to determine equation that was applied to the incoming data to ensure a biden win. No assumptions were made for the proof.
What you wrote is incomprehensible gobbledy-gook.
The Shiva analysis is flawed – I’ve looked at it. Take an example precinct with 5 Republican (R) and 5 Democrat (D). If 4R vote straight party (SP) and 1D – you get an X axis of 80%R on his chart. Then for the non-(SP) 1R votes Trump and 4D vote Biden – you get -60% on the Y axis of his chart. However think about it, this would be a 50/50 precinct with a 50/50 vote for Trump, this guy would label this as a strong 80%R precinct with a huge differential in (-60%) the NSP votes. Yet all the R voted Trump and all the D voted Biden.
No it isn’t.
When I said “no it isn’t,” I was responding to the Joemama comment that my comment was gobbledygook. I agree with John R. that the Shiva analysis is flawed. Review the Shiva video. See what he puts in for the X axis and the y axis. Then read my comment again.
Even if there are flaws, it’s quite interesting that all the precincts begin to skew at around the same point–20%. Something kicked in; it IS a red flag, just as there was that red flag that caused FOX to call AZ for Biden and all the other states to miraculously and simultaneously stop counting.
Georgia’s hand count then becomes the key to the puzzle if the computer ballot reader is a fraud device. The Wayne County review made perfect sense to me because the graph is based upon per cent, not the number of votes. You would not need the formula because they expected so few Trump votes. One robs were there is the greatest return. It’s when R% increases beyond 20-30%, that the formula kicks in. For Wayne I would expect the fraud to happen by other methods. The reader is one method of many ways to have election fraud. I expect Georgia will prove Dr Shiva correct.
Tell us why the linear transform is coded into the program, as described by the user manual. There’s few conclusions to be had about that.
Everyone here seems to be missing this fact.
Speaking of evidence of election fraud, anybody have an update on this chart of “Bellwether Counties”? IMO stunning prima facie evidence of fraud. I began verifying numbers, all correct, but I had no easy way to look up the numbers and gave up. Also a reverse image search (Tinyeye, google) came up blank. Big surprise. Sooo… anybody got a sauce, or corroboration?
https://twitter.com/SelimSeesYou/status/1325138990580727813/photo/1
It is not an accident that it is nearly impossible to independently verify election number. Massive systematic fraud would become immediately obvious if the data were made available.
But, yes, if the table you posted is accurate, that is another sign that there was something really weird going on.
The national bellwethers must heavily correlate with each other, since they are chosen for all having high correlation to the election. So this is evidence in favor of Trump, but not as strong as it looks because winning one or two of the bellwethers makes it much more likely to win all the rest.
State bellwethers are a different story. It would be very damning if Trump won most of the bellwether counties in Michigan that predict victory in Michigan, and the same in all the other swing states, yet somehow lost those same states in the middle of the night in the Democrat controlled precincts.
I read your reply 3 times. I cannot find any logic in it. Or comprehension of the chart I linked.
Perhaps make your thoughts more clear?
Any two bellwether counties have voted in almost exactly the same way for the previous 10 or 20 elections. This means, in addition to luck (many similar counties are not on the list because they got one election wrong), that the counties strongly resemble each other on systematic factors that drive the vote such as demographics, number of D vs R voters, population density, education and many others. Your county may be near the median on some measures, but the bellwethers are in the sweet spot of perfect mediocrity on a much longer list of metrics.
So they will tend to be right together or wrong together in every election. That most of them switched from right to wrong isn’t an unusual event where large number of coins being tossed and all coming up Heads (under 1 in 1000 chance that this happens for 10 coins). It’s a row of coins whose faces are all together with strings, being flipped as one collective supercoin in which most of them will fall the same way every time. Before the 2020 election the probability might have been, let’s say, 80 percent that whoever wins most of the bellwethers wins the election, but 93 percent that at least 3/4 of the bellwethers vote the same way as each other. So the odds that most of the bellwethers go to the loser of the election are actually pretty high, it is somewhat surprising in that it indicates some shift happened (that much we knew before the votes came in), but not super-surprising in a way that makes it a tell-tale indication of fraud.
On the other hand, if you do the state by state bellwether analysis, that is 50 independent coins being tossed and if the only ones that went against Trump were in the states suspected of fraud, that is a giant smoking gun. This is why I say the state analyses are important.
Thank you for taking the time to expand on that! 🙂
I find this county Bellwether chart valuable because it is *not* a state level comparison, rather a reality check to state results. It is smaller venues which, for the many good reasons you elaborated, strongly tend over time to vote for the national winner. But this year that correlation evaporated. Safe to assume the majority of Bellwether counties did not change fundamentally over the last few years. And if we can trust they remain Bellwethers, and the national results are wildly different, then fraud obviously occurred elsewhere. To me that shows the cheating happened at strategic bottlenecks, either at the state level or in counties with large metros, and not in these widely dispersed and difficult to manipulate Bellwether counties.
That the characteristics of the bellwethers didn’t change much but they failed as (national) predictors does say that something changed this election. But the change could, in theory, be something innocent like COVID or Trump fatigue or massive GOTV efforts by Dems or some other explanation that applies across the country.
The only way to refute the idea of a nationwide effect is to go down to the state or regional level; if “Trump fatigue” or “COVID” fail to make Biden win the state-level bellwether counties then those are invalidated as explanations of the national change and fraud in swing states rises in credibility. So if both analyses point in the same direction they reinforce each other but for the national one alone the Democrats can make up any number of stories to fit the data.
I am sure many other people have thought of this who have access to the state historical data, and hope we will see that analysis soon.
A similar analysis is to not focus specifically on the idea of bellwethers, but track the correlations between different states and counties, and show that the parts of the USA where there is no suspicion of fraud kept their historical patterns of correlation with each other, while the fraud states/counties suddenly dis-correlated with all the places they formerly were in line with.
No. They show its not “normal” starting at about the 45 minute mark.
And it’s important to remember who is responsible for driving the WuFlu fear porn to justify the ‘main in ballot’ (not absentee ballot) scam. https://nexttobagend.blogspot.com/2020/11/face-it-weve-been-had.html
Do you have a clear signature that there is a “floor of at least 20% for Biden”? Your standard of ending all debate is unrealistic. How many debates are ended even in the mildest political climate?
The obvious counter-argument is the Romney theory–“Republicans don’t like him, either”. The presenter tackles this argument and shows a theoretical example of what the data distribution would look like if that were true.
Makes sense logically, but a hypothetical data plot is not dead nuts proof.
The most convincing way to blow that theory out of the water would be to show the ACTUAL data distribution in a select number of counties from other places where vote switching is not in question.
if the curves don’t match up when the hard data is compared side-to-side, then the Romney theory, (like the man himself) is proven to be full of shit.
Well that was pretty conclusive. Massive election fraud in the presidential race.
Now we need to analyze down ballot races, starting with Senatorial and House races.
When those are done, we need to go all the way down to the bottom of the ballot in every state.
I have sent an email to Sidney Powell, which I had originally addressed to PDJT to ask if the entire time series data for the 2020 election be made available to the US public at no cost. It would cost about $1.25 million, given that the WI data cost about $25K. I wrote up a proposal for forming an open source group of volunteers to do election fraud analysis on all the data, in every race, in every state. I haven’t sent that out to people I know who program in Python, yet.
After seeing this video, maybe I need to add this type of analysis to the project. I am currently proposing to do Benford’s law analysis to flag states and precincts which likely have fraud. The benefit of doing this analysis first is that it could provide guidance to PDJT’s legal team as to which places it makes the most sense to push for surveys and hand recounts..
A person on thedonald.win that goes by the handle TrumanBlack has performed an analysis of the presidential race, using state level data garnered from the NYT online database and has produced a table flagging where votes for PDJT were transferred to biden or were outright deleted. Here is a summary of his results:
I call his analysis “vote switch & deletion analysis” (VSDA), just to give it a name. This analysis is nice, because it gives you an estimate of the magnitude of the fraud. If the entire vote data for each state, broken down by precinct were available, VSDA could be performed for every precinct in every state to detect big instances of fraud in each precinct.
Dr. Shiva and team have this wonderful analysis which can detect where automated fraud was performed. This could be applied in every state for the presidential race, but I don’t think the NYT data breaks the data set down into the categories needed. Again, I think the whole data set would be needed to do this valuable analysis. It seems to me that by doing some digital calculus, this type of analysis might also be able to estimate the magnitude of the vote transfers/deletions.
Ultimately, I would like to see all of these kinds of fraud detection done in every election in every state and in every precinct, automated and visible to all for every race in the USA in the future.
But maybe starting with the current election would be worthwhile.
Any Python people out there with time and motivation? I have used Python a little, but I’ll pick it up fast. I have downloaded TrumanBlack’s code, Python 3.2.8 and am starting by trying to reproduce his results.
Be careful – I’m posting this message several times and we are going to lose credibility by pushing the Shiva analysis. This analysis is flawed – I’ve looked at it. Take an example precinct with 5 Republican (R) and 5 Democrat (D). If 4R vote straight party (SP) and 1D – you get an X axis of 80%R on his chart. Then for the non-(SP) 1R votes Trump and 4D vote Biden – you get -60% on the Y axis of his chart. However think about it, this would be a 50/50 precinct with a 50/50 vote for Trump, this guy would label this as a strong 80%R precinct with a huge differential in (-60%) the NSP votes. Yet all the R voted Trump and all the D voted Biden.
I disagree your statement of being careful. Yes, 5 votes for Trump and 5 votes for Biden, but the reader would report something different, we are only dealing with about 2% overall transfers. The machine would take 2 percent (in general, it may be higher as %R increases) of 1 vote and add it to the 5 of Biden for 5.02 votes. That is what is happening with bigger numbers. The Georgia hand count will prove DR Shiva correct.
Wait, wait. I’m not saying there isn’t a problem with Old Dominion software, etc. I’m not saying that the vote changes didn’t happen (maybe they did). What I’m saying is that the Shiva analysis doesn’t say what he thinks it does and is based on flawed assumptions. I’m just illustrating with a real life example of something that would be reasonable and would show up as an outlier based on his assumptions (he would call it proof of an algorithm and then he would take those NSP votes and project them back to Trump). I’m all for hand recounts, I hope that there is also an analysis of the voter forms / because once the ballot goes through is there a way to trace back to who voted. Dr. Shiva could be correct that there is fraud going on with the machines, but that can still make his analysis wrong.
Also by the way. I was posting here on the weekend my suspicions. A recap, I believe that the widespread (Democratic sponsored systemic fraud) could have targeted Republican counties. I see strange results in Kent County like 74% of the new votes over ’16 going to Biden in a Red County. This can be explained by a lot of Trump voters in ’16 coming back and voting for Biden, which I don’t think is likely. But I do think the Detroit, Philly, Atlanta, fraud is more of the historical crap they always pull, just on steroids now because of the mail in vote. They used to need to get warm bodies to the polling stations, now they can just go around neighborhoods and round up ballots from people.
I find your writing very hard to follow. Wouldn’t it be 4/(5+5) = 4/10 = 40% on the x axis?
No – I watched the video a couple of times, what shiva’s X Axis shows is the % voting Republican Straight party versus Democrat straight party. It tells you nothing for example about what percentage of votes in that precinct were straight party versus non straight party. So in my example Trump wins 80% of the Straight Party vote, and that is the X Axis. (4 of 5 vote R) on the other side 1 R votes NSP and 4 Dems NSP but they still choose Trump if they are a Republican and Biden if they are a democrat. This results in Trump winning 1/5 or 20% of the NSP. the difference between 80% and 20% is what Shiva is plotting on the Y axis. So -60% in this hypothetical.
Let me give you another hypothetic example and how it would show in Shiva’s analysis: A precinct with 1000 voters 550 Republican and 450 democrats. Assume all republicans vote Trump and all Dems vote Biden. If 80 R vote SP and 20 D vote Straight party, then Trump wins Straight party vote 80% (x axis). Now NSP – 470 remaining R vote for Trump and 430 remaining D vote for Biden. Trump wins 52% (470/900) 52% is 28% lower than the SP vote so -28% on the Y axis. Its a problem Shiva’s own creation that even when all republicans vote trump and all dems vote biden in a 55% Trump area, his results will make it look like Trump underperformed the non-SP vote. yet the vote still turns out Trump 550 votes for 55%.
I re-listened a number of times and this is what I came up with. I have also sent a message to Dr. Shiva to see if he can clarify.
I think in mathematical formulas. Expressing math in words I find extremely confusing. Here is how I think the math works.
X = RSP/(RSP + DSP) and
Y = RV/(RV + DV) – RSP/(RSP + DSP) = RV/(RV + DV) – X
Where RSP = number of Republican straight party ticket votes in the precinct.
DSP = number of Democrat straight party ticket votes in the precinct.
RV = total number of votes for the Republican candidate in the precinct.
DV = total number of votes for the Democrat candidate in the precinct.
At 20:00 minutes: “Of the people who voted straight party vote, 60% voted republican.”
This means that for the X axis, the demominator of the ratio is the number of people who voted straight party, i.e. 100 votes.
For the Y axis, the demominator of the ratio is the total number of votes in the district, i.e. 1000 votes.
X axis = Republican straight party (RSP) vote percentages = #RSP/(#RSP +#DSP) = 80/(80 + 20) = 8%
Y axis = %Trump individual candidate votes – %RSP votes = 550/(550+450) – 8% = 55% – 8% = 47%
The example you have chosen is not realistic. Dr. Shiva’s analysis uses straight party votes as a proxy for identifying hard core republicans and democrats. When 55% voted R and 45% voted D, while only 8% voted RSP and 2% voted DSP is not realistic. It is tantamount to 90% of the precinct identifying as Independent.
I feel that this analysis is useful, but it would not be possible to use in every state, since many states have eliminated the ability to vote straight party ticket. I believe that the democrats have decided that this puts them at disadvantage for down ballot races, so they got rid of it.
Ultimately, if things go like that have for years and years, the democrats will eventually eliminate it everywhere. That’s all they do: they live like leaches off of tax dollars and scheme constantly on how to usurp all power and they seem to always eventually win.
Listen around 19:50. He describes the Y Axis as the % who vote RSP – the % who vote Trump split party (not to total % in precinct who vote Trump). Which is as I described above.
sorry jeez now I’m inverting it (he describes it as % who vote Trump split party – RSP%. As I describe above.
John I think your missing something …. The Biden straight party vote is already accounted for in Shiva’s horizontal red line … that line covers the total straight party vote, including the democrat straight party partisans when you see 15% Republican for the percent of straight party voting, remember that also indicates the 85% democrat portion of the straight party vote … the only thing causing the variation are the leftovers … the ticket splitters. … this percentage pretty consistently clusters around 50% no matter how many votes are cast … and so each candidate should have a consistent share (as a percentage) of those leftovers … with normal variations from precinct to precinct.
And that’s exactly what Shiva’s graph shows … up to the point where 25% of straight ticket voters are Republican, then the graph takes a sharp turn downward and continues to -30% as the percentage of straight party Republican votes increase … It should not … it should continue straight across … because we’re just looking at percentages of leftover votes … not the actual number of votes.
As a check, I did the same graphs for Macomb and Oakland counties for the returns from 2016 (the only other election where straight ticket votes were reported). I expected to see the regular horizontal distribution for total vote percentage in relation to straight ticket voting but was surprised to see a distribution similar to the 2020 data …. except …. the bend started at 30% on the x-axis maxed out at -15% on the y-axis. This indicates Clinton cheated too, just not as much as Biden … Recall that she lost Michigan in 2016, and the Bidens weren’t about to let that happen again so they upped the numbers.
Not missing something – I think you meant to say Democrat straight party vote, but yes that is included in the X axis so that when on left at say 20% Republican straight party (Trump vote) that means that 80% of total straight party votes when to Biden. When you get ot 80% x axis that means 20% went to Biden. His math isn’t wrong that’s not the problem. I did much the same as you in Kent County. You will get the same plots as he did. The issue is that this line down left to right is what you would expect to see because of what he does with the Y axis. I think some people thought he was using the total percentage of Trump votes when calculating the difference, he’s not – he is using the direct Trump vote percentage versus biden (not including straight party votes from the x axis). Most precincts has around 1,000 to 2,000 people, so first it will depend on the ration of straight party to split party tickets. If only 10% of people vote straight party, and you compare this to the remaining 90% of the vote with an assumption that you should get a similar result plus or minus 5% based on whether Trump is liked or not doesn’t make sense. Now I don’t see in the data 10% straight party. In Kent county I see Grattan Township p1 at 55% and Ada Twp 5 at 46%. So it seems to range around 50% in Kent county. which is why my illustrative example is making the point that even in a 50 / 50 county where every republican votes for Trump and democrat for biden with Shiva’s formula you get the linear line down, because it varies based on the proportion of Trump in the straight party votes. It is natural that when Trump get’s 80% of straight party vote that there will be a large negative difference, because for Trump to win 80% of the split ticket would be nearly impossible as in includes independents and more democrats because we are assuming a 50/50 county. In other words take a county in an election you have confidence in, run the same data for the democrat or republican and you will get the same result as Shiva. In fact Biden’s graph if you did the same would look exactly the same. Again I grew up in West Michigan and I don’t trust the detroit counties, I think there is fraud in those areas, but now it is on steroids because of unleashing the mail in ballots.
https://theconservativetreehouse.com/2020/11/12/mit-phd-and-statistician-outlines-algorithmic-fingerprint-within-vote-data/comment-page-2/#comment-9144743
DHS Assistant Director of Cyber security just forced to resign.
https://www.cisa.gov/bryan-s-ware
The tell tale grin.
https://www.govconwire.com/2020/11/dhs-cyber-official-bryan-ware-to-leave-post-for-private-sector/
Once the Trump-Biden race is finally resolved (if ever), citizen patriots should probably look for earlier instances where vote-switching could have occurred.
Obviously perpetrators they had to test out the vote-switching beforehand, probably in state races and earlier elections.
Somebody mentioned the Doug Jones victory after Session’s appointment.
Later, many odd 2018 Blue Wave victories, such as in AZ, CA, WI and MN.
Computer fraud as ballots tabulate.
At the end of Shiva’s podcast, they suggest votes are valuable like money. And consequently need protection and regular audits. Maybe we need CPAs to peruse election returns!
1. I think President Trump Knows. Let’s focus on a recent tweet.
“DATA ANALYSIS FINDS 221,000 PENNSYLVANIA VOTES SWITCHED FROM PRESIDENT TRUMP TO BIDEN. 941,000 TRUMP VOTES DELETED.”
*https://twitter.com/realDonaldTrump/status/1326926226888544256
221,000 SWITCHED
941,000 DELETED
2. I think that analysis can be found here:
https://thedonald.win/p/11Q8O2wesk/happening-calling-every-pede-to-/
https://noqreport.com/2020/11/11/data-deep-dive-on-dominion-voting-systems-offers-incontrovertible-proof-of-election-hack/
3. Here is an excerpt from the analysis: (lost=deleted)
“Pennsylvania: Switched: 220,883 Lost Votes: 941,248”
220,883 SWITCHED
941,248 LOST VOTES
*The second link is a duplicate of the first, it’s an easier read.
As I watched the good professor for a second time I wondered whether he had done the same analysis for the democrat vote. I’ve concluded he has, and that analysis should eliminate any doubt whatsoever about his conclusion. I hope his life insurance is current, his tenure secure and he has an ex-SEAL as a best friend.
I’m sharing Prof. Shiva’s video with as many as I can, while I can.
We need someone else to look at this data. At the end, they come up with the “Workers of the World Unite” slide and that alone shines a bad light on everything else. Also the suspicion that Republicans were in on the Coup attempt. Say what? How could they come to that suspicion?
Something is fishy
Yes, the mere thought that both parties united against Trump to steal the election is pure BALDERDASH, my good man. How could that be? Why, they would have to be a Uniparty, a swamp of some sort, an entire City or indeed a global cabal aligned for its own interests, against the very people of this nation.
Or, maybe its just corporate America protecting its investment in China.
Preposterous! You have convinced me. Good day, sir.
You have obviously never heard of the UNIParty.
Dr Shiva did show a chart of a Biden area as an example of a NON manipulated count. He then shows the Trump areas and the chart is radically different. That is the big point they were making. I watched all of the video.
LIARS!
I mean the Democrats, of course
I ran the same analysis using Dr. Shiva’s algorithm but I used the data from Antrim County, MI. This is the county that was hand counted after the “glitch” was discovered, so it should have been more or less flat according to Dr. Shiva’s analysis. I indeed found that it was flat and did not show the “red flag” curve that the good doctor found in counties like Wayne. I can email you the data / graph if you like.
Larry
I had a look at that too … problem is there’s only 15 precincts reporting there … that’s a lot of soup from one oyster 🙂 … But what data is there is pretty random (though it is below the red line, not above it) …
These men are our American Cream.
Thank you, gentlemen!
So much concern trolling on this article. So many expert statisticians on this site to tell us how and why this analysis is wrong. Why, it is almost as if the democrats had paid operatives coming onto websites to discourage and confuse republican voters as a part of their massive fraud effort to steal the election.
Indeed.
2,066,000 registered voters in Maricopa County, AZ and the New York Times count had 2,035,000 of them voting (99+%), with 2% to go? The odds of that being real?
Democrat County Recorder and Democrat Secretary of State, and the Republican responsible for certifying the Dominion machines skipped the inspection. This is where McCain and Flake have done the most damage.
Meanwhile, in the state’s second most populous, heavily Democratic Pima county, only 509,000 of the 602,000 (80%) voters cast votes. Nothing to see here, move along.
https://recorder.maricopa.gov/voterregistration/redirect_new.aspx?view=city
It would be interesting for the people that say the analysis is flawed to email their reasons to Dr. Shiva and get his response. If he responds please post. Thanks
It doesn’t require a PhD to recognize the tilted scale whether it is from criminal deep state professionals in law enforcement, intelligence etc. or logic defying election numbers. If it looks like a duck, quacks like a duck, it’s a duck. The election looting is widespread, do we ever round up the looters and identify each one and exactly what they stole. Not that I recall.
I don’t count votes, but I have conducted many, many inventories. This is such a simple process, no reason why it has to be made so complicated with black box software. Basically, there are a number of very simple cross checks of independent sources. Voters are validated against the voter rolls. The number of voters in a precinct must = good ballots + bad ballots+provisional ballots. The total number of ballots issued = ballots cast by voters+spoiled ballots. If there are “lost” ballots (issued but not returned) we should know how many and which ones they are. Every ballot is serialized and individually trackable in the database, so we can verify what was tabulated against the actual source document. We can also identify duplicates and fakes with phony serial numbers. And tabulating is as easy as counting up 1+1. I cannot think of any reason to build in a feature such as a “weighted election.”
And yes, we need a fully human readable audit trail. Hand-marked paper ballots are a must. Once you get to the point where you’re just manipulating numbers in a spreadsheet or database the data can become corrupted easily and you get lost very quickly, with no means to find your way back.
I would like to see an analysis of Biden individual vs. straight ticket vote . I’m trying to think of a reason why this very odd pattern appears, it looks too perfect but without seeing the other side I can’t say if it’s a natural feature of the dataset or manipulation. If, for example, we see no pattern or the OPPOSITE then that’s a pretty much ironclad confirmation to me there are shenanigans afoot.
An important point of the analysis is that the slope (decline in Trump votes) was nearly identical in the counties used. The chances of this happening naturally are near zero. I, too, would like to see this theory tested by a hand recount, and it will be. Lin Wood seems to have his finger on the pulse of the magnitude of the fraud and he says (via tweets) that we should all be patient. Hard to do given what we have witnessed. At the end of the day Biden’s huge, late breaking numbers in PA, WI, and MI defy common sense, the smell test, or however else you want to describe it, particularly when POTUS crushed it in Ohio. Once more fraud is revealed, more conclusively, the tide should turn.
The chance of a single occurrence is far fetched, multiple occurrences… enemy action.
Here is Gallup’s Donald Trump National Job Approval by Party Identification among REPUBLICANS Oct 16-27, the latest poll on Gallup’s web-site: Republicans 95%, Independents 41% and Democrats 3%. I cannot find a Republican approval poll for Michigan only. I would believe it would be in line with the National poll results. That approval rating would reinforce Dr.Shiva’s and this team’s analysis that it makes NO SENSE that the more Republican a precinct is, the less they like Trump, and the less Republican the more they like Trump.
It’s many moons ago that I studied statistics, but these guys would seem to be on the right track to look at straight-party versus non-straight-party votes. The logical place to switch votes would be in a Republican area without affecting down-ballot candidates, who would be more likely to demand a recount. Easy then to use the Mitt Romney explanation for seemingly odd results. They want it to appear as if Republicans repudiated Trump.
Thing is, Shiva covered the “Mittens” explanation too … If it were true, it would just show a horizontal scatter below the red line.
Here are themost recent (Oct 16-27) Gallup Poll Results for Donald Trump Job Approval by Party Identification for the: Republicans; 95%, Independents; 41% and Democrats; 3%. I cannot find a Party indentification poll for Michigan only but would believe it is in line w/the National poll. The National approval poll reinforces Dr.Shiva’s analysis that the data reveals irrefutable evidence that the Voting System is moving votes from President Trump to Biden. It makes NO SENSE that the more Republican a precinct, the LESS they like Trump. The voting system is 180 degrees out from the numberous Gallup Party indentification polls taken since May, 2020 when Trump was at 85% Republican Approval to today wjere he is at 95%.
How many state did Biden win where the Dominion voting machines / software wasn’t used? How many states did Biden lose that Dominion wasn’t used?
Can someone put together a chart / graph on that?
Dominion is used in 28 states. The Dominion website highlights the 12 states where they are most proud of having the business: https://www.dominionvoting.com/about/
Of those 12 states, Biden is winning 10:
California Biden
Nevada Biden
Utah Trump
Arizona Biden
Colorado Biden
New Mexico Biden
Alaska Trump
Illinois Biden
Michigan Biden
Georgia Biden
New York Biden
New Jersey Biden
“compelling” is an understatement, it’s astounding. It’s alarming.
One could provide supporting evidence for the claims by canvasing the identified counties to confirm how republicans actually voted.
It seems there are some people in this thread missing the point. It is not a matter of demonstrating expected or unexpected results, but a matter of demonstrating trends that are too perfectly sloped to have occurred naturally, regardless of expectation. Hence the term “fingerprint” used by SD. Even if you think Shiva’s data shows expected results for Trump (which it doesn’t IMHO), it does so too perfectly. That is the point.
Thank you. THAT is exactly the point.
Excellent and frightening presentation.
When I voted here in NC, the voting machine churned out what appeared to me to be a blank piece of paper. I had no idea whether what I’d seen on the screen had been transferred to that paper. And further, when that paper was fed into a second machine, I have no idea how the “marks” on it were interpreted/manipulated.
My takeaway? Turn the lights out. The country is done with.
My main takeaway is that the raw data shows declines in Trump votes. There are videos of news coverage that shows the feed numbers actually dropping for Trump. No PHD needed to ask the simple question. Why? How is that possible?
A drop in count could be voters that voted absentee and in person. Strike those.
A poll worker ran ballots through twice, system adjusted.
Votes came in that were invalid, delete those after the system checked.
Likely more reasons that need to be explained.
Or, most importantly, the system MAY have a weighting function that was working as it was supposed to. These weighting functions do exist in voting software. The number 1 question should be, “Does the function exist on the systems used where we see the anomaly?” If it exists, was it turned on? If it was turned on, why, and who?
The data and charts shown by Dr Shiva could be evidence of a weighting. They simply asked the question. And as we all know, statistics will show that if your head is in a bucket of ice water, and your ass is on fire, you should feel fine!
I’m not talking about hackers or some internet transfer problem or nefarious software code. I’m talking about a simple setting by anyone with access to the software.
If in fact a weighting function is available in Dominion software, it was known when it was evaluated and purchased. This could be sold as a feature that allows the software to be used for many types of voting and the evaluators saw that as a bonus. The fraudsters would see that as an opportunity.
What Shiva’s plotting shows is that the more Republican-leading the precinct, the more likely that split ticket voters voted for Biden. But ONLY in precincts with a threshold minimum of 23-24% of “Republican-leaningness”. Further, that relationship is roughly linear with a slope of -.5. More Democrat-leaning precincts (23-24% Republican or less) had split ticket voters giving Trump votes at roughly 3-10% above what Trump garnered from Republican Straight Ticket voters.
What is special about the inflection point of 23-24% Republican-leaning, when Trump votes suddenly leak into Biden votes?
It’s counter-intuitive that Trump would bleed votes in a perfect negative linear proportion to the precinct’s Republican identity.
Also, the same pattern is repeated in three MI counties.
We need to see what all MI counties look like – especially the Detroit precincts.
I think the Trump layers are concentrating their efforts on Wayne county. Perhaps the computer fraud group should include those other counties. By the way, it looks like GA may have had the algorithm turned on also. These analytical chaps may want to look there also.
That would be Trump lawyers.
iirc, the researchers said the slope was as expected in Wayne Co., in which case maybe the polling place problems in that county were head fakes–deliberate attempts to invite allegations to serve as distractions.
We don’t know how audits or recounts will turn out, if they happen, but suppose they purport to show that the Republicans were crying wolf. Maybe it’s all designed to distract from the actual computer fraud perpetrated in other areas. iow, it’s a trap to discredit all allegations of fraud. https://www.mlive.com/politics/2020/11/new-lawsuit-alleges-detroit-wayne-county-manipulated-election-results.html
They seem to already have an “explanation” for everything: people just misunderstood what they were seeing. Just speculating. (And they’re “racist,” too.)
We’re dealing with deep state spooks who know how to deceive and misdirect. Evil.
Shiva charts Trump votes that are sourced from Split Party ballots and compared to Trump votes sourced from R Straight Party ballots.
There are 2 types of ballots in MI: Straight Party and “Independent” (which is easier to think of as “Split Ticket”).
Of course Democrats registered as Dems have the same options. He is NOT considering Democrat straight party tickets, except when discussing how “AP reports”.
What he plots shows (for each precinct) how many votes for Trump came from split tickets (all regions above the x axis) as a percentage of all R votes and how many votes were cast against Trump (thus for Biden) came from split tickets (all regions below the x-axis). Basically it’s a scatter chart to show deviation towards Trump or away from Trump compared to the straight party ticket support for R party in each precinct.
thx for your explanations Apfelcobbler, I’ll come back and re-read and re-read them until I understand them fully
… like running the same stack of Biden ballots thru the machines 2 or 3 times
I’ve read a few of John R’s comments and I read the medium.com critique by Naim Kabir
I don’t currently have the mental horsepower to sort thru this (retired 9 years from my scientific career and brain gone somewhat to mush)
… but I do recollect cautions about using arithmetic differences and ratios (like percentages) in regression analysis
There’s a thing called self-correlation or spurious self-correlation that can result from plotting a variable x versus a difference that also contains the term x, like y-x
I can’t figure if that’s at play here, but maybe a couple of you who have gone back and dissected exactly what the variables are that Shiva is potting can comment
I think I understand that he’s plotting percentages against differences in percentages or something, but I can’t quite bear down mentally at the moment
Here’s a reference, more can be found by keyword searching
Beware of Spurious Self-Correlations!
https://www.deepdyve.com/lp/wiley/beware-of-spurious-self-correlations-vYaSi80hpt
Spurious self-correlations arise when two parameters (sums, differences, ratios, products, or single variables) that are used in a linear regression analysis have a common term.
I could be wrong, but the analysis for each of the precincts was very straightforward and I don’t think linear regression plays into this. Basically, I think where the critics are going is to say the flaw in the analysis is amplified which is why he gets a straight line downward trend. I don’t think the analysis is flawed or amplified. But, my brain is also fuzzywuzzy.
The way I understood it was much simpler. The voting system software has a “feature” that can be turned on that allows for the system to count each vote as a fraction, so pretend the system counts 1 vote as 1.2 or 1.7. Those additional fractions add up pretty quickly to additional whole votes. In addition to that harmful function, the system was programmed to shift votes from one candidate to another when a certain threshold was reached (according to some eyewitnesses who are calling it a “glitch” – computers don’t glitch, they follow instructions). Example, take 6000 votes away from Trump, and give them to Biden. The combination of these two things could create that artificial straight line downward.
thanks for your explanation of the fractional apportionment, gramma, yes – I’m open to the likelihood that chicanery was carried out by this software
as far as linear regression, no – I wasn’t contending that Shiva and his associates carried out linear regression analyses
what I was pointing out is this phenomenon of self-correlation when a variable is plotted against an algebraic function of itself, like a sum, difference, ratio, or percentage
because the variable ends up being represented on both axes, results will tend to show a slope, like the slope Shiva is making much about
that slope will be quite amenable to regression analysis, but regression need not be run, necessarily …. one can make a case just pointing to the evident slope
but in this topic of spurious self-correlation, the slope is an artifact of choosing or defining variables inadviseably
plot (y + x) versus x
with increasing x, (y + x) will also increase … perhaps because of y, but certainly and trivially because of x, because it’s also in the summed term
larger values of x will also produce larger values of (y + x)
it’ll have a slope
positive for a sum
and I think negative for a difference (y-x)
there’ll be different effects whether the 2nd variable in the parenthesis is a sum, difference, ratio, percentage, etc
the key point is that x is getting represented on both axes … the self-correlation of x with itself will impart a slope of some sort … a line, a curve, something
that”s why I’m suspect about plotting percentages against a difference in percentages
for one thing, percentages are ratios
for another thing, depending how they’re constituted, there may be representation of some variable on both axes
I’ve got other things to do and can’t devote a half day to sorting Shiva’s variables out to my confidence
…. so I was hoping someone who’s already carefully dissected his variables would see my raising this self-correlation aspect and comment further
like “no, that doesn’t appear to be confounding things here because his x variable is yada yad and his y variable is yoda yoda
thanks again
and to be clear, I did watch the entire video the other night, I was the first to comment on it on the youtube page
and I was nicely convinced that his contentions were reasonable and complimented him and his team on their analysis
but then today, I browsed over John R’s’ comments – didn’t concentrate hard enough to fully understand, just get the gist and noted others disagreeing
… and I browsed the medium.com critique by that Naim guy, calling Silva’s analysis a “parlor tridk”
And caused me to remember this bit about self-correlation from my career days and was able to find the article on spurious self-correlation that I remembered, and that it comes about from plotting a variable against some algebraic function that also contains that same variable
but again, I don’t have the mental horsepower to sort thru it without devoting a day or half day to it
also, as part of the “fractional” weights / votes
I commented at the youtube page but that comment got pushed way down the page by newer comments – search for “dimbulb” to find it
my comment was that decimal fractions should have absolutely no place in vote-tabulating software
voters and ballots are inherently integer
counting and storing operations can be integer operations on integer-defined variables in the software
the only computational operation should be addition
with results stored as counts or sums
and stored as integer
I also questioned why a software writer would do such a thing as define variables as decimal fraction variables
One can only expect mischief to result
And I questioned whether software “engineers” have the same ethics expectations as other engineering disciplines (the kind who build bridges that can fall down, for example) and asked where the hell is the ethics board for this software
Today I read that the decimal business may have arisen from a special case of a Sacramento election
That may be so, but the same software shouldn’t be used for other elections
They could disable a subroutine or two and issue clean, wholly-integer software for all other “normal” elections where we don’t have decimal fraction voters and “fractional” votes
alright, rant off
@Apfelcobbler
Exactly! And there is no plausible explanation why the relation is changing when getting into more republican precincts. There are to possible explanations:
1. Trump looses support in more republican leaning areas – all of the sudden, at a a threashhold of 20% and from that forth deeper and deeper the more he is on his home turf. Again: He must be underperforming heavily with republicans in a proportional way (but only beginning at a certain threashhold), republicans love him, when they have democrats as neighbours, hate him if they don’t have democrats as neighbours, independens and democrats love him. (unlikley nonsens, trump underperforming in rural precincts? )
2. Votes from Trump are converted into votes for Biden, but only in precincts with at least 20% republican votes, so that it doesn’ t look to obvious. (likley)
I watched the whole thing and I actually got an A in Advanced Statistics at the doctoral level, but it took me 2 tries… His analysis made perfect sense to me. The deviation is not natural and I’d venture to say that it happens 0% of the time, but I think one of the presenters said 1% to avoid saying 0%. COMMON SENSE tells us that Trump won in an epic landslide across the country with the exception of a tiny handful of states. The enthusiasm for the People’s President has been off the charts. So using our common sense, WE KNOW THE ELECTION WAS RIGGED, and don’t really need guys from MIT to tell us that.
What was very interesting was “questions for future discussions”: Were/are both Democrats AND Republicans in on it? I say why yes, yes they are.
You can see my above many comments, my total agreement with Dr Shiva. I agree the election was rigged but I disagree that republicans were in on it. They in general, don’t understand what is going on. The ballot counter is a magic box to everyone. Look at how we on this thread are having problems understanding what happened. It took you to two tries. Which Republican would have the fight and ability to understand. After all is only about 2% in general of votes being transferred. The people involved knew what they were doing, no transfers below about 20%R, then harvest the rich number of R votes. I bet that below 20% R, you would get crazy numbers that would make it obvious what was going on.
It stands to reason that if you were going to flip an election to a Democrat with an algorithm, the best way to do it without making it too obvious would be to go to high population counties that skew Republican. That’s where you can steal the most votes on an absolute basis with the highest liklihood of not being detected. If they tried this in Dem skewing precincts in low population counties they’d have to (out of necessity) steal almost EVERY Trump vote which would make it so obvious that it wouldn’t take an MIT genius to figure it out. We are lucky Dr. Shiva DID figure it out and we’d better hope the judges and legislators figure it out as well.
good point, Nick
So the question is whether hand recounts would reveal that the machines are changing the tallies?
Is there a paper trail?
If they truly have the paper ballots in hand to count …
… then a hand count would catch the kind of fraud where they take a 200 stack of Biden ballots and run it thru the machine 5 times to make a thousand
… the total counts won’t add up
… 1,000 tally but only 200 in hand
This is the kind of thing that tripped up the Michigan recount for Jill Stein in 2016 before they could even get started
They started to open sealed boxes that had the total ballots tally and the candidate tallies written on the seal
First box or two they opened, there were hundreds more ballots in the box than what was written on the outside
So they couldn’t legally recount because some mischief / fraud was already revealed simply from the total ballot tallies not agreeing
So the recount was over before it could even get underway, at least in that precinct … can’t recall how widespread that dysfunctionality was
Don’t know the overall answer to your question, but ballots are going to have to agree with tallies, tally sheets, machine printouts, etc
I’d expect trouble right off the bat, just like that Michigan example
We need this analysis for many other counties.
Why on earth is there such a thing as a “weighted feature” built into a vote counting machine?
Ok, I did go back and search for his x and y variable and it didn’t take me half a day, it took about a minute to locate the table in the video and another couple minutes to do a screen cap and crop it … oh, and I see other people upthread discussing the variables but this graphic should also aid
X-axis: Republican Straight-Party *RSP” vote percentages (%)
Y-axis: Difference of % Trump Individual Candidate votes minus %RSP votes
Abbreviating and de-cluttering, we get
X-axis: %RSP
Y-axis: (%TIC – %RSP)
And so %RSP is represented on BOTH axes, as I was rambling on about upthread
… %RSP is plotted on the x-axis
… AND it’s represented on the y-axis in the algebraic difference term
As x increases along the x-axis, larger values of x will produce smaller and smaller values in the difference term … or more strongly negative, considering both the positive and negative extent of the y-axis
Hence the negative slope
Basically, x is correlating with itself, but in a negative slope way due to the difference
I’m no expert, but I’m afraid Dr Shiva may have set himself up a nice case of spurious self-correlation
… by having %RSP on both the x-axis and as part of the difference term on the Y-axis
I’ll have to re-read that Naim “parlor trick” article and see if what he was going on about is akin to what I’m saying here
I was a rudimentary investigator in water-resources studies but I did remember this one important paper about spurious self-correlation from my career
So I’d appreciate any of our higher-powered math or stats guys chiming in here
Anyone?
… correction:
Basically, x is correlating [ and trending ] with itself …
… part of that trend or slope may be due to a real reationship with the other variable, but some degree of it will be attributable to the self-correlation
That’s as much as I can conjecture from my fairly feeble expertise
Would really like to see an expert here tear into this
oh, and here again is a reference I posted upthread, dragging it down here for proximity
Beware of Spurious Self-Correlations!
https://www.deepdyve.com/lp/wiley/beware-of-spurious-self-correlations-vYaSi80hpt
“Spurious self-correlations arise when two parameters (sums, differences, ratios, products, or single variables) that are used in a linear regression analysis have a common term.”
And I’d clarify that even if no “linear regression analysis” is being done in Shiva’s presentation, the same relationship in variables will produce a scatterplot with evident slope that a regression analysis would operate on
hey Nimrodman – yes, this is correct. I think that you can see it best in the analysis of Biden and Straight dem tickets. I did this with the Kent County data – you can get it online for all precincts in the county. I get the same scatter chart as Shiva. What is interesting is that since I have all the data for Biden as well, I created an X-Axis for Democrat (Biden) straight party votes, and a Y Axis for Biden Split Party votes. Same data same counties. The line is exactly the same for Biden starts top left and drops to bottom right. This makes sense right because of what you not above it’s self correlated. Don’t know how to post pictures here but have both scatter plots. So what is clear at a minimum is his conclusion based on this analysis does not hold because you could say the same thing for Biden i.e., the downward sloping line indicates an algorithm was used to shift Biden votes to Trump. It’s like a parallel universe. I don’t know about all the comments angry with me for pointing this out, or people saying they also have PHD in stats and he’s right. The only thing he is right about is that the data is plotted correctly as he said, but his conclusion on what it means is wrong. Again, I believe there is fraud in this election and that there even may be votes being shifted by the Dominion voting machines, but this graph doesn’t prove it.
Thanks, John
Yeah, I’ll have to go back and listen to Shiva’s presentation, but at this point it’s not clear to me how these scatterplots and trend lines are diagnostic that “vote shifting” has occurred.
I’m viewing the scatterplots in a strictly algebraic sense, based on how he’s defined his x and y variables:
X-axis: %RSP
Y-axis: (%TIC – %RSP)
And – reminder – %RSP is represented on BOTH axes … explicitly on the x-axis and in the difference term on the y-axis
For increasingly large values of %RSP, they y-axis values get increasingly large %RSP comes to dominate the plotted values
… large %RSP on the x-axis
… large difference term on the y-axis, and negative because large %RSP is subtracted
For largest values of %RSP, %RSP is swamping out and dominating the relationship
%RSP is trending with itself, basically
That’s the “spurious self-correlation” spoken about in the literature
… and to me, it looks like Shiva has built it into his scatterplots as a result of how he defined his variables
And a follow-on conjecture, I wrote this last night but held off posting:
—–
Another glimmer of a thought, extracting from above:
I said:
Basically, x is correlating [ and trending ] with itself …
… part of that trend or slope may be due to a real reationship with the other variable, but some degree of it will be attributable to the self-correlation
It now occurs to me to speculate that perhaps this is what’s happening in that 20% or so threshold area that we’ve puzzled about
I’d conjecture that the early left-most part of those scatterplots may represent the true relation between %RSP and %TIC
… for small values of %RSP in the difference term (%TIC – %RSP)
… and that there isn’t much of a relationship, as evident from the scatter but with little apparent trend
Now, as we move to the right, we’re getting larger and larger values of %RSP
… and perhaps greater self-correlation between
%RSP and (%TIC – %RSP)
Remember, these are mostly-Republican precincts, right?
… so %RSP is the reservoir with larger magnitude values – I think
So, moving to the right, at some point the increasingly larger %RSP values start to outstrip the influence of the %TIC values in the difference term on y-axis
… and so %RSP starts self-correlating and trending with itself
… call it the “hinge” in the line
Thereafter, to the right, larger and larger %RSP values self-correlate with more strongly negative values of the difference term (%TIC – %RSP)
You could test some of this intuitively by simply inspecting the values
Sort the dataset by the difference term (%TIC – %RSP)
And compare the %TIC values with the %RSP values for largest values of (%TIC – %RSP)
I think you’ll see that (%TIC – %RSP) values in that upper range have very little to do with %TIC and everything to do with %RSP
That’s self-correlation and it has everything to do with including %RSP on both axes
Shortcut for you – the two ends of the line for Trump and a Biden graph are two precincts: In Kent County the outlier precincts are Grand Rapids City Ward 3 Precinct 70. Straight Party Rep (Trump) 48 Dem (Biden) 671, Split Party Trump 26, Biden 139. This gives Trump scatterplot a point at X axis 7%, Y Axis 9% and the same graph for Biden gives Xaxis 93% YAxis -9%. This is a heavy Dem area, and Biden underperforms his straight party vote by -9%.
The other outlier precinct (other end of the lines) is from Byron Township Precinct 4 with Straight Party Rep(Trump) 1,003 Straight Party Dem 169, and Split Ticket Trump 460 versus Split Ticket Biden 265. This gives Trump scatterplot a point at X axis 86% Yaxis -22% and the same graph for Biden X axis 14%, Y Axis 22%.
Plot these two points on a graph for Trump, you get downward sloping line left to right. Do same for Biden and you get the same. So Shiva’s conclusion would be on Trumps Graph – Trump votes were shifted to Biden, and on Biden’s graph Biden votes shifted to Trump. There you have it – I’m giving you all you need to check for yourself – if you can’t do it, that’s on you.
1. They flip votes in heavy Republican counties, that because that’s where the votes are for them to harvest.
2. This puts the fraud in the areas you would not expect to see it. When you look for fraud and Democrat strongholds there’s nothing to see; while the fraud is in the areas your least likely do expect.
3. If you do discover the fraud, you’re stuck fighting about frauds and areas that you ostensibly control. So, well you’re pointing out the election is fraudulent, they’re blaming it on you. This is their area they own this. Look what they’re trying to do, they’re manufacturing votes in the areas where they control. Etc.
This is downright diabolical, and in a close election it probably would have worked. The only reason it’s being exposed is the victory margin was too great and they had to employ other measures and Democrats strongholds where we were watching for that. Absent the “emergency Democrat ballots” being brought in to Democrat strongholds during the dead of night, this probably would have slipped past.
As a physicist who uses statistics all the time, I’m unimpressed by this analysis, which, among other things, lacks comparisons to demonstrate its correctness. The link below has a nice summary of deficiencies
https://kabir-naim.medium.com/dr-shiva-ayyadurai-the-danger-of-data-charlatans-4f675ffe793c
The analysis is flawed, they performed a mathematical operation on percentages with no regard as to what the relative sample sizes are, you can’t do that, and the result is a line with a negative slope. If you plot the same data for Dementia Joe and the communists you get a very similar result. While I do believe there was massive election fraud this isn’t it.
Dr. Shiva metrics
x = %SP
c = %Trump
y = %Trump – %SP = c – x
for each precinct
Establishes a line tendency as it sould be, y = – x + c. If y = 0 x = c, if x = 0 y = c**, the slope is always the same (45º) and crosses x-axis and y-axis on c.
No relationship, a priori, between %SP and %Trump. Not calculated over total republican votes or republican+democrat votes – first the data afterwards the hipothesys -.
Changing the metrics:
x = %SP
y = %Trump = y-shiva + %SP = y-shiva + x = c
New line tendency looks like without slope. You have just to, visually, sum up %SP to y-shiva to see the parallel line to x-axis with a constant value of c**. This accomplishes Dr. Shiva hipothesys (line above or below y-shiva = 0% but, always, parallel to x-axis)
Macomb county according to Dr. Shiva metrics:
Early voting: Tendency line crosses x-axis at 40% of %SP. Implies c (%Trump – %SP) in y-axis or in x-axis about 40%**, I mean, a straight line, in new metrics, of y = %Trump parallel to x-axis with a value of 40%. This value, 40%, is the absolute (not decreased by %SP) direct vote to Trump and should be the same along de %SP (x-axis), as it is. So, %Trump does not depend on %SP.
Election day: Tendency line crosses x-axis at 75% of %SP. Almost duplicates the c before (40%). This value, 75%, is, again, the absolute (not decreased by %SP) direct vote to Trump
Republican present voters choose Trump almost the double (> 75%) to those who did by post-mail (40%). We talk about a specific behaviour inside the same party, inside the same type of voter. It´s difficult to believe. Even though post-mail vote was acreditative (with permanent identification), wich I don´t know, what would make republicans to vote SP instead to Trump, the difference do not justify such a gap (unless republicans vote for SP being afraid than their vote would be launched to the bin or being signaled as a Trump voter).
Conclusion:
– %Trump over %SP does not vary from what should be expected (no correlation)
– Dr. Shiva metrics obliges the line slope to be, always, 45º.
– We find the real and absolute %Trump direct-vote at x = 0, wich implies y = – 0 + c = c = %Trump
– In Macomb county, republicans, early voted round 40% to Trump independently of %SP in each precinct.
– In Macomb county, republicans, election day voted round 75% to Trump independently of %SP in each precinct.
– There are not strong hypothesys for this duplicated statistically behaviour.
– USA is adopting “new ways of democracy” coming from the old Europe frigtening, in the streets, the lovely american republican grandparents.
So the transfer from one candidate to the other. in Macomb county, would start changing the metric, at 40% of %SP in early voting and at 75% of %SP in election day because at these points %Trump equals %SP. I think, Dr Shiva has showed the better graph possible to demonstrate the shifting between candidates. Two points of discordance, early voting to election day, and Dr Shiva steady line for Trump direct vote fact as there should be no dependence.. The greater the %SP, the same %Trump vote.