Claim: Ballots in Wayne County were run through the tabulator and counted as many as 4-5 times

[Akton]

It's hard to keep track of all the various claims of election fraud, but this one in particular caught my eye when I saw it on twitter because it was one of the first ones that I have seen so far which seems at least plausible to me and is not able to be very quickly debunked. It doesn't seem to have spread very far, so I haven't found any debunkers who have had their way with it yet. I figured I would give it a go and let everyone else check my work.



The full affidavit quoted in the tweet makes a lot of allegations, and most of them don't seem like the kind of thing that ought to be taken seriously (including a claim that cars that dropped off lunch must have been dropping off fake ballots because the witness did not see any lunch). The first allegation got my attention though because of the technical nature of it. The witness seems to be alleging that many ballot counters were not properly discarding batches of ballots before resubmitting them to the scanner after removing a problematic ballot, resulting in the batch being counted twice. In this case, "discard" does not seem to mean throwing out the ballots, but simply not submitting them to be counted and allowing them to be run back through the scanner.

Looking into this myself, I found the Michigan government website that includes information on which voting systems each county uses. We can see that in Wayne county, the tabulator was the "ImageCast Precinct". In order to understand how it is used, I found what I believe to be the user manual for this system here. On pages 26-28 you can find information about discarding batches.

This section in particular stood out to me because it seems to describe roughly the situation the witness is talking about, but it describes a different protocol than the one she says is normal. More specifically, it seems to recommend one that could easily be mistaken by an observer as double-counting the ballots. I will admit the writing is not super clear, but it seems to be saying that when you go through the output bin and remove the offending ballot, you are supposed to send all the ballots removed from the tray during your search back through, since the system will not have scanned them yet (it stopped when it detected a bad ballot). It does not instruct the user to discard the batch before rescanning the entire batch. Now this section of the manual is not describing a jam, but it is describing something very similar. CTRL+F'ing the manual doesn't show any results for "jam", though this section does talk about something "stopping" the machine, which is what jam does, so I figure it is the most applicable.

(edited the above paragraph because I realize I was misunderstanding something about the system. It makes more sense to me now)

This seems to cut against the allegation, but I am posting this here because I am not 100% sure I have found the correct manual, and don't have much knowledge of how voting machines work. Furthermore, the manual itself is confusingly written in my opinion. I actually can't tell if it's talking about putting only the ballots that haven't gone through the machine back through, or if it's talking about putting all the ones that already have back through. The text seems to imply that the user is taking things from the output tray and putting them back in the input tray, but that they are somehow "non-scanned" even though they have gone through and are now in the output tray. Perhaps when the system detects an error, it deletes the scans from its memory of all ballots in the batch, so they are now "non-scanned" despite being in the output tray. In any case, however, the batch is not to be discarded and scanning should continue.

I have only posted an except of what I consider the most relevant portion here, but I invite everyone to read pages 26-28, as well as any other parts they think may be important to better understand this.

 

[Mendel]

Looking into this myself, I found the Michigan government website that includes information on which voting systems each county uses. We can see that in Wayne county, the tabulator was the "ImageCast Precinct". In order to understand how it is used, I found what I believe to be the user manual for this system here. On pages 26-28 you can find information about discarding batches.

It says on page 28:

Article:

In some cases, a user may want to discard a batch. An example of such a situation is when a batch is
accidentally scanned twice, or when the quality of the batch is brought into question.

Source: https://www.sos.state.co.us/pubs/elections/VotingSystems/DVS-DemocracySuite/documentation/ICC_UserGuide.pdf

So allegation 3 hinges on what a "jam" is: if it's merely an unreadable ballot, batches aren't supposed to be discarded. If poll workers re-scan the whole batch, they should discard the batch first; if they rescan just the part of it that wasn't processed, then no discard is necessary. It seems to me that all ballots will have been pulled through the scanner, but the workstation will not have processed the ballots after the erroneous ballot.

Allegation 4 depends on how the software is run. The technician assumes that the poll workers use the software with a batch size of 50 when the workers might simply aggregate several physical batches into a single "software batch". If the limit of 50 ballots per batch is enforced, then surely a batch rescanned 8 times will lead to a number of votes in excess of 50; and that will definitely stand out further down in the process and should be detectable with a simple audit. So even if that did happen, I believe it's likely that this would have been caught and corrected at some point. But it's certainly possible to check this now.

 

[Akton]

It says on page 28:

Article:

In some cases, a user may want to discard a batch. An example of such a situation is when a batch is
accidentally scanned twice, or when the quality of the batch is brought into question.

Source: https://www.sos.state.co.us/pubs/elections/VotingSystems/DVS-DemocracySuite/documentation/ICC_UserGuide.pdf

So allegation 3 hinges on what a "jam" is: if it's merely an unreadable ballot, batches aren't supposed to be discarded. If poll workers re-scan the whole batch, they should discard the batch first; if they rescan just the part of it that wasn't processed, then no discard is necessary. It seems to me that all ballots will have been pulled through the scanner, but the workstation will not have processed the ballots after the erroneous ballot.

Allegation 4 depends on how the software is run. The technician assumes that the poll workers use the software with a batch size of 50 when the workers might simply aggregate several physical batches into a single "software batch". If the limit of 50 ballots per batch is enforced, then surely a batch rescanned 8 times will lead to a number of votes in excess of 50; and that will definitely stand out further down in the process and should be detectable with a simple audit. So even if that did happen, I believe it's likely that this would have been caught and corrected at some point. But it's certainly possible to check this now.

Interestingly enough, the instructions also only refer to discarding batches once the entire batch was scanned, indicating that even in the case of an error, the ballots should all be put through before being discarded (contrary the allegations which indicate scanning should stop). This putting through of ballots even after an error might be mistaken as double-counting.

 

[dogweather]

Excellent - thanks for doing the work and posting this.

 

[Akton]

I've been wondering to myself if there is a simpler and less opaque way to disprove this claim (and related claims) using election data. Surely if there were precincts where many batches of ballots were counted as many as 8 times that would create some sort of statistical anomalies we could observe.

The simplest way to show this would be to look at the number of people who are marked on the voter registration rolls as having cast a ballot and compare that to the final vote tallies. If the two don't match up, then there's an issue. I assume internally that those who run the election would have been able to notice this in their internal tallies and catch it, which is why I personally disbelieve this accusation, but to a conspiracy theorist that won't do as an explanation since all the election workers are presumably corrupt. To prove this using open sources would require access to some source recording the number of people registered to vote who actually cast a ballot. One definite way to calculate this number would be to look at the vote status of everyone on the voter registration rolls, which this website claims are not publicly available in Michigan (the voting status is not available publicly, though much information about registered voters is).

Are there any other methods/sources that could be used to show a match or mismatch between the number of votes counted and the number of people who cast ballots? Information on the number of people who cast ballots definitely exists within the system, as it is used to make sure nobody attempts to cast a ballot twice. It's just not publicly available in the Michigan context it would seem, so we might have to rely on proxies. I'm hoping that somewhere there might be some public datasheet that shows the number of registered voters who cast ballots.

Edit: It would appear that the Michigan SOS website does include a page that includes a number of "total voters", as well as links to the results in individual counties, which often include a raw number of voters who cast ballots, as well as the totals recorded for each candidate. This is a lot of data, but cross referencing it should demonstrate whether there were significant amounts of ballots that were counted multiple times.

Edit 2: Something even simpler actually occurred to me. The allegations are specifically about the counting of mail in ballots, so to make the relevant comparison one would simply need to look at the total number of mail in ballots received vs vote totals when broken down by ballot type. It’s late for me so I maybe I will take a closer look at those numbers tomorrow and make a post about them.

 

[Mendel]

Are there any other methods/sources that could be used to show a match or mismatch between the number of votes counted and the number of people who cast ballots?

There's a paper trail for this; thanks to the good work by Ed Felton of Princeton and others, voting machines without a paper trail are on the way out. Election officials do check the tallies; if it's been properly set up, with a bipartisan set of four eyes. That's all part of the procedure, and probably the most common check. It's just not open to the public, except maybe by the FOIA.

If you want to rig an election, your best bet is to do it before the ballots get counted: stuff the urns, strike voters from the rolls, intimidate them into not going, make voting involve an hours-long wait, make it hard to register or simply don't process the registrations, or get absentee ballots lost or delayed. After they've "entered the system" and get counted, these votes are being tracked very well. (This does not apply to a "state party" -style "democracy".)

 

[Akton]

There's a paper trail for this; thanks to the good work by Ed Felton of Princeton and others, voting machines without a paper trail are on the way out. Election officials do check the tallies; if it's been properly set up, with a bipartisan set of four eyes. That's all part of the procedure, and probably the most common check. It's just not open to the public, except maybe by the FOIA.

If you want to rig an election, your best bet is to do it before the ballots get counted: stuff the urns, strike voters from the rolls, intimidate them into not going, make voting involve an hours-long wait, make it hard to register or simply don't process the registrations, or get absentee ballots lost or delayed. After they've "entered the system" and get counted, these votes are being tracked very well. (This does not apply to a "state party" -style "democracy".)

Thanks, I don’t actually know that much about the nitty gritty of how elections are carried out in America. I’m learning a lot reading about all these various electoral fraud claims. Is there any kind of centralized source/sources I could read that give a general outline of how election results are tabulated and certified in the average state according to best practices? I know that the US election system is a heterogenous hodge-podge so such a thing might not exist, except in such general terms that it’s not informative.

Edit The manual for county canvassers in Michigan describes exactly the auditing process I described on Chapter 4 pages 6 and 7, so if we trust the certification process, this sort of error will definitely be spotted. I wish this was done in some way that was publicly accessible though.

 

[Akton]

Today the Wayne County Board of Canvassers did not certify the results of the election in a 2-2 split along part lines. Does this vindicate the allegations that massive amounts of ballots were counted four to five times?

I would argue no, based on everything we know so far. Apparently (as I have learned) Detroit has a persistent problem with poll books not matching vote totals. There were evidently similar mismatches between poll books and election totals in both the 2016 election and the latest primary in Michigan

The situation in August and earlier imbalances in 2016 were not enough to keep the same board from certifying the results in the August and November 2016.

A full audit was performed of the 2016 results which found an overall net total of 40 extra votes cast, and chalked the mistakes up to human error, finding no evidence of widespread voter fraud.

Given this history, it would be reasonable to think that current discrepancies are of the same order of magnitude. Furthermore, it is reported of the recent decision not to certify:

The decision came after absentee ballot poll books at 70% of Detroit's 134 absentee counting boards were found to be out of balance without explanation. The mismatches varied anywhere from one to more than four votes.

Source

If 134 counting boards each had a discrepancy of "more than four votes" the maximum amount of extra votes we are dealing with would be around 500 (134 times 5 equals 670, times 0.7 equals 469), with the actual number likely being significantly smaller depending on how many counting boards had only 1 or 2 extra votes, how many actually reported undercounts (fewer votes cast than voters) instead of overcounts, and so on. Taking into account the fact that at least some overcounted votes would probably be Trump votes, the extra margin Biden would have gained as a result of these errors would be negligible, probably around 100 if I had to guess, and definitely far smaller than the 146,000 he currently leads by. If batches of 50 ballots each were being double, triple, or sometimes octuple counted four to five times per hour across roughly 2 days of counting and presumably multiple tabulating machines, we would expect to see the total number of extra ballots be at least in the thousands.

Though I find it regrettable that the Wayne County Board of Canvassers has chosen to make this decision given the effect it will have on the public's perception of the credibility of Trump's voter fraud claims, the information that has come out from this decision so far actually seems to be strong evidence against this particular claim of voter fraud. The next step in the certification process is for the results to be double checked by the state board of canvassers. They may take a different stance on these small discrepancies and agree to overlook them the way the Wanye County board overlooked them in past elections.

Edit: It would appear that the Canvassing Board has unexpectedly reversed their decision at the last minute. I take it that they realized the significance of what they were doing and the flimsy evidentiary basis it rested on. I applaud them for this decision, even if much damage was already done.

 

[TEEJ]

Trump immediately tweeted Trump must have been livid when the board reversed their decision!

 

[Hevach]

The Wayne board didn't specify why they changed their minds, but it's still probably not that mysterious.

If they refused to certify, then it gets passed up to the state board, which immediately stated they would certify the county as it stood, as the number of ballots out of balance wouldn't even affect any of the smallest local municipal votes let alone anything county wide or bigger. The county board could either let the state do so, or agree on some remedy themselves (generally a recount or audit of some sort - in this case it is an audit of the most out of balance precincts) with the condition that they must certify the results of that remedy.

Also, the Republican chair of the board and mastermind of this plan got notice that her hearing for *prior* election related ethics violations (conflict of interest as she was running a dark money PAC involved in school board elections under her supervision) has been set for this week, and going into that particular hearing with this on top was probably not going to end well.

 

[Mendel]

From the article that Akton linked:

Article:

A county board that fails to canvass within 14 days after the election must give all of its documentation to the Secretary of State's office and Board of State Canvassers, which then has 10 days to complete the work, canvass and certify the results, according to the board’s canvassing manual.

Wayne County would have been required to pay for the state canvassing work, according to board guidance.

When Kinloch protested that additional county tax money was going to be spent on the ongoing canvass, Palmer said she would be open to certifying much of Wayne County with the exception of Detroit.

Source: https://eu.detroitnews.com/story/news/politics/2020/11/17/wayne-county-canvassers-deadlock-certifying-november-3-election-results/6324274002/

Monika Palmer is the Republican chair of the board who Hevach refers to in the previous post.

 

[Akton]

Any now they have gone back on their going back on their decision not to certify. Ah well. Having a protracted fight over this is terrible for democracy but at least it will probably end up surfacing clear numbers and stats in the public record that can be used to disprove this claim.

I will also add that this washington post story gives us more information on the numbers involved.

Mahar said, however, that there had been “out of balance” precincts, where there were discrepancies between the number of voters and the number of ballots counted. This type of error is common, Mahar said, and often due to human error or computer malfunction. A jammed machine might cause a ballot to get counted twice. Out-of-date software might result in incorrect totals.
According to Mahar, the counts in 28 percent of Detroit’s precincts and absentee counting boards were still out of balance without a clear explanation. But the size of the errors was small — a difference of one or two votes per precinct without a clear explanation of why the imbalance occurred, affecting a ballpark amount of about 450 total votes.

(Mahar is the county director of elections)

450 votes is around my estimate, and the article doesn't tell us if that is total votes that mismatch or just overcounts, but I assume it's total mismatches. That means the net overvotes is going to be even smaller, and some of those overvotes will be for Trump.

Interesting that the quote mentioned jamming machines causing double-counting. It's possible that a machine might have jammed once or twice and the scenario Carone described might have happened to a very limited extent, though the number cited indicates it was not to the extent Carone alleges and certainly not to a degree that it would effect the outcome of the election.

 

[Hevach]

Sounds like Palmer waited until the deadline hit to actually try to rescind her certification, and she might not actually be allowed to. Either way, if she did it just gets bumped to the state board who already telegraphed their intentions.

As a Michigan resident I've voted in these machines many times. In 2004, my ballot even jammed the machine (Saginaw Township Precinct 3 at the time), and our machines are very similar to the ones I see in videos from the Detroit counting centers. It wasn't counted twice - it was counted zero times. It was extracted and the damage ballot was given back to me to inspect. I could void it and fill out a new one, or if I felt it was clearly readable I could give it back in its envelope and it would be placed in a box for uncounted ballots - they said it would be transcribed onto a clean ballot after polls closed, my original destroyed and the transcribed one counted.

 

[Akton]

Sounds like Palmer waited until the deadline hit to actually try to rescind her certification, and she might not actually be allowed to. Either way, if she did it just gets bumped to the state board who already telegraphed their intentions.

As a Michigan resident I've voted in these machines many times. In 2004, my ballot even jammed the machine (Saginaw Township Precinct 3 at the time), and our machines are very similar to the ones I see in videos from the Detroit counting centers. It wasn't counted twice - it was counted zero times. It was extracted and the damage ballot was given back to me to inspect. I could void it and fill out a new one, or if I felt it was clearly readable I could give it back in its envelope and it would be placed in a box for uncounted ballots - they said it would be transcribed onto a clean ballot after polls closed, my original destroyed and the transcribed one counted.

That's for in person voting. I believe the situation might be different for the process involved in counting large batches of mail in votes quickly.

 

[Hevach]

Michigan uses (or at least this year used - this was my first year doing so) the same ballots for mail as in-person (but a different envelope). If you got a mailed ballot this year, one of your options was to take it to the polling place, where after checking in you took the ballot straight tot he same machine and inserted it.

 

[Akton]

I know it's the same ballots but I imagine the process for feeding them through the machine and dealing with errors is different when you are doing entire batches at once vs one single ballot at a time.

 

[Mendel]

Any now they have gone back on their going back on their decision not to certify. Ah well.

Article:

Trump also called two Republican canvass board members from Wayne County [on] Tuesday to offer his support, the person said, after they went back and forth on voting to certify the election results from the state's largest county, which includes Detroit. The board members filed affidavits Wednesday seeking to "rescind" their votes to certify the election result.

Source: https://edition.cnn.com/2020/11/19/politics/gop-michigan-results-trump/index.html

Second source:

Article:

Earlier this week, the county’s two Republicans canvassers blocked the certification of votes there. They later relented and the results were certified. But a person familiar with the matter said Trump reached out to the canvassers, Monica Palmer and William Hartmann, on Tuesday evening after the revised vote to express gratitude for their support. Then, on Wednesday, Palmer and Hartmann signed affidavits saying they believed the county vote “should not be certified.”

They cannot rescind their votes, according to the Michigan secretary of state. The four-member state canvassing board is expected to meet Monday and also is split with two Democrats and two Republicans.

Source: https://apnews.com/article/trump-allies-try-overturn-biden-victory-29da6aac9cc41e47f3095855e7af7031

Apparently Monica Palmer has confirmed this.

Article:

Monica Palmer, the Republican chairwoman of the Wayne County Board of Canvassers who initially voted on Tuesday against certifying the county's election results, then reversed her vote, said Thursday she received a phone call from President Donald Trump Tuesday evening after the meeting ended.

Palmer said she did not know how long the call lasted, saying there was a lot of stress and adrenaline that night.

"He was checking to make sure I was safe after seeing/hearing about the threats and doxxing," Palmer wrote in a text message, referring to a firestorm of information released about her on social media.

Palmer earlier told the Free Press her family had "received multiple threats." "The threats have been made against myself, my daughter and my husband," she said. "Reports have been filed with Grosse Pointe Woods police and the FBI."

Source: https://eu.freep.com/story/news/local/michigan/detroit/2020/11/19/trump-monica-palmer-wayne-canvassers-certification-election/3776190001/

I don't know who or which supporters made death threats, but that's unacceptable.