reSee.it - Related Video Feed

The speaker noticed irregularities with the absentee ballots. The ballot numbers were in sequence, which is unusual for mailed-in ballots. The speaker also discovered that the envelopes lacked a date and the ballots were all from the same area. The signatures were similar, and there were no date stamps. Additionally, the ballots were not showing up in the system and had to be entered manually. When the speaker questioned these issues, they were met with resistance from the supervisor. Despite wanting to stay in the room, the speaker chose not to challenge further.

I'm here to submit my ballot. Is this the deposit box? Yes, your ballot goes in here. It's not electronically counted; it will be secured with others. How can I be sure it's counted? Once it's in the locked box, it will be counted after 8 PM. The computer will read the cards, provided they are filled out correctly. There was a problem with the machine earlier. Occasional jams can happen, but it's not frequent. Thanks for the help. Any other questions? The election board is next door. I think my ballot didn't work because I voted for Trump. Thank you.

The speaker explains that when a ballot jams in the tabulating machines, an error message appears with the options to discard or recount. However, instead of following the correct procedure of discarding the entire batch and rescanning the already tabulated ones, they were repeatedly rescanning the jammed ballots without discarding them. This led to the ballots being counted 8 to 10 times, indicating a lack of understanding and confusion among the individuals involved.

The speaker expresses their intention to vote for Jason Shaw and confirms their ballot. They discuss the process of adjudication and admit to voting for someone who did not vote. The speaker, who is the election supervisor, mentions deciding to count only one candidate in the Trump and Biden race. They then express their preference for Biden and complete the ballot accordingly. They proceed to decide that a particular ballot should be left blank. The speaker confirms that they have scanned the same package multiple times. The transcript abruptly ends with the speaker mentioning having more to do.

There are two batches of votes being examined. The speaker points out that there are identical markings on different ballots, suggesting duplication. They show examples of the same markings and even a dot in the same spot. The speaker mentions having a total of 62 images, but didn't go through all of them. The numbers of the batches are not remembered, but it is clear that duplicate ballots were scanned multiple times.

An observer in the duplication room witnessed the preparation of a new ballot because the original was soiled. The observer read a Trump Republican ballot to their Democratic partner. When the partner entered it into the system, the ballot defaulted to a Biden Democratic ballot on the screen. This was reported to supervisors, and others claimed to have witnessed the same manipulation. No corrective action was communicated. The next day, the observer was scolded for saying the machines were not working correctly and was told not to discuss anything. Many people were threatened with being removed from the room or not being allowed to work there again if they spoke out. The observer believes it is their duty to speak the truth.

Speaker 0 mentions the possibility of printing out something. Speaker 1 asks if there is an explanation for the uncounted votes. Speaker 2 clarifies that there is no concrete explanation for why those votes were not counted by the machine in the first place. Speaker 1 confirms that they don't know why the votes didn't get scanned. Speaker 2 asks if the Dominion Tech guys have figured out the reason, but Speaker 0 says they are not allowed to comment. Speaker 1 believes it wasn't a memory card issue. Speaker 2 asks if memory cards can be ruled out, and Speaker 1 agrees. Speaker 0 suggests it may be human error, but they don't have evidence to confirm it. Speaker 2 questions if it could be a software issue, but Speaker 0 refuses to speculate. Speaker 2 acknowledges the lack of a definite answer.

The speaker discusses the process of adjudication for a batch of votes. They mention a double vote and express uncertainty about whether completing the adjudication will accept both candidates or reject the vote altogether. They also mention the option to hide the box and express uncertainty about whether the check mark was accepted. The speaker then decides to vote for Jason Shaw and completes the process.

Speaker 0 mentions the possibility of printing out something. Speaker 1 asks if there is an explanation for why certain votes were not counted. Speaker 2 clarifies that there is no concrete explanation for why those votes were not counted by the machine. Speaker 1 confirms that they do not know why the votes were not scanned. Speaker 2 asks if the Dominion Tech guys have figured out the reason, but Speaker 0 says they are not allowed to comment. Speaker 2 asks if it could be a memory card issue, but Speaker 1 and Speaker 0 both say they don't think so. Speaker 0 suggests it could be human error, but they don't have evidence to confirm it. Speaker 2 questions if it could be a software issue, but Speaker 0 avoids speculation. They admit they don't have a definite answer yet. Speaker 2 acknowledges this and thanks them.

A person attempted to vote early, but the scanner was broken. Staff said they would scan the ballot later, after everyone left, and offered to put it in a box. The person declined and said they would return another day.

The speaker expresses frustration about a blank ballot that was not voted on. They mention being able to stop adjudication and scan and adjudicate all ballots, including the blank one. Misty asks if a ballot can be scanned more than once, and the speaker confirms that they have done it. They explain that they kept scanning the same batches of ballots. The speaker mentions that they have set the system to handle ambiguous marks and overvotes, but it should also handle blank ballots. They scan a blank ballot and accept it into the system, noting that the system does not know who touched the ballots during adjudication.

Trucks from a company called Runbeck were continuously delivering ballots for several days. The speaker is unsure of the exact duration but mentions that they were still coming in on their last day, which was the 10th. Runbeck is a high-speed scanning company that handles the scanning and printing of duplications, possibly including military ballots. The speaker is uncertain about the purpose of scanning the ballots off-site and whether they were printed or scanned. They mention that the high-speed scanning happens at Runbeck and that there were no observers present. The speaker also raises questions about the origin of the ballots and suggests asking county employees for clarification.

In the video, the speaker describes their observations during a ballot verification process. They noticed a yellow banner indicating "low confidence" on some ballots, but the signatures being compared were illegible and didn't match. The speaker asked about it, but was told not to worry as it was a new program being tested. Later, there was a server outage, but the lights were still on. When the computers came back up, a person who previously had a yellow banner now had a green one indicating "high confidence." The speaker observed conversations among the screeners and heard one person say they were now working on high confidence instead of low confidence.

The speaker expresses their intention to vote for Jason Shaw and confirms their ballot. They discuss the process of adjudication and admit to voting for someone where no vote was cast. The speaker, who is the election supervisor, mentions deciding which candidate to count in the Trump and Biden race. They express their preference for Biden and disregard votes for the other candidate. They proceed to mark a ballot as blank and discuss the ability to scan a ballot multiple times. The speaker mentions scanning the same package multiple times and receiving additional ballots later. The transcript ends abruptly.

The speaker believes that the rejected ballots were placed in a separate box to be later counted at the headquarters. The rejection happened at the voting center due to invalid ballots that wouldn't match any tabulator's program. The question arises if Maricopa County was contacted to clarify their processes. It is mentioned that the rejected ballots would be sent to central tabulation to be duplicated onto readable ballots and inserted into the system. However, there is no way for voters to confirm if this process was actually carried out, which raises concerns.

The speaker witnessed someone using correction tape inside a tabulator machine, which caused the ballots to get jammed. The tape was placed back in the machine after being peeled off. The speaker explained that if a ballot goes through the machine with the tape on it, the machine won't read anything, and the user can override it. The speaker also mentioned that the ballots had no numbers at the top, but they assumed the machine wouldn't read them and the user could manually input the ticket type. The person operating the machine repeatedly put the same set of 27 ballots in instead of separating them, resulting in an inaccurate count. The speaker explained that the machine doesn't have a way to detect if the same ballot has been inserted multiple times, as the sensor and reader are covered.

A scanner can only scan one dollar per second, but there are 4,000 ballots with the exact same time stamp of November 5th at 11:19:40. This suggests that these ballots may have been inserted and cannot be authenticated. The main question is whether this time stamp, which shows an unbelievable speed of scanning, is from the EMS system and not something we created.

- The discussion centers on ballot processing in Maricopa County, with several shipments arriving after the initial belief that counting was near completion. Speaker 0 notes that the Wednesday before the Friday they quit voting, and ten days before they quit tabulating, more truckloads of ballots came in, leading to the question: “how can you not know how many ballots are still out there?” - Speaker 1 asks for clarification: “They thought they were done.” The conversation confirms multiple times that those running the counting rooms believed they were almost done, or would be done, on Wednesday morning, then Thursday morning, then Friday morning, and the process extended into the next week. - Trucks bringing ballots arrived on the third, fourth, and fifth days, continuing throughout the last week. The last day mentioned is the tenth, with ballots still arriving. The company involved is Runback, described as doing high-speed scanning and printing of duplications and military ballots. There was no observer presence at Runback, and Speaker 0 indicates she had not been called to work there; she does not know exactly what Runback was doing (printing vs. scanning). - It is stated that all high-speed scanning occurs at Runback, and the ballots go to Runback. There is uncertainty about off-site scanning and whether Dominion equipment was involved. Speaker 0 clarifies: “They were duplications, the ballots that wouldn’t read through the tabulation machines. They were ballots that came in from military and overseas.” The number of additional sources for ballots beyond military/overseas is unknown, and Speaker 0 suggests this is a question for county employees to explain. - About the counting process: Speaker 0 confirms that the ballots went through tabulation machines and that adjudication work took place for those late arrivals. They observed the ballots being processed, but did not know the exact totals for certain days. - Daily volumes are described. Speaker 0 estimates: one day a shift might handle 90,000 ballots, and some days had similar volumes across three shifts; other days had fewer. There were days when as few as 15,000 ballots were processed. The “back door” arrivals are contrasted with the front door, with Speaker 0 noting that all back door ballots were received through back entries, not the front door. The remaining ballots in the latter part of the period continued to come in and be tabulated, with ongoing full-time shifts through the eighth, ninth, and tenth days. - The episode concludes with Speaker 1 seeking further explanation, and Speaker 0 indicating that some of the details were not fully known and that a county employee should clarify where the incoming ballots came from during the latter part of the period.

The speaker noticed irregularities with the ballot numbers and names on absentee and mailing ballots. The numbers were almost consecutive, and some envelopes lacked a date. When the speaker questioned this, they were met with resistance. The ballot numbers were all from the same area, with similar signatures and no date stamp. None of these details were entered into the system, and they were being manually entered. The speaker suspected something was amiss but didn't challenge further to avoid being kicked out.

The speakers discuss the flow of ballots and the involvement of a company called Runback. Trucks delivering ballots arrived on the third, then the fourth, and the fifth, continuing for days. The last day of the speaker’s involvement was the tenth, and trucks were still coming in. The ballots were coming from Runback, a company that does high‑speed scanning and printing of duplications, and the speaker mentions military ballots being produced or processed by Runback, though there is uncertainty about exactly what Runback was doing. When asked whether the ballots were printed or scanned off-site, the speaker is unsure. It is stated that all the high‑speed scanning occurs at Runback, and that those ballots go to Runback. There were no observers at Runback, and the speaker had not been called to work there. The question is raised about whether the scanning was done on-site at the Maricopa County structure, but the response indicates that scanning was not on-site and occurred at Runback where there are very high‑speed scanners. The question of whether Dominion equipment was involved is addressed: the ballots being scanned were not related to Dominion. The purpose of scanning the ballots in advance of tabulation on Dominion equipment is then explained: they were duplications of ballots that would not read through the tabulation machines, specifically ballots that came in from military and overseas. However, the speaker notes there were more ballots than just those, with trays of ballots being brought in, and uncertainty remains about where the rest were coming from. The speaker suggests that the remaining questions about the sources of these ballots should be answered by the county employees. In summary, the discussion centers on: a sequence of ballot deliveries over several days; Runback handling high‑speed scanning and duplications off-site; uncertainty about whether ballots were printed or scanned and by whom; the absence of observers at Runback; scanners used were not Dominion; the purpose of off-site scanning was to duplicate ballots that wouldn’t read through the tabulation machines, including military and overseas ballots; and unresolved questions about the origin of additional ballots, which require explanation from county staff. The exchange ends with a note that the remaining questions about the ballots’ origins are for the county employees to explain.

Speaker 0 mentions the possibility of printing out something. Speaker 1 asks about the reason for the uncounted votes. Speaker 2 clarifies that there is no concrete explanation for why those votes were not counted by the machine. Speaker 1 confirms that they don't know why the votes didn't get scanned. Speaker 2 asks if the Dominion Tech guys have figured out the reason, but Speaker 0 says they are not allowed to comment. Speaker 2 points out that it hasn't been confirmed if it was a memory card issue. Speaker 1 agrees and suggests it could be human error. Speaker 0 says the ballots didn't transfer over correctly, but they don't have a definite answer yet. Speaker 2 asks if it could be a software issue, but Speaker 0 refuses to speculate. They conclude that they don't have a pinpointed answer at the moment.

The election supervisor wants to vote for Jason Shaw. During adjudication, the supervisor states they want to vote for Biden, overriding the original marks on a ballot. The supervisor decides someone "don't deserve no votes" and chooses not to count votes on that ballot. The supervisor then declares another ballot should be blank. The supervisor admits to scanning the same batches of ballots repeatedly.

Senator Bennett and others discuss how ballots are read by tabulation machines and what causes rejections. The corner marks (timing marks) determine top/bottom of the ballot, while the side marks determine the alignment of the bubble to be counted for a candidate or measure. The machines reject ballots when timing marks are incorrect or too small. The speakers explain that if timing marks are not exactly the right size and they are not perfectly black, the software will reject the ballot; blotchy printing can also trigger rejection because timing marks must be 100% black and of the correct size. Madam chair notes that many unreadable ballots were attributed to printers being set incorrectly and printing too lightly. The first reason ballots are rejected is timing marks being too small. A presenting slide shows that the bottom ballot has timing marks that are 5% larger on the edges, but the computer program requires precise, perfectly black timing marks; any white specs in the black box render the timing mark invalid and the ballot ejected. The blotchy printing is linked to printer fuser temperature settings; ballots printed on heavier paper (80 or 100 pound) require higher fuser temperatures. If the temperature isn’t high enough, toner doesn’t bind and appears blotchy, causing invalid timing marks. Speaker 1 asks whether central count machines or other systems could be set to accept lighter print, vs. those used at voting centers. Speaker 0 rejects this, stating the same ballot definition file is used and central count did not allow looser image requirements. The discussion clarifies two types of marks: timing marks (must be perfectly black and correctly sized) and voter bubbles (the threshold for counting a vote). The threshold setting on the scanner determines whether a bubble is counted: typically between 14% and 35% fill. If a bubble is less than about 14% filled, it is considered a stray mark; if more than about 35% filled, it is counted for the selected option. Some voters may partially fill an oval, or fill more extensively, and the software uses this threshold to decide if a vote is cast. In summary, ballots are rejected if timing marks are not perfectly black or incorrectly sized, or if printing is blotchy due to inappropriate printer settings. Central count uses the same ballot definition file, so lighter prints are not accepted. The bubble threshold determines which marks are counted as votes, with explicit percentage cutoffs to distinguish stray marks from valid votes.

The speaker discusses the concept of overvoting, which occurs when multiple parties are voted for in a section where only one vote is allowed. They mention scanning a new batch of ballots and scanning them again, even though they have already been scanned once. The speaker remarks that all the ballots went through the system without any apparent issues. They then mention going back to RTR.

In the video, the speakers discuss the issue of 17,852 ballots that lack corresponding ballot images. Speaker 0 acknowledges the significance of this number, prompting Speaker 1 to explain it. Speaker 1 clarifies that while they can explain the number, they cannot explain how it occurred. They express hope that more information will be provided on Tuesday regarding this matter.