Debate Now MIT Analysis of Voting Machine "Fraud" Analysis Thread.

[Oddball]

Has anyone done a forensic of the program itself to see if the code was written to create fraud?
Then if so, get coders or execs to talk who ordered it and was Pelosi or any agent of a politician involved in the scheme, including CIA agents or FBI and were they threatened or compromised not to talk?

I've heard that any number of them were connected to the internet for "updates"....If that's true, then someone who knows what they were doing could install the hack right then.

 

[flacaltenn]

Has anyone done a forensic of the program itself to see if the code was written to create fraud?
Then if so, get coders or execs to talk who ordered it and was Pelosi or any agent of a politician involved in the scheme, including CIA agents or FBI and were they threatened or compromised not to talk?

If you watch the video, the Dominion software is not open code so not available for analysis.

So could they court order the company or question the code and program creators to get them to talk?

You have to have credible evidence to get a court order or any other asses in gear.. That's what this thread is LASER focused on this one MIT theory.. And believe me when I say this -- I WANTED this to be true because it's the ONLY way forward to GETTING a hearing on this whole thing.. But I think Dr Ayyaduri is short of the mark in proving ANY machine malfunction here..

 

[Oddball]

Has anyone done a forensic of the program itself to see if the code was written to create fraud?
Then if so, get coders or execs to talk who ordered it and was Pelosi or any agent of a politician involved in the scheme, including CIA agents or FBI and were they threatened or compromised not to talk?

If you watch the video, the Dominion software is not open code so not available for analysis.

So could they court order the company or question the code and program creators to get them to talk?

You have to have credible evidence to get a court order or any other asses in gear.. That's what this thread is LASER focused on this one MIT theory.. And believe me when I say this -- I WANTED this to be true because it's the ONLY way forward to GETTING a hearing on this whole thing.. But I think Dr Ayyaduri is short of the mark in proving ANY machine malfunction here..

You prove these things by gaining access to them.

This isn't a standard legal proceeding where the accused has certain rights....This is supposed to be a free, fair, open, and transparent operation.

 

[flacaltenn]

Has anyone done a forensic of the program itself to see if the code was written to create fraud?
Then if so, get coders or execs to talk who ordered it and was Pelosi or any agent of a politician involved in the scheme, including CIA agents or FBI and were they threatened or compromised not to talk?

I've heard that any number of them were connected to the internet for "updates"....If that's true, then someone who knows what they were doing could install the hack right then.

When you and I discussed this before -- I thought being "connected" WOULD be a requirement. Because I thought you would need "real time" feedback to adjust your "skimming and weighting"..

But since I invested at looking at THIS presentation from Dr Shiva-- It's pretty clear that if you HAVE machines that mysteriously have a fractional "weighted" voting feature -- you don't NEED the feedback if you're "not too greedy" with vote shifting..

Simple solution man.. Outlaw fractional math in voting software that does tallying and reporting.. Let the MEDIA do the math to get accurate decimal point percentages for results. Then it's easy to review code and updates from all manufacturers...

So investing the 8 hours thinking this over wasn't a complete loss. Just wish some of the "STEM students" on USMB would join in here and noodle on it..

We DO have about a dozen of them.. They just don't LIVE or camp on USMB..

 

[HaShev]

Has anyone done a forensic of the program itself to see if the code was written to create fraud?
Then if so, get coders or execs to talk who ordered it and was Pelosi or any agent of a politician involved in the scheme, including CIA agents or FBI and were they threatened or compromised not to talk?

If you watch the video, the Dominion software is not open code so not available for analysis.

So could they court order the company or question the code and program creators to get them to talk?

You have to have credible evidence to get a court order or any other asses in gear.. That's what this thread is LASER focused on this one MIT theory.. And believe me when I say this -- I WANTED this to be true because it's the ONLY way forward to GETTING a hearing on this whole thing.. But I think Dr Ayyaduri is short of the mark in proving ANY machine malfunction here..

Isn't past performance admissable as evidence?
They've had many many previous problems that can raise evidence of a pattern that opens up other avenues of investigating like: why would they use an often flawed company's software for this election, *why did they get so many contracts* with such problems existing over various years usages?
That's when I'd submit previous abuse complaints regarding politicians cashing in on this software, like *Feinsteins Husband and Pelosi who both own shares of the Co.*

 

[flacaltenn]

Has anyone done a forensic of the program itself to see if the code was written to create fraud?
Then if so, get coders or execs to talk who ordered it and was Pelosi or any agent of a politician involved in the scheme, including CIA agents or FBI and were they threatened or compromised not to talk?

If you watch the video, the Dominion software is not open code so not available for analysis.

So could they court order the company or question the code and program creators to get them to talk?

You have to have credible evidence to get a court order or any other asses in gear.. That's what this thread is LASER focused on this one MIT theory.. And believe me when I say this -- I WANTED this to be true because it's the ONLY way forward to GETTING a hearing on this whole thing.. But I think Dr Ayyaduri is short of the mark in proving ANY machine malfunction here..

You prove these things by gaining access to them.

This isn't a standard legal proceeding where the accused has certain rights....This is supposed to be a free, fair, open, and transparent operation.

Not bullying your way in without evidence.. And ANY substantial vote shifting CAN be detected by data forensics that Dr. Ayyaduri tries to do here. He KNOWS what to look for -- he just set up the problem wrong..

THIS IS the way you access to do massive invasive investigation.,.

 

[Oddball]

Has anyone done a forensic of the program itself to see if the code was written to create fraud?
Then if so, get coders or execs to talk who ordered it and was Pelosi or any agent of a politician involved in the scheme, including CIA agents or FBI and were they threatened or compromised not to talk?

I've heard that any number of them were connected to the internet for "updates"....If that's true, then someone who knows what they were doing could install the hack right then.

When you and I discussed this before -- I thought being "connected" WOULD be a requirement. Because I thought you would need "real time" feedback to adjust your "skimming and weighting"..

But since I invested at looking at THIS presentation from Dr Shiva-- It's pretty clear that if you HAVE machines that mysteriously have a fractional "weighted" voting feature -- you don't NEED the feedback if you're "not too greedy" with vote shifting..

Simple solution man.. Outlaw fractional math in voting software that does tallying and reporting.. Let the MEDIA do the math to get accurate decimal point percentages for results. Then it's easy to review code and updates from all manufacturers...

So investing the 8 hours thinking this over wasn't a complete loss. Just wish some of the "STEM students" on USMB would join in here and noodle on it..

We DO have about a dozen of them.. They just don't LIVE or camp on USMB..

It's pretty clear that if you HAVE machines that mysteriously have a fractional "weighted" voting feature -- you don't NEED the feedback if you're "not too greedy" with vote shifting...

Which 'splains why the counting was shut down right as the vote numbers and percentages of the total were reaching the critical point where states have been called in the past.....Then the pallets of ballots were rolled in during the dead of night, and the process on Wednesday was done with the most minimal of oversight.

IMO, those pulling the strings got the results back from Florida and shit chickens.....Subtlety was no longer an option.

 

[flacaltenn]

Has anyone done a forensic of the program itself to see if the code was written to create fraud?
Then if so, get coders or execs to talk who ordered it and was Pelosi or any agent of a politician involved in the scheme, including CIA agents or FBI and were they threatened or compromised not to talk?

If you watch the video, the Dominion software is not open code so not available for analysis.

So could they court order the company or question the code and program creators to get them to talk?

You have to have credible evidence to get a court order or any other asses in gear.. That's what this thread is LASER focused on this one MIT theory.. And believe me when I say this -- I WANTED this to be true because it's the ONLY way forward to GETTING a hearing on this whole thing.. But I think Dr Ayyaduri is short of the mark in proving ANY machine malfunction here..

You prove these things by gaining access to them.

This isn't a standard legal proceeding where the accused has certain rights....This is supposed to be a free, fair, open, and transparent operation.

The companies themselves have PROPRIETARY rights. The precints, districts, election officers have rights to be innocent until some CREDIBLE guilt is presented.. You're not gonna get there tossing MORE shitballs at the wall.. THIS kinda thing in THIS thread is the key to opening up the path..

 

[flacaltenn]

BTW-- Dr. Ayyaduri and crew leapt the shark and spent the last 20 minutes of their "science/engineering" presentation to blow off an incredible amount of partisan steamed shit.. THa't s a bonehead move if you're trying to PROVE voting machine hijinks.. Because in the heated 'near war' political backdrop -- you've just buried your "science/engineering" objectivity...

 

[Oddball]

Has anyone done a forensic of the program itself to see if the code was written to create fraud?
Then if so, get coders or execs to talk who ordered it and was Pelosi or any agent of a politician involved in the scheme, including CIA agents or FBI and were they threatened or compromised not to talk?

If you watch the video, the Dominion software is not open code so not available for analysis.

So could they court order the company or question the code and program creators to get them to talk?

You have to have credible evidence to get a court order or any other asses in gear.. That's what this thread is LASER focused on this one MIT theory.. And believe me when I say this -- I WANTED this to be true because it's the ONLY way forward to GETTING a hearing on this whole thing.. But I think Dr Ayyaduri is short of the mark in proving ANY machine malfunction here..

You prove these things by gaining access to them.

This isn't a standard legal proceeding where the accused has certain rights....This is supposed to be a free, fair, open, and transparent operation.

The companies themselves have PROPRIETARY rights. The precints, districts, election officers have rights to be innocent until some CREDIBLE guilt is presented.. You're not gonna get there tossing MORE shitballs at the wall.. THIS kinda thing in THIS thread is the key to opening up the path..

But there are no aspersions being cast upon any of the election officials....They're just doing their jobs with the equipment provided.

As a Linux guy, I know far more about what makes computer equiment tick than the average Joe...But I still wouldn't know my ass from a hot rock when it came to something as specialized as a tabulator.

 

[Thunk]

BTW-- Dr. Ayyaduri and crew leapt the shark

Don't discount the POSSIBILITY they may be a red herring.

Time is of the essence right now...any time wasted chasing untamed geese is a win for one side.

 

[toobfreak]

I'm going to make a few general broad observations here first based on memory having watched the video last night and with Flacc already kind of making some of the points I had intended.

• One problem I have with these things is some of the untestable assumptions they are based on, such as the x axis of the scatter-plot claiming a percentage of Republicans in each precinct. Percent based on what? Percentage of total eligible voters? Percentage of registered voters? Then from that deriving that the balance then MUST be the other party! What about independents?
• Likewise, I only watched the vid once and last night, but I also took issue with the assumption as I understood it of separating people who vote SPV vs. voting individual races. Maybe it should be mandated that you have to enter a preference in EVERY race. Perhaps: R, D, I, and NP (no preference).
• As I understood it, I also had an issue with the chart counting independent votes for Trump (no other races called) as counting AGAINST him? How can you get a vote that counts against you?
• I did not like that the y-axis of the chart DOES NOT GIVE any label to what it is plotting against % of republicans! +20% of WHAT? -30% of WHAT?
• If someone votes straight republican, we have a HIGH DEGREE of confidence that this person is registered republican (though I'm not sure why we must guess this and don't tabulate that data during the voting process).
• I know from personal experience that those times I've voted by individual race (not SPV), that I still voted republican. I tend to register for the party I support most (though I voted straight R for many years before changing my registration from D to R as well), and contrary to what some here might think, I don't register R or vote R due to partisanship-- -- I would gladly vote D IF THEY ONLY PUT UP PEOPLE with less harebrained schemes! I mean, there was a third person in with Hillary and Bernie in 2016 that actually made some sense and this time there was Tulsi Gabbard who wasn't so bad, so, it would be helpful if there was some data on percentages of people who cross vote: registered for one party while voting some/all candidates in the other party. That data should be party of the equation.
• Finally, the one other general observation I want to make is that if Dr. Shiva really wanted to do this right, he could have had at least data from ONE OTHER ELECTION as a baseline reference and test of his data and methods in THIS ONE. How would the plots look for these precincts for the 2016 election? If he did plots of same precincts in 2016 using the same calculations and they gave similar point distributions, that would say quite a different thing from doing it all and the 2016 election (where presumably, little/no fraud occurred) gave a totally DIFFERENT point-spread distribution! Without looking at other election years like 2016, I think it makes it rather hard to interpret what these plots are telling us FOR SURE.

 

[colfax_m]

The idea that this is somehow proof of election fraud is laughable in my mind.

His analysis splits the electorate up into two groups. One group votes straight ticket and one votes for each candidate individually. His assumption is that people who vote straight ticket and people who vote for each candidate individually should vote for Trump at equal rates.

This assumption is completely untested and unfounded. He barely tries to defend it. He spends no time attempting to demonstrate this is what we should assume is correct. There's no normal comparison group. Nothing. We are just supposed to believe it's true and use this as a basis for saying that votes were shifted.

I think, instead, he's just demonstrating human nature.

I think it's pretty obvious that the people who vote straight ticket are more partisan than those who do not. Whether they are Democrats or Republicans, they're partisan. That means those in the middle are more likely to be swing voters. I think that swing voters are less likely to vote heavily one way over the other, their vote is more evenly distributed between Biden and Trump.

Therefore, when a precinct is highly partisan Republican, when a precinct has a high percentage of people voting straight party ticket Republican, the swing voters are unlikely to be equally heavily Republican and vice versa. This gives the exact same outcome as his data without the need for fraud.

Using SPVoting as proxy for the "republicanism" of any particular precinct IS problematic.. Because it's ASSUMED that "in some mathematical fashion" SPV will go up "correlated" with the partisianship of that precinct. We know that intuitively that's true. But SPV doesn't prove that.
Will work for Dems as well..

But it's NOT a definitive measure of the partisanship.. I can give you a couple better ones..

1) Use the percent REGISTERED Repubs in that precinct.. It's available -- have at it..
2) Better choice. Because this MIT analysis never uses any metric other the Repub votes. And No TURNOUT factor. And turnout boosts the vote for one party or another. So use the relative TURNOUT for each party.. This metric isn't part (to my knowledge) of a General election race. Because no one is queried for party at polls or by mail.. But it can be DISCOVERED thru inquiries to the registrars or approximated by the percentages posted in the results for the Prez race..
3) Just use the final vote for Trump/Biden in that district as your "measure of partisanship".

So much for his "x-axis" independent variable.. Now look at the Y axis.. (See OPost for definitions)

(OTV - RSPV) / RSPV

This choice ONLY means anything in terms of the method voters use to choose to fill out a ballot.. NOTHING AT ALL to do with the actual competition or imbalances in Turnout. And actually the "minus" seems kinda arbitrary at first glance. But it's NOT.. The way it's formulated -- as RSPV goes UP -- the numerator goes downs linearly, but the fraction goes down even faster.

Seems to me the clearer metric is to use --
OTV + RSPV) / RSPV

In this case the Y value has strategic meaning.. The numerator is TOTAL VOTES for Trump in that district. It wasn't the other way 'round.. And in this case -- as RSPV goes UP -- the NUMERATOR goes up -- but the fraction goes down even faster.

MY BET IS -- if you used the latter Y formula -- the curves would be FLAT like he postulates they SHOULD BE -- if there was no "machine hanky panky"..

I'm looking at how much work it would be to get Michigan precinct data for one of Dr. Shiva's examples.. And it would take maybe an hour to RE-RUN a scattergram with that change..

To be honest, I don’t really see any value in this data at all other than being descriptors. What I think it really shows is that people of a district who don’t vote party line are less likely to vote as overwhelmingly for Trump or Biden as those who do vote party line.

 

[toobfreak]

Looking at two of the plots Shiva put up, one for a mostly R country (left) vs. one for a very D county (right):

• It seems patently impossible that several other R-leaning counties would follow nearly THE EXACT SAME PLOT DISTRIBUTION CURVE as this one, all with the transistor switch around 20% republican taking them from a positive plot straight down a slope very negative. That the negativity would be a function of how red the precinct was, that it would be so consistent with other counties and follow a STRAIGHT LINE screams of the kind of superimposed simple weighting algorithm one would use if trying to institute a fast and dirty change to the outcome! Of course, better algorithms could be written to disguise and randomize themselves with similar effective results, but that takes skill and why take the time if you don't expect a wholesale AUDITING of all the precincts for just this sort of thing? Once again, this needs tested against other calculations like that offered by Flacc in post #20 and at least one other election year IMO in order to have any meaningful interpretation, especially in getting a court to herald a large scale investigation.
• Shiva rambled on a bit and tended to jump around from topic to topic, so now a day later, I'm a little vague now on remembering his methodology of how these Trump individual votes were counted to make them more or less positive, but one good sign is that the 2nd chart shows what you'd expect! Detroit is very blue, most precincts average about 2-5% red precincts, and the point spread was rather wide and random, diffusing to the right positively. But I'm not sure sure of their 10% graph line, nor how precincts with ZERO republicans could score up to 50% POSITIVE (far left line).

 

[toobfreak]

The one thing that occurred to me was what if these plots REALLY WERE DUE to whether republicans were still voting R, but SIMPLY NOT VOTING FOR TRUMP? But Shiva seems to have accounted for this in the following plot.

Again, at this point having only watched it once last night and being tired, I've lost how he is figuring on the y axis WITHOUT SEEING BIDEN VOTES and factoring them into the equation, but basically, this is more what one would THINK you'd get -- -- actually, if Trump lost a lot of votes with republicans, the plot should probably look more like THIS:

He would lose votes but still have positive votes as well as not everyone (obviously) was soured on DJT. Again, the fact that we DIDN'T see something more like this (more like Detroit above) is significant, but statistically inert without more data, tests, analysis and a BASELINE ELECTION YEAR for 2016 to COMPARE IT TO to SHOW a judge a concrete change tied more to election PERIOD rather than scoring methodology and statistical error or bias in measuring.

 

[toobfreak]

I agree that someone with more of an interest in statistics and knowledge in elections should look at this. At the very least, I think a panel needs convened to take a hard and thorough look at the Dominion software. If Trump and the GOP were on the ball, this might have been a good thing to look into over the PAST FOUR YEARS. But I do agree with some of the conclusions by Shiva:

1. The inputs and outputs of the system are unverifiable, a fatal flaw in the system that never should have happened rendering it nearly impossible to show and PROVE fraud.
2. The code being used here in any electronic/computerized medium counting votes needs examined and any potential for weighting eliminated.
3. Ballot images should not be discarded for at least two years.
4. Voting should be audited.
5. The system should be TRANSPARENT to not only keep it honest, but to keep it free of bugs and errors.

The fact that so many would be opposing this stuff much less the attorney general or whomever Shiva said deleted his Twitter account when he tried to share the data goes a long way of raising questions as to why so much SECRECY has crept into our voting methods!

Rather than go back to watch the video and worry about the math he used here, my opinion isn't what matters, this stuff should be getting looked at and analyzed by those actually leading these legal challenges in court.

 

[Oddball]

"Fire the democrats...Fire the republicans."

Now there's a policy prescription I can live with.

 

[WelfareQueen]

This is a STRUCTURED DEBATE thread.. I get to pick 2 rules in addition to the Zone2 regular board rules.. Here they are:

1) Discussion in this thread is ONLY about the posted analysis from Dr Shiva at MIT.. If you're not willing to watch the presentation, understand his definitions, analyze his results - dont bother posting here. I assume just a dozen or so USMB members will make that effort.. Others will be thread banned and/or warned.

2) Not the place to go into gory detail about the IMPLICATIONS of finding systemic voting machines in America.. Just jumping the gun BEFORE any evidence certain.. We all understand that FINDING systemic voting machine issues is an Armageddon type situation and a potential nation ending disaster,.. THIS THREAD is for pure analysis of ONE THEORY...

This presentation by the MIT folks was the FIRST AND ONLY assertion of machine voting issues that used ACTUAL open sourced election data. And it's a repeatable experiment.. Those are the kind of assumptions that ultimately matter. It's gotten SCADS of attention in the MMedia and SoshMedia in the past 4 or 5 days.. And I initially thought something was proven here.. After watching the vid several times and portions MANY times - trying to understand his variable definitions and assumptions -- I'm having doubts at what he's proven here.

My current thinking is that because of the way HE CHOOSE his variables and the ASSUMPTIONS that he made -- that ALL he's really proven here is that "partisans in states that allow "STRAIGHT PARTISAN" voting options -- prefer that method over checking boxes on a "normal" full ballot.. I'll flesh out that analysis later on..

But simply -- Straight party voting (SPV) preference IS prefered by partisans over the effort of splitting votes on a normal ballot.. His SCATTERPLOTS use Republican SPV (RSPV) as a measure of the PARTISIANSHIP of that precinct on the Xaxis.. But that is not the partisan SPLIT in that precinct. His Y axis takes "Other Votes for Trump" (OVT) in that district where the voters choose NOT to SPV.. He Subtracts OVT - RSPV to use on that axis,

Given that problem definition -- The slope of ANYTHING plotted in that space is -- (OVT - RSPV) / RSPV..

ALL THAT IS a "proof" that partisans of ANY PARTY prefer straight party line voting.. A "Duhh" moment.. And the negative slope on all his scattergrams just shows that "open ballot voting" drops off the more a district leans Republican.

My contention is -- that if he ran the SAME TEST on DEMOCRATS -- you'd get the same results..

I'm NOT done yet. Still arguing with my wife and biz partner. We both have EXTENSIVE work and publications in signal and image processing data analysis and can handle "HSchool Algebra 2" math and graphing -- which is what this exercise really is.. It's just REALLY TRICKY to set up the variables and definitions. After that, it's simple statistical graph analysis.

Welcome all folks that are capable of reviewing this.. If you dont know what a scatterplot is or independent/dependent variables on a graph -- please dont even try.. There's DOZENS of OTHER threads on this..

The presentation is at the link below and COULD (given the SoshMedia MMedia tyrants) disappear tonight from the web.. So be quick to watch it and take notes if you want to join in..

Hey man, you probably think I'm an idiot because I generally don't take things too seriously. But I have a graduate school research background and I know statistical analysis and regression pretty well. I know you understand this as well from your posts on climate change and what the data really shows.

Having said that my take on the MIT professor's data is this. I find it troubling, but I cannot say conclusive. Statistically virtually anything is possible given enough trials. Quantum mechanics comes to mind. The data set he presents is troubling because it does seem to correlate to potential voter fraud. But to prove that case in a Court of Law the correlation would need to be damn near 100%. He has not done that.

However, I think this professor is correct in bringing this information to the public. It does present a pretty strong case for possible voter fraud.

As an aside, I find it interesting that all the identified glitches in this election seem to trend in only one direction (I.e. favoring Biden). That could be reporting bias ( Trump is the one alleging fraud), but at some point the "glitches" should start to include some errors that hurt Biden. Thus far, I have seen none, which like the professor's data is at very least highly suspicious.


Regards

 

[flacaltenn]

BTW-- Dr. Ayyaduri and crew leapt the shark

Don't discount the POSSIBILITY they may be a red herring.

Time is of the essence right now...any time wasted chasing untamed geese is a win for one side.

I've invested the time because there are NO OTHER data forensics studies I've seen in the media finding EVIDENCE of machine vote skimming.. THIS IS IT.. And it's gotten TOO MUCH media play on the right wing side and very little scrutiny.. I WANTED this to be true -- but the methodology is so silly and poorly designed -- that this Dr. Ayyaduri study is pretty useless. Only thing they've proven is that as partisans INCREASE in a district -- they take advantage of Straight Party voting if the state makes it available.. It's a dud -- dude..

 

[flacaltenn]

Hey man, you probably think I'm an idiot because I generally don't take things too seriously. But I have a graduate school research background and I know statistical analysis and regression pretty well. I know you understand this as well from your posts on climate change and what the data really shows.

Don't think you're an idiot and I'm not shocked that you're a STEM type of hooligan.. LOL...

Having said that my take on the MIT professor's data is this. I find it troubling, but I cannot say conclusive. Statistically virtually anything is possible given enough trials. Quantum mechanics comes to mind. The data set he presents is troubling because it does seem to correlate to potential voter fraud. But to prove that case in a Court of Law the correlation would need to be damn near 100%. He has not done that.

It's troubling because it CANT prove vote shifting. It's only variables used are SPVoting and "Other Trump votes".. HE LEAPS to conclusions about the number of votes "stolen" with NOTHING in either axis that relates to "differential partisan turnout" or "partisan race totals" or even the strength of Republicans in that precinct.. His use of Repub SPVoting as a proxy for Republican STRENGTH in that district is even a stretch. The two variables are VOTING CHOICES given to Republican voters in Mich. Not anything COMPETITIVE related to the race..

Of COURSE the slope of line is gonna go down.. Because he SET IT UP to go down.. ANY data spilled into those axes has a slope of (OTV - RSPV) / RSPV. BY DEFINITION it's gonna go down because as SPV GOES UP -- the Y point goes DOWN !!!! Because of the subtraction in the numerator and a LARGER reduction from the denominator..

Might work if the minus became a plus a bit better. And THAT would be a clear variable definition -- because with the + sign -- that's the Total Trump vote in that district..

OTV would only INCREASE in precincts that AREN'T predominately leftist,. And the 4 largest Mich counties - like any state -- ARE PREDOMINATELY left leaning by wide margin..
It's completely those 2 simplistic variables. His use of Repub SPVoting as a proxy for Republican STRENGTH in that district is even a stretch.. So -- it's easy math.. But the SET-up to the problem was bungled badly..