Discover more from Delaware Counts Newsletter
Open Letter to The Department of Elections, The Board of Elections and The Public
There was a meeting of the Board of Elections Monday 9/19/22. In the allowed 3 minutes, I respectfully informed the board and commissioner that I had concerns with regard to our election systems both in terms of vulnerability and non-compliance. Due to the logistics and communications restrictions imposed by the board that prevents normal communications, I let them know I had placed an open letter to the board, the DOE and the public regarding such concerns at DelawareCounts.com/open-letter
An open letter to: The Department of Elections, The Board of Elections and The Public
Over a year ago I started looking into our election processes here in Delaware. I believe I have found many vulnerabilities and some inefficiencies in our election systems. I am concerned that in some cases State Processes may not be in conformance with State Law. Even if the vulnerabilities have never been exploited I believe that to the extent possible they should be eliminated. I have proposed many solutions to eliminating those vulnerabilities. A summary of observations is presented in Exhibit 1 attached.
With observations, concerns and solutions identified my next goal was to communicate and interact with personnel at the Department of Elections (DOE) and those on the Board of Elections (BOE). I wanted to receive more information and I wanted to make them aware of the items presented in Exhibit 1.
I have tried to interact through the casual and more formal methods. As a member of the public I have found it difficult to interact with the DOE or BOE on anything but the simplest of issues. As illustrations of this difficulty I provide Exhibit #2 titled My Experience Trying to Interact with the DOE and BOE.
As a fellow citizen I believe the concerns presented Exhibit 1 still warrant review and action to eliminate vulnerabilities. I am still requesting that the BOE and / or DOE report back or interact with me and or the public in a way and at a level of detail that demonstrates that they are aware of the issues and with respect to what they plan to do to resolve vulnerabilities if anything.
If there are any questions or if you would like to interact regarding solutions please feel free to contact me at email@example.com. I am willing to help. Thank you for your consideration.
Observations and Concerns Regarding Election Systems and Processes in Delaware:
This attachment summarizes observations detailed in a two presentations that can be found on Delawarecounts.com In the two presentations (total 2+ hours) many items are surfaced but those related to the vulnerability of the system are summarized below for convenience.
1. Primary Results 2022: Absentee and Early Voting Ballot Vulnerability? In the presentations I argue that the Absentee Voting system is vulnerable to “stuffing” of the Absentee Ballot box. In item 9 and 10 below, I express concern that both Absentee and Early Voting ballots could be vulnerable to “Date Trigger Flip” type programming to effect fraud. In short a “Date Trigger Flip” algorithm is one that within the computer could flip for example every 15th vote for candidate A, to vote for candidate B. Such a “Flip” algorithm would not be turned on until a specific “Trigger Date” or time. The Trigger Date would be a date or time after which the DOE has already completed the Logic and Accuracy Tests (LATs) designed to approve the machines for use in the election.
If the two theories are above are correct could we see hints of such evidence in election data? Wouldn’t the concerns above be consistent with data that show that winners in elections have a higher percentage of their (Absentee + Early) votes vs that same data for their competitors? Said a different way, the more (Absentee + Early) votes an enemy or a cheater could get into the system the more they could flip votes, and when they flipped votes for a candidate A wouldn’t that skew that same data for candidate B?
We just had a primary election. If every primary contest had two candidates and if the occurrences were random and there were no other influences we would expect that 50% of the time the winner would have a higher percentage of (Absentee +Early) votes than their competitor and the other 50% of the time the winner would have a lower percentage of (Absentee + Early) votes than their competitor. Since some contests in the recent primary elections had more than 2 candidates, if my calculations are correct, the number is more like 43%. Said again we would randomly expect the winner of the contest to have a higher percentage of their votes being (Absentee or Early) votes about 43% of the time. If the number is less than that, it would not be consistent with the “stuffing and flipping” thesis. Of the 31 contests reported in the recent primary elections, the winners had a higher percentage of their own vote being (Absentee + Early) ballots 67% of the time. In the case of Democrat contests the number was only slightly higher at 69% and for Republican races the number was only slightly lower at 63%. Given the few small number of races and other factors this information should not in any way be presented as “proof” of any nefarious action but perhaps it suggests we should not dismiss such vulnerability concerns as unthinkable.
2. LAT Test Design: The State of Delaware Title 15, Section 5523 (b) (1) (a) indicate that Logic and Accuracy Tests (LAT’s) shall ensure that the voting system is properly programmed. The actual LATs conducted and witnessed for the absentee ballot scanners as performed by the State of Delaware (Department of Elections = DOE) are not capable of detecting even two of the most trivial examples of improper and coincidentally nefarious programming. The examples are presented in videos but for simplicity let’s call these “Date Trigger-Flip” programming for convenience. The State may not be in conformance of its own laws based on design of test. Subsequent to the test for the absentee scanners, I attended an LAT for the “in person” machines. The tests for those machines could not “catch” Date Trigger-Flip programming.
3. LAT Test Results: The actual LATs performed by DOE personnel on the absentee ballot scanners in preparation for the 5/10/22 school board elections revealed additional concerns regarding the system. The scanners could not read between 13-25% of the test ballots that were properly filled out by DOE personnel. DOE personnel postulated that the failure to read was because of some “light or flawed timing marks” on the perimeter of some of the printed ballots. I postulated that these flaws could be related to mechanical printer, or improper programming concerns but underlying source was not resolved as part of the LAT. A key point is that there was nothing that touched these ballots that were not part of the State of Delaware Election System and as a result one may interpret the failure to read as an error rate. The State of Delaware is not allowed to purchase equipment unless it is certified as meeting EAC and VVSG performance standards. VVSG standards require less than a .001% error rate. The state may be in a second non-conformance of its own laws based on actual performance of the tests.
4. Actual counting of ballots: School Board Election: The actual counting of ballots in the school board election included a “pre-screen” step. DOE experts inspected 100% of the ballots, not for count, but rather to see if there were any errant marks on the ballots placed there by the voter or other items that might prevent proper subsequent scanning of the ballot. If there were such issues those ballots would be removed from the population, “duplicated” and then the duplicate would be returned to the ballot population in preparation for scanning. The end result was that after pre-inspection there a stack of ballots that one would expect to run perfectly through the scanners without problems. The result? Even after pre-inspection and duplication, approximately 12% of the ballots could not be read by the system. If we include the ballots that were pre-screened we are back to a situation where up to 18% of the ballots could not be processed by the system and required duplication. These results are not inconsistent with the results of the LAT.
5. Actual Counting of Ballots: 2020 General Election. A ballot that could not be read by a scanner would require “adjudication”. Specifically, a group of humans would need to determine the intent of the voter for that ballot. In practice it was observed that ballots that could not be read by the scanner were “duplicated”. With this background, Patriots for Delaware submitted a FOIA request to the State of Delaware. They asked the State of Delaware how many ballots required “adjudication” in the 2020 election. The State of Delaware responded that the number was 47,205. If we make some qualifying assumptions regarding “write in” votes and also assume that the majority of issues were related to mail in or absentee ballots (“MIABs”) this again gets to the order of about 20% of the MIABs. That is similar to the magnitude observed in the LATs and the counting of the School Board Election ballots 1.5 years later.
6. No Audit: MIAB’s Scanner: In Delaware, in 2020 this class of scanner counts the highest density of ballots. By density I am referring to the highest number of ballots processed in the fewest number of locations and fewest number of machines. In Delaware in New Castle county that would have been about 93,000 +/-ballots in 1-2 office suites. As employed in Delaware, unless I am wrong, this class of machine is not subject to any type of post-election audit in the course of a normal election.
7. Efficiency: In Delaware the inability of the machines to be able to read voter intent on the first pass results on about 20% of the ballots, results in DOE human personnel doing 220% by hand, by eye inspections than they would if they just counted the ballots 100% by hand, by eye in the first place.
8. Update: Good and Bad News: Actual Counting of Ballots: I recently attended a session for counting of some of the Absentee ballots for the 2022 Primary election (9/12/22 3:00 PM). I watched two batches of actual ballots being processed through the MIAB’s. In one batch there was failure to read 3 of 90 or 3.3% of the ballots. In a second batch the scanner failed to process 4 of 116 ballots or 3.4% of the ballots. Upon discussion, DOE personnel informally indicated that again one concern was “bad timing marks” down the left side of the ballots. I was shown examples of a flawed ballots that wouldn’t run through the scanner. Again, informally a DOE employee indicated he believed that the problem had been associated with 1 printer and DOE had stopped using that printer at some point. That said it was not understood at this point what was causing the printer to print ballots with “bad timing marks”. I applaud the DOE personnel for partial diagnosis, and taking corrective action on an issue that appears to have been present at least since 2020. While taking some corrective action appears to have improved the most recent results over the prior results reported in items 2-4 above the results still show the system does not come close to achieving VVSG performance standards.
9. VULNERABILITY Summary Items 2,6: As a security measure the legislature built into the law a pre-condition to using certain machines. The legislature allows the use of machines to count our ballots based on the assumptive pre-condition that a test be performed (LAT) that can “ensure that the system is properly programmed”. The LATs the State employs don’t do that. I am even concerned whether it is possible to design a test that could “catch” variations of Date-Trigger-Flip programming. Given this reality perhaps the ONLY protection Delaware would have is human inspection either on a 100% or sample basis.
While Title 15 Section 5012A and the DOE recognize inspection solution this with respect to the “in person” “voting devices” (ES&S Express Vote XL Machines) I find it curious, and a grave omission that the Title 15 in effect has limited the definition of “voting device” to the in person voting machines. The DOE has also “in practice” excluded the MIAB’s scanners from the post -election random audit process. Let’s review that again for emphasis. THE LATs for the CLASS OF MACHINE THAT PROCESS THE HIGHEST DENSITY OF VOTES, IN THE FEWEST LOCATIONS, ON THE FEWEST NUMBER OF MACHINES, CAN NOT DETECT EVEN THE MOST TRIVIAL TYPES OF NEFARIOUS IMPROPER PROGRAMMING, AND IN LAW, AND IN PRACTICE THEY HAVE BEEN EXCLUDED FROM THE POST ELECTION AUDIT PROCESS. If I am correct and I as an individual member of the public with no special resources can observe this vulnerability what are the chances that any foreign state or domestic enemy with significant resources has not also recognized this vulnerability?
In videos I propose a hypothetical attack strategy a single individual might employ given the observations above. In this strategy, the primary purpose of the defective timing marks on Absentee Ballots would be simply to delay the process and afford the cheater the time needed to cheat elsewhere in the system. In a second element of the strategy the hypothetical enemy would employ Date-Trigger-Flip algorithms in machines that are not subject to post election audits.
10. EARLY VOTING PROCESS: DESIGNED TO AVOID THE AUDIT? The DOE has determined that for the early voting in the 2022 election it will use the standard ES&S Express Vote XL machines to mark ballots. This machine is the same style machine that is used in normal “in-person election day” voting. These machines are capable of both marking ballots, counting ballots, and then recording that information on media (a vendor proprietary thumb drive). This is how they are typically employed in Delaware Elections yet for early voting they are NOT going to be deployed in this manner.
In the 2022 election starting with the primary’s the DOE has decided to only have voters “mark” the ballots with this “in person” machine. At that point the ballots will be extracted from the machines, and then transported to a different location where they will be counted on the MIAB’s scanners. What an unfortunate coincidence for those concerned about vulnerability of our elections that the DOE has chosen to scan the Early Vote Ballots on the machines that are not subject to post-election audit.
11. Post-Election Audits: Express Vote XL machines: It has been confirmed by DOE officials (email) that the DOE does not have written Standard Operating Procedures (SOP’s) for the post-election audits. It follows that each audit could be performed differently each time and nobody would have any idea if it was being done correctly or not. It also follows that the only way to determine what is actually done is to attend such an audit. I have attended such an audit and discovered vulnerabilities. In an effort to show the importance of SOP’s, I postulated five examples for defeating the methods I observed being used at the post-election audit. Under section 5012A the DOE is required to post the audit results on their website. I am not implying, nor could I prove that any of the defeat strategies I proposed were employed. That said given the detail level provided in the audit reports I do not believe the DOE could prove they weren’t. I don’t believe that is a good situation. I believe with a few simple modifications to process it would be easy for the DOE to prove they could not have been employed. Even if the vulnerabilities are not being exploited we should move to eliminate them from our processes.
12. Solutions: There are ways to change the DOE processes to address many of the observations and vulnerabilities presented above. They range from things as simple as changing the room location and times when post-election audits are performed, to additional redundant in person inspection of the ballots either on a 100% or sample basis. There are many feasible solutions and improvements that can be implemented by the DOE under the authority of section 5522(b) of Title 15 and as a result no additional enabling legislation is necessary.
My Experience trying to interact with the Department of Elections and The Board of Elections
1. Interactions at Public Demonstrations:
Interactions with personnel at public tests such as the Logic and Accuracy Test’s (LAT) and the post-election audits have typically been good. Except for one recent incident individuals explained what they were doing and they answered the simple questions I could think to answer at the time.
2. Email Interactions: As I started to understand more of the process I started to ask more difficult and perhaps more sensitive questions. A few examples of what might be sensitive questions follow:
Do you agree that the LAT couldn’t catch these specific examples of improper code? (i.e Date Trigger Flip) type algorithms.
Are you allowed access to the underlying code in the machines, and if not, who is and can I get their contact information?
During the test it was indicated that “they” set up the machines, who was the “they”? Was it another DOE group, an outside vendor or someone else?
The response is that I was referred to the DOE link for FOIA requests. I clicked on the link. It didn’t work. I saved the screen shot of the error message. While it didn’t work, the error message directed me to a back-up email address (firstname.lastname@example.org).
I reviewed the FOIA process and looked at the example questions above and others and realized many were not appropriate for a FOIA request. I was not asking for documents, just opinions and directions to other experts. I made a second plea by email regarding the questions that I felt were not appropriate for a FOIA request. To my knowledge I never received a response from that email.
3. FOIA Request: In parallel to my second email request I submitted a FOIA request to the “back-up” email address and from that back up address I received an acknowledgement that the request had been forwarded to the FOIA coordinator for elections. That was in May 2022. To my knowledge I never received a reply of any kind to my FOIA request.
4. Section § 302 Procedure: I looked for additional means of contact. I found section § 302 of Title 15 (highlights mine).
§ 302. General powers and duties of Commissioner.
The Commissioner shall have the following powers and duties:
(13) To establish procedures to allow citizens to report possible violations of this title to the Commissioner, to any county director, or to deputy county director, which shall include procedures for anonymous reporting of possible violations thereto;
I thought that such a procedure might be appropriate since in some cases I am trying to report on “possible violations” to Title 15. I found something that I first thought might have been intended to meet the intent § 302. It was a pamphlet that was titled: State of Delaware Department of Elections: Administrative Complaints: “Following Federal and State Procedures”. There was similar language in the code. Upon review of this pamphlet I noticed that it did not allow for anonymous reporting, and did not focus on the possible violations of Title 15 but rather appeared to try to restrict the scope of concerns to those defined in the Federal Help America Vote Act of 2002. It would be somewhat ironic that if this were the intended to be the procedure required under §302, in that the procedure itself may be in violation of Title 15. In any case it appeared that the turnaround time for using this procedure was 90 days + 60 days which rendered it quite worthless in the context of the 2022 elections.
5. The Board of Elections (BOE) and Commissioner: Website: My next stop in trying to get information acknowledged was the BOE and the Commissioner. After all they are ultimately responsible for oversight and repair of any vulnerabilities or inefficiencies in our elections systems. If there are such concerns they would clearly want to know about them. As a citizen I would also expect them to have a positive attitude and a scientific with respect to investigating and vetting concerns. I would also expect them to be curious with respect to proposed solutions if someone had taken the time to not only raise concerns but also to actually propose solutions.
I found the link to the BOE https://elections.delaware.gov/aboutagency.shtml . While this had the name of Board Members there were no contact phone numbers or email addresses. I looked up the site for the meeting minutes and written agendas of the BOE. I was surprised to find that with all the attention and focus on “Election Integrity” that, at that time, the BOE had not met in 7 of the prior 12 months, and had not met in the past 4 months (March-June) straight. According to the notes they had not met due to board vacancies. That led to even more questions. Were candidates submitted in a timely manner? Does it just take that long to find them?
6. Board of Elections Virtual Meeting: As it turned out the BOE had announce a meeting for July. Great! An opportunity to get the message to both Republican and Democrat board members and the Commissioner. I called in and waited patiently for one hour and 23 minutes. I was then given 3 minutes. This is what the minutes of that meeting indicated that I said:
Public member Mr. Dan Lesher offered his email and phone number and requested the
emails and contact information of the Board members. Mr. Lesher voiced his concerns about
the error rates of voting devices and scanners and commented that he felt the
implementation of the voting system was not meeting standards.
My statement was a bit more detailed and I indicated that I had significant additional concerns and that the information I presented was only 1 portion of 1 concern. Upon completion Mr. Pasquale stated to Mr. Albence. “Anthony, I’ll let the department send whatever they want to send to the board members along with yourself” to which Mr. Albence replied “OK…” I am writing this on 9/18, and to my knowledge have had no contact from anyone, and have no confirmation as to whether Mr. Albence has sent anything to the BOE members at all or researched concerns at all.
From my perspective as a citizen it got worse. In a later discussion another member of the public expressed frustration with regard to not being able to contact the board without being filtered through the commissioner’s office. This was on a completely independent matter I believe from a member of the Libertarian party. They expressed that for them, the process had been problematic in the past. In a response to both a board member responded.
“this is a board that meets, and it is very carefully balanced and it doesn’t meet unless the balance is available, individual contact with any individual board member to bring up an issue, I don’t believe is appropriate”. Another board member of opposite political party agreed and indicated that if there were issues the public needs to bring up, bring them to the monthly meeting.
If that statement reflects the opinion of the board which I hope it does not I would like to make a few recommendations. That statement would imply that the only way anyone would able to communicate with BOE members, or that board members could communicate with each other is at the monthly meetings. If that is the case I would recommend immediately that the board meet at least monthly and preferably every two weeks. I would request that they meet even if there is not a quorum or even if the board is not at full strength. Why? While the board may not have its balance to discuss topics at least they are providing the public or others a means of presenting information that can be recorded in the minutes and get on the agenda and studied for when the board does get back to full strength. There may be things they need to know that can not wait 5 months. A second practical suggestion is that the minutes of the then current meeting should be posted to the public within a week of the meeting. Given recent history, it is not reasonable to have the public wait 4-5 months for approval of minutes at the next meeting to learn what was discussed at a prior meeting. Finally even if a full attendance of the board is not possible the DOE should provide status reports to the BOE. We the public and the BOE wouldn’t want to wait 4 months only to then learn then that plans and implementations for whatever reason are getting off track.
As a citizen I disagree with the reluctant attitude the two board members expressed regarding their desire to not communicating with anyone unless there was a balanced full attendance of the board. Does it follow from that logic that a board member shouldn’t research, investigate or do anything, or talk to anyone regarding DOE business unless there is a full balanced meeting of the board? That could be a neat trick. As a board member you could do absolutely nothing between board meetings and that logic would allow you to feel noble about not doing anything? I wonder if the other members of the board, and / or nominating chairpersons feel the same way and would agree with that attitude? I have served on boards. Perhaps not as prestigious as the BOE but I thought the time between board meetings was the time to research, investigate, and interact with people from many backgrounds to understand what is going on and to determine if there are any issues or points of view that should be added to the agenda or expressed in discussions when the board does get back together to meet. I might be a little out on a limb here but it might be possible that the Democrat members would want to make sure Democrat concerns are represented and Republican members might want to insure Republican concerns are represented and each might want to talk to people in the public and their respective parties between meetings to understand those concerns. I acknowledge and support the concept of “balance” but would argue that the “balance” is achieved by the legal required 5-5 split of the board members by political party. I would also posit that if members of one party think they shouldn’t talk to people between meetings because the balance is not there, and they are under the assumption that the other party isn’t doing that either, they may want to reconsider that assumption.
I felt a little disappointed after my first BOE meeting but I am not discouraged. In fairness I imagine phone calls such as mine in the past have not been the norm. Like many I took our election processes for granted. In the last 1.5 years myriads of evidence and data have surfaced regarding elections and election processes and it is hard to keep up. Even given frustrations I vent here, I am still optimistic that if the BOE and DOE become aware of the vulnerabilities they will want to both interact, and act.