eric76 said:

---

The Debt said:

---

TravelAg2004 said:

---

Here's a quick webpage I put together showing 4 data sets:

• Oakland (President)
• Oakland (Senator)
• Washtenaw (President)
• Washtenaw (Senator)

https://travelag2004.github.io/michigan-votes/

You can turn each one on and off to see how they interact.

When you put them on the same time scale, I'm not seeing anything as dramatic as what they are showing. Not saying something isn't going on here, but it's not nearly as cut and dry as it first appeared. I've made it pretty easy to add more data, so I'll add a few more counties later on.

---

Not to diminish your work, but MI does have a very contentious Senate race. If someone were to play with the numbers, you would hit two birds one stone. This provides a cover for the fraud.

Are you to tell me that as precincts get more conservative, fewer republicans are voting for John James who does not have the negatives of trump? Surely he would over perform trump and even the straight ticket voters.

---

Remember that nothing in those graphs give any clue at all to how many voters voted for the candidate in question. There does not appear to be any possible way to read anything from the graphs regarding the existence or non-existence of fraud.

---

Sure. But the high correlation along that slope is the problem. We can look at the 4th graph (detroit co?) and other metros in BG states without dominion to get a control group.

If we keep doing the same analysis on counties in MI, of course it will look like a trend of what really took place.

The thing that bugs me about those graphs is that if there was no tampering, why is the line against trump so straight, no curve, no second horizontal slope, no messaging affecting X precinct rather Y precinct.

You mean to tell me that all the none straight ticket voters acted near uniform regardless of messaging? That's a hard sell.

This thread was getting interesting with TravelAg doing the work of verifying the plots and then plotting other counties. Then some came in to derail the discussion by making personal attacks against the scientists who are presented the method and long rambling posts about why it hypothetically cannot work.
Had we had a chance to plot more data from various counties, then we could have had a real discussion of what, if anything, the data represents. We could have then decided for ourselves if this method has value.

I thank TravelAg2004 for your plots and others who provided the data. They started an interesting discussion. It is a shame that we didn't have the opportunity to continue discussing these plots and others that might have followed.

It seems like this is a microcosm of what is happening nationwide. Everything needs to be finished immediately. Many don't want to do any analysis to see if things are working the way they should.

The Debt said:

---

eric76 said:

---

The Debt said:

---

TravelAg2004 said:

---

Here's a quick webpage I put together showing 4 data sets:

• Oakland (President)
• Oakland (Senator)
• Washtenaw (President)
• Washtenaw (Senator)

https://travelag2004.github.io/michigan-votes/

You can turn each one on and off to see how they interact.

When you put them on the same time scale, I'm not seeing anything as dramatic as what they are showing. Not saying something isn't going on here, but it's not nearly as cut and dry as it first appeared. I've made it pretty easy to add more data, so I'll add a few more counties later on.

---

Not to diminish your work, but MI does have a very contentious Senate race. If someone were to play with the numbers, you would hit two birds one stone. This provides a cover for the fraud.

Are you to tell me that as precincts get more conservative, fewer republicans are voting for John James who does not have the negatives of trump? Surely he would over perform trump and even the straight ticket voters.

---

Remember that nothing in those graphs give any clue at all to how many voters voted for the candidate in question. There does not appear to be any possible way to read anything from the graphs regarding the existence or non-existence of fraud.

---

Sure. But the high correlation along that slope is the problem. We can look at the 4th graph (detroit co?) and other metros in BG states without dominion to get a control group.

If we keep doing the same analysis on counties in MI, of course it will look like a trend of what really took place.

The thing that bugs me about those graphs is that if there was no tampering, why is the line against trump so straight, no curve, no second horizontal slope, no messaging affecting X precinct rather Y precinct.

You mean to tell me that all the none straight ticket voters acted near uniform regardless of messaging? That's a hard sell.

---

You are trying to read far more into the graphs than can be determined from the graphs.

About all you can tell about the number of votes from the graphs would be from the number of dots, not the distribution of the dots. Since each dot represents one precinct, fewer dots means fewer precincts and that presumably would mean fewer dots. A county with 1,000 precincts would have 1,000 dots (maybe less if some precincts map to the same dot). A county with 5 precincts would have 5 dots. That's it. if you had 5 precincts with 200,000 voters per precinct in one county and 1,000 precincts with 1,000 voters per precinct in a second county, the graphs would have very different numbers of dots but would represent the same number of votes.

As for the slopes, the differences are pretty much meaningless. They tell you absolutely nothing about how many votes someone did or did not get and how many votes, if any, might have been stolen from them. There is nothing about it that can be used to infer that fraud existed.

One of the stupidest things in the video was where he had a line that started across horizontally and then shifted and went down. It was clear that the horizontal lines were there only to confuse people. In the two graphs I noticed where he did that, one of the graphs would have made more sense just continuing the same line on up. The horizontal line didn't track anything and was completely worthless. On the other, there was a bit more fuzziness on the left side, but continuing the sloping line would have made just as much sense. Again, the horizontal line he drew was there to confuse people.

As far as why the graphs looked similar, that is what one would expect -- a somewhat elongated cloud of data points going somewhat downward. With just a few precincts, there would be just a few points in a very sparse cloud that might be more difficult to see. And, of course, the slope of the line graphed to show the trend would likely show more variation. The more data points, the more consistency you would see in the cloud.

Other than showing the general trend, the line itself is not all that meaningful.

Another thing the graphs don't show you is anything relating to the relative frequency of straight party voters vs non-straight party voters. 90% of the people could have voted straight party or 10% of the people could have voted straight party and you couldn't determine that from those graphs. Remember, all the graphs tell you is what percentage of those who voted straight party for the Republicans and what percentage of those who didn't vote straight party voted for Trump. And even that was done in such a way to get the shape of the graph he wanted.

BuddysBud said:

---

This thread was getting interesting with TravelAg doing the work of verifying the plots and then plotting other counties. Then some came in to derail the discussion by making personal attacks against the scientists who are presented the method and long rambling posts about why it hypothetically cannot work.
Had we had a chance to plot more data from various counties, then we could have had a real discussion of what, if anything, the data represents. We could have then decided for ourselves if this method has value.

I thank TravelAg2004 for your plots and others who provided the data. They started an interesting discussion. It is a shame that we didn't have the opportunity to continue discussing these plots and others that might have followed.

It seems like this is a microcosm of what is happening nationwide. Everything needs to be finished immediately. Many don't want to do any analysis to see if things are working the way they should.

---

You are obviously no mathematician. An individual analysis of the methods leads you to understand what is happening. A graph basically verifies what you learned from the analysis. The analysis explains the graph -- the graph does not explain the analysis.

You are perfectly free to plot the data for every candidate from every county provided that you can find the data. Have at it.

The real microcosm of what is happening nationwide is that people often believe unquestionably what they wish to believe instead of trying to determine whether or not it is nonsense.

In my opinion, any freshman math, statistics, or physics student at A&M who is unable to figure out what he is doing with his analysis should probably change majors as soon as possible.

I guess I am late to this party. There are a number of different ideas thrown out in the video besides the effort to compare individual voters for Trump with the % of votes that are straight party Republicans. A key point that continues to come up in the various claims and counter-claims of fraud is the different types of fraud. It is important to note the difference between Voter fraud and Election fraud.

I think the recommendations in his conclusions at 56:54 are quite useful. We need a verifiable system if we want people to believe the system is sound and not corrupt.

As for the charts Shiva spends the bulk of his video on, his argument seems to be a mixed bag. There are some things that look fishy. OTOH, I don't think he makes the case the sloped line shows what amounts to "point shaving." The screwy thing is the purported algorithm may explicitly escape this analysis.

Here is my rather hurried attempt to get a feel for what to expect from this type of analysis.....

I put together a simple spreadsheet model. In this model there are 6 types of voters.

Straight party Dems
Straight party Reps
Dems splitting their ticket voting Trump
Dems splitting their ticket voting Biden
Reps splitting their ticket voting Trump
Reps splitting their ticket voting Biden

Those are the model inputs and highlighted in pink.


I assumed across the population the % in each group were constant within a party. Then the real question is what happens to the Individual votes for Trump as you vary the mix of Democrats and Republicans.

If my analysis is correct, we probably should expect a downward sloping line across all the range.
Thus any hockey stick patterns are potentially anomalous and suggest one of the segments may be suspect.

The biggest caveat is that this is a model. I'm not sure how well it correlate with reality.

There are some more observations but I will save those in case y'all show my effort to be flawed.


I will add this model will always yield a line that is a specific function of 3 of the model inputs. It is left as an exercise for the reader to derive the equation for the slope of the line. :-)

Suppose that a candidate got 100% of the vote, but nobody in the precinct voted a straight party ticket for his party.

Then, his point on the graph would be (0,100).

Conversely, if all his voters voted a straight party ticket, but none voted directly for him, then his point on the graph would be (100,-100).

That's the furtherest apart that any two points on the graph could possibly be. Yet, nothing about that can be logically construed to indicate which one got the most votes. The results could be really lopsided or they could be completely tied.

There is, I think, precisely one point on the graph in which it is possible to figure out how many votes the candidate received: (0,0). At that point, he got no votes from straight party tickets and no votes from voters voting directly for him. At any other point besides (0,0), it is mathematically impossible to determine how many votes they got.

I don't have much to add to this mathematical breakdown of dr shiva's argument. However, based on the fact that Eric Willfully Refuses to accept and continues to argue:

Concerning MIB in Pennsylvania

1) There is NO CURING of MIB in Pennsylvania.
2) MIB cannot be opened prior to 7AM on Election Day

It is very clear. GO READ THE LAW.

I personally place ZERO confidence that an MIT mathematician is wrong and Eric is right.

aggiez03 said:

---

I don't have much to add to this mathematical breakdown of dr shiva's argument. However, based on the fact that Eric Willfully Refuses to accept and continues to argue:

Concerning MIB in Pennsylvania

1) There is NO CURING of MIB in Pennsylvania.
2) MIB cannot be opened prior to 7AM on Election Day

It is very clear. GO READ THE LAW.

I personally place ZERO confidence that an MIT mathematician is wrong and Eric is right.

---

First of all, you are clearly wrong about curing of mail in ballots in Pennsylvania. Please check the other thread.

Second, the ballots do not have to be opened to examine the outside of the envelope. Whether or not examining the outside of the ballot is permitted remains to be seen.

Third, the guy in the video has zero degrees in Math.

Fourth, any freshman Math, Physics, or Statistics student at A&M who cannot see through the errors in the video should probably find a different major.

I think someone else did the graphs. The dude in the video just presented them.

eric76 said:

---

aggiez03 said:

---

I don't have much to add to this mathematical breakdown of dr shiva's argument. However, based on the fact that Eric Willfully Refuses to accept and continues to argue:

Concerning MIB in Pennsylvania

1) There is NO CURING of MIB in Pennsylvania.
2) MIB cannot be opened prior to 7AM on Election Day

It is very clear. GO READ THE LAW.

I personally place ZERO confidence that an MIT mathematician is wrong and Eric is right.

---

First of all, you are clearly wrong about curing of mail in ballots in Pennsylvania. Please check the other thread.

Second, the ballots do not have to be opened to examine the outside of the envelope. Whether or not examining the outside of the ballot is permitted remains to be seen.

Third, the guy in the video has zero degrees in Math.

Fourth, any freshman Math, Physics, or Statistics student at A&M who cannot see through the errors in the video should probably find a different major.

---

Everyone Ignore this guy.

He is not competent to judge a 8th grade reading comprehension test much less a statistical analysis by a guy that holds 4 degrees from MIT, is a Fulbright Scholar, and has started 7 successful hi-tech companies including EchoMail, CytoSolve and Systems Health. He is currently the Founder and CEO of CytoSolve, Inc., which is discovering cures for major diseases from Pancreatic Cancer to Alzheimer's. He is also the Founder of Center for Integrative Systems that is the home of Innovation Corps and C.L.E.A.N./R.A.W. certifications.

Guy is literally a clown.

If you think that you're so great in Math, how about a little Math problem that was taken from a test for advanced high school students?

Let x, y, and z be real numbers > 1. Prove that if x^y=y^z=z^x, then x=y=z.

Enjoy.

Agree

I guess I'll just post this here.



Take from this what you will.

akingery said:

---

I guess I'll just post this here.



Take from this what you will.

---

Good find.

akingery said:

---

I guess I'll just post this here.



Take from this what you will.

---

I've seen this guy's videos before. I really appreciate how delicately he treats the assorted mouth breathers, moon bats, and charlatans who get angry at him for showing real math. It's really kind of charming that he tries to actually engage them.

I watched this last night (the first one, not the one above). As usual the old guy late to the party. I saw potential minor flaws but the overall findings appear to be pretty solid. The bottom line was that the more red the precinct in percentage, the lower the percent of those not voting straight party line that voted for Trump. Remember, the charts are in percentages. It would appear to be a brilliant scheme using a percentage modifier of actual votes for Trump to move them to Biden (total votes have to match). I am not familiar with the vote counting machines--it would seem this had to be programmed into each vote counting machine or else, if all votes in the county were counted in one machine (I don't see how that would be the case) then that machine could be programmed.

He made the point early that the votes were shown in decimals. I don't recall him saying this specifically, but effectively it seems each vote would be split with a portion to Trump and a portion to Biden. There seems to have been a great deal of resistance to poll watchers and answering questions. Some have surmised that some folks were watching results and turned on the switch or modified the algorithm in the early morning hours of the 4th. I don't see how that would be possible. If an algorithm was used, it likely was developed based on 2016 votes and the algorithm developed to provide the winning margin to Biden. If there were early morning dumps, etc., then those likely were made in order to offset the effects of very high turn out for Trump that had not been estimated correctly in the original algorithm. I am not much into conspiracy theory, but that high turnout could have explained the "closing" of counting that actually continued. Again, I think there may have been some valid explanations for that, at least in some states.

After seeing 4 years of absolute harassment of President Trump and every effort to hamstring his presidency, I unfortunately would not be surprised if all of this did happen. Frankly, I am a little surprised that the President has survived these past 4 years. I find that terrifying to acknowledge that I think someone would try to assassinate the president. In my lifetime I don't believe I have ever seen anyone being attacked so thoroughly over a long period of time. The President has a very strong constitution to have withstood that continuous barrage.

The 'y' increments were not equal on his plots to show Biden's data plotted same as President Trump.

law-apt-3g said:

---

The 'y' increments were not equal on his plots to show Biden's data plotted same as President Trump.

---

I did not understand that Dr. Shiva's team had plotted Biden's data. Has that been plotted by Dr. Shiva or anyone else? That would be interesting as one would probably expect, dependent on how the software is written, that there should be a similar and opposite effect on Biden's vote count.

Or was this in regards to the video on this page? I have not yet watched it.

I'm guessing they are in talks with Sidney Powell and company now.

The video posted on this page explains it pretty clearly. They plot the Biden data. The exact same thing happens. There is no difference between Trump and Biden. The slopes are the same. It's just bad analysis.

Then he goes on to explain why their methodology was pointless and plots the two data sets correctly, again showing no difference between the two candidates in this data. No evidence of vote switching in this context.

Cactus Jack said:

---

The video posted on this page explains it pretty clearly. They plot the Biden data. The exact same thing happens. There is no difference between Trump and Biden. The slopes are the same. It's just bad analysis.

Then he goes on to explain why their methodology was pointless and plots the two data sets correctly, again showing no difference between the two candidates in this data. No evidence of vote switching in this context.

---

Thank you.

I am working on the same data and it will be different. It looks as if the presenter may have made some bad assumptions and also used different axis scales. I should have that a bit later--just got back in from church services.

If I can figure out how to post information from excel I think I can show you that Dr. Shiva is correct and that there is a very good correlation between what might be net losses from Trump to net gains for Biden. All of this would need to be confirmed by actual examination of the ballots.

I believe Dr. Shiva is correct and the second poster is not. His graphs are correct but he fails to note that Bidens are nearly all in the positive while Trumps are nearly all in the negative, such that in an 70% Republican precinct Straight Party Vote Trump gets 37% of the vote by candidate while Biden gets 60.4%. That type of disparity covers at least 60%-70% of the precincts. For each precinct, Biden outperforms the Democratic straight ticket vote by about the same number of votes Trump underperforms the Republican straight ticket vote. That is repeated over and over again.

jgriffith73 said:

---

If I can figure out how to post information from excel I think I can show you that Dr. Shiva is correct and that there is a very good correlation between what might be net losses from Trump to net gains for Biden. All of this would need to be confirmed by actual examination of the ballots.

I believe Dr. Shiva is correct and the second poster is not. His graphs are correct but he fails to note that Bidens are nearly all in the positive while Trumps are nearly all in the negative, such that in an 70% Republican precinct Straight Party Vote Trump gets 37% of the vote by candidate while Biden gets 60.4%. That type of disparity covers at least 60%-70% of the precincts. For each precinct, Biden outperforms the Democratic straight ticket vote by about the same number of votes Trump underperforms the Republican straight ticket vote. That is repeated over and over again.

---

Hope you post it. This is exactly what I was thinking when I watched that second video.

I just took screenshots of excel and posted the pictures. Was easier than trying to post whole excel files.

Regarding the second video, Standup Maths presenter Matt Parker makes an extremely important point that most people seem to be overlooking. He kind of mentioned it in passing and didn't concentrate on the point.

That is, that you cannot add fractions together unless they are fractions of the exact same thing.

If you have three apples and I give you two apples, then you have five apples. If you have three apples and I give you two oranges, you do not have five -- you have three apples and two oranges.

Remember the move Little Big Man? If I remember correctly, at one point the Indian guy says something like "I have four horses and one wife" to which the "Little Big Man" replies with something like I have four wives and one horse.

If you could add them together than each would have 5 of the same thing. But wives aren't horses -- one horse and four wives is not the same thing as one wife and four horses.

Similarly, if you have one car and six cookies, that's a whole lot different from having six cars and one cookie. Yet, if you added 1 to 6 or 6 to 1, you would get 7.

I assume that everyone here learned that, at the very least, by the time they were in high school.

In percentages, it is a bit more removed. Remember that a percentage is a ratio of X/100 expressed as X%. If the Aggies win 80% of their football games this year and 75% of their baseball games in the spring, you cannot add the percentages together and say they won 155% this school year. That would be a preposterous statement. You cannot mix percentages like that. It would be equally silly to subtract one from the other, say 80%-75% to arrive at 5%.

Yet, that is precisely what is done in the first video -- they are subtracting percentages of straight party votes for the candidate's party from percentages of individual votes for the candidate. That is, if the candidate got 60% of the straight party votes and 40% of the individual votes, they are subtracting 60% from 40% to get -20% which is completely meaningless because they two are percentages of different things.

When you are subtracting vote percentages like in the first video, it is particularly misleading also because there is no reason for the number of straight party votes to be anything close to the number of individual votes in any one precinct and also there is no reason that they should be the same ratios in different precincts.

And as I have pointed out earlier, there is only one point on the graph that you can tell how many votes the candidate got in that precinct -- the point (0,0) where he would have gotten no votes via straight party votes and no individual votes.

If you did the plot for my county for Biden, you would see the point (0,0) marked -- in my precinct, neither Biden nor any other Democrat even got one vote.

The point (100,0) would mark a point in which a candidate received 100% of the 100% of the votes on ballots that were voted as straight party votes and 100% of the individual votes from those who did not vote straight party. In spite of that, it is mathematically impossible to determine from the graph how many votes the candidate received. For most points on the graph, it is mathematically impossible to even determine from the chart which candidate got the most votes in that precinct.

To try to claim that it shows vote stealing is absurd. Going further and trying to claim that votes were stolen from one candidate based on the graphs or the other is beyond absurd.

jgriffith73 said:

---

If I can figure out how to post information from excel I think I can show you that Dr. Shiva is correct and that there is a very good correlation between what might be net losses from Trump to net gains for Biden. All of this would need to be confirmed by actual examination of the ballots.

I believe Dr. Shiva is correct and the second poster is not. His graphs are correct but he fails to note that Bidens are nearly all in the positive while Trumps are nearly all in the negative, such that in an 70% Republican precinct Straight Party Vote Trump gets 37% of the vote by candidate while Biden gets 60.4%. That type of disparity covers at least 60%-70% of the precincts. For each precinct, Biden outperforms the Democratic straight ticket vote by about the same number of votes Trump underperforms the Republican straight ticket vote. That is repeated over and over again.

---

I would bet that you are doing something wrong.

In any event, you cannot make any legitimate conclusions of the number of votes received from the graphs except at (0,0). After all, the graphs are crazily subtracting percentages of different things and never show anything about the number of votes anyone received.

I'm no mathematician, just an engineer with a decent grasp of math and statistics, but I've been trying to figure out if this argument holds water or not and I just don't see how it would. The StandUpMaths guy's video is a really good explanation on some of the problems the argument has, with the point that Biden's chart behaving in a similar manner being quite convincing. I haven't tried to chart it out myself yet based on the raw data but might to verify.

I just plotted all the results for the precincts in Kent County based off of the data published on their website. My numbers match exactly the numbers in the "debunking" video and the plots look the same as both videos. I would post the plot but I don't have anywhere that I can publicly post the screengrab so, oh well, you can take my word for it.

BOTH plots demonstrate a similar negative slope (top left to bottom right), just as the "debunking" video shows. But as some noticed, the Y axis on the 2 plots in the "debunking" video are different, and that is significant I believe. I plotted them together so any differences would be evident. There is a vertical shift in the Biden data set vs the Trump data set, and the shift causes the majority of the data for Biden to be in the "plus" territory or above the "0" line. This is not that surprising because it is a 2 choice election for most people so you are going to vote for one or the other.

The reason for "subtracting" the percentages is to attempt to "normalize" each precinct to its own straight party voting percentages. The assumption here (valid or not) is that the straight ticket rep/dem split would be representative of that entire precinct. The subtraction then shows how Trump/Biden did compared to the straight party splits.

Example - If the Rep/Dem split was 60/40 in straight party voting, it is assumed that the remainder of the votes would be roughly 60/40.

What is surprising is that in 205 of the 252 precincts in Kent county, Trump UNDERPERFORMED the straight party voting by an average of 10% and conversely, in 207 of the 252 precincts Biden OVERPERFORMED the straight party voting by an average of 9%.

That means that, according to the results, on average in Kent County Michigan, Biden was roughly +10% on crossover voters.

I also plotted the straight correlations (without the subtraction of %) for those who want to nitpick, and all of the above is still true. Biden significantly overperformed the non-straight party votes when compared to the straight party ticket percentages.

Example - Ada Township Precinct 2:
Straight Party (Rep/Dem) = 53%/45%
Not Straight Party Votes (Trump/Biden) = 36%/60%

So is it possible that Biden was +10% in Kent County non-straight party voters? Yes.

Is it also possible that someone was somehow systematically skimming Trump votes? Yes.

Ultimately, this data can not tell you one way or the other.

Liars figure and figures lie.

MEENAGGIE09 said:

---

The reason for "subtracting" the percentages is to attempt to "normalize" each precinct to its own straight party voting percentages. The assumption here (valid or not) is that the straight ticket rep/dem split would be representative of that entire precinct. The subtraction then shows how Trump/Biden did compared to the straight party splits.

---

Here's an example of why there could be vast differences between the two major parties in voting:

Suppose that in the general election, there is a very popular local candidate who is a Republican. The party of local candidates is not as likely to matter very much, if any. It isn't that unusual to find candidates in one party or another who have broad support in their communities in the other major party as well as their own.

So if a Democrat wants to vote for a popular local Republican candidate, he can't vote straight party while a Republican wishing to vote for the same local Republican candidate could vote straight party.

In this case, you would likely see a high percentage of Republican Party voters and a low percentage of Democratic Party voters voting straight party. With a large number of Democratic Party voters voting for a local Republican candidate, they would have to vote directly for their Presidential candidate since they aren't voting straight party.

Thus, there would likely be a very great and completely legitimate difference in numbers between the numbers for straight party votes and the numbers of votes directly for the President.

What's more, there is no way to deduce from any of this anything about the actual number of votes as well as which candidate won the precinct.

Of course a similar graph for a Biden will show a similar result - it is two sets of two variables, where both sets add to 100. So the ratio of one set will show the same result as the other set. Too busy to work out the details now, but I believe that's the way it is. Surprised Matt Parker didn't get that relationship

FYI: Dr.SHIVA LIVE: MIT PhD is live now:

MEENAGGIE09 said:

---

I just plotted all the results for the precincts in Kent County based off of the data published on their website. My numbers match exactly the numbers in the "debunking" video and the plots look the same as both videos. I would post the plot but I don't have anywhere that I can publicly post the screengrab so, oh well, you can take my word for it.

BOTH plots demonstrate a similar negative slope (top left to bottom right), just as the "debunking" video shows. But as some noticed, the Y axis on the 2 plots in the "debunking" video are different, and that is significant I believe. I plotted them together so any differences would be evident. There is a vertical shift in the Biden data set vs the Trump data set, and the shift causes the majority of the data for Biden to be in the "plus" territory or above the "0" line. This is not that surprising because it is a 2 choice election for most people so you are going to vote for one or the other.

The reason for "subtracting" the percentages is to attempt to "normalize" each precinct to its own straight party voting percentages. The assumption here (valid or not) is that the straight ticket rep/dem split would be representative of that entire precinct. The subtraction then shows how Trump/Biden did compared to the straight party splits.

Example - If the Rep/Dem split was 60/40 in straight party voting, it is assumed that the remainder of the votes would be roughly 60/40.

What is surprising is that in 205 of the 252 precincts in Kent county, Trump UNDERPERFORMED the straight party voting by an average of 10% and conversely, in 207 of the 252 precincts Biden OVERPERFORMED the straight party voting by an average of 9%.

That means that, according to the results, on average in Kent County Michigan, Biden was roughly +10% on crossover voters.

I also plotted the straight correlations (without the subtraction of %) for those who want to nitpick, and all of the above is still true. Biden significantly overperformed the non-straight party votes when compared to the straight party ticket percentages.

Example - Ada Township Precinct 2:
Straight Party (Rep/Dem) = 53%/45%
Not Straight Party Votes (Trump/Biden) = 36%/60%

So is it possible that Biden was +10% in Kent County non-straight party voters? Yes.

Is it also possible that someone was somehow systematically skimming Trump votes? Yes.

Ultimately, this data can not tell you one way or the other.

---

I have been really busy and not able to respond. Dr. Shiva has posted a new video with further explanations and explaining the the single slope line is the pattern that is not normal. The normal first curves upward to a flatter section and a curve downward. In regards to the plot of the Biden and Trump votes, please note the difference in those two plots against the y-axis. The shift is the claimed difference in votes (expressed in percentages) that were moved from Trump to Biden. Dr. Shiva is still right. As he notes, this is not simply a math problem--it requires an understanding of the specific data and what that data represents. Again, this is all further explained in his most recent video. Gig 'em.