September 13, 2023


The Editor


A post-election audit report for the November 2022 election was hidden by Secretary of State Maggie Toulouse Oliver and her office for eight months. The report, dated January 16, 2023, is called a “Risk Limiting Audit” (RLA). The report was only posted by the SOS to her website in the last week of August after being pressured by grassroots researchers to release it. The RLA was written by Zlotnick and Sandoval, PC and it shows massive errors in the tabulation of write-in votes and errors 64 times higher than the legal limit for the election overall. The SOS was warned by multiple sources before the 2022 Midterms that the machines were not counting write-in votes correctly, but she ignored them.


The RLA process is done on a limited number of races following general elections. Dice are used to “randomly” choose a small number of precincts prescribed by statute. Ballots from those precincts are then collected and run through a Dominion tabulator. The same ballots are also hand counted. The machine and hand counts are compared, and the results are compiled in a report.

According to statute, only federal and statewide offices are ever “audited.” No down-ballot races are ever checked for accuracy. The machines could be programmed to miscount school board, sheriff, district judge, and county commission races and the RLA would never catch it.

The “audit” isn’t a true audit because the procedures are under the full control of the SOS and the “auditors” take no responsibility that what they report has any meaning or is sufficient to catch issues with the election if they are present. From the RLA authors:

We were not engaged to and did not conduct an examination or review engagement, the objective of which would be the expression of an opinion or conclusion, respectively, on the accuracy of voting system machine tabulators used in the conduct of the 2022 General Election held on November 8, 2022. Accordingly, we do not express such an opinion or conclusion. Had we performed additional procedures, other matters might have come to our attention that would have been reported to you.


In other words, the accountant that did the “audit” followed the instructions given to them by the person they were auditing. They were not free to investigate anything else or make any conclusions.


There were two races with write-in candidates included in the RLA for the November 2022 election: US Representative District 1 – with write-in candidate Victoria Gonzales, and US Representative District 2 – with write-in candidate Eliseo Luna. To vote for a write-in candidate, voters must fill out the oval for the write-in candidate and write the candidate’s name in the box provided.

The RLA showed massive discrepancies between the hand counted votes for write-in candidates versus those counted with machines. The RLA checked only 5 precincts statewide for the US Representative District 1 race. The Dominion machines undercounted Victoria Gonzales’ votes by 33 percent.

The RLA checked only 109 precincts statewide for the US Representative District 2. The Dominion machines undercounted Eliseo Luna’s votes by 200 percent.

Since the SOS tightly controls the auditor’s methods for investigating the election results, this gross error in counting write-in votes was not even flagged in the RLA report. The total number of votes missed or flipped by the Dominion machines for all the races were simply lumped altogether and the error rate was then calculated on the total number of votes tallied during the RLA irrespective of the individual races or the presence of write-in candidates.

Since the number of write-in votes was small compared to the number of votes cast for traditional candidates, the gross error being made by machines was covered up by the accountant who was following directions from the SOS.


Getting on the ballot under one of the major parties can be expensive and time-consuming, requiring collection of thousands of signatures and a pre-primary process. Getting on the ballot as a declared “write-in” candidate only requires filing a declaration of intent. The traditional path to candidacy is typically reserved for those within an elite club of insiders with access to ample funding and campaign staff. The ability to declare a write-in campaign is viewed as an opportunity to bypass the major parties and give candidates with grassroots support an opportunity to become elected officials.

In 1980, New Mexico elected U.S. Congressman Joe Skeen as a write-in candidate. The Democrat Attorney General at the time was trying to prevent Republicans from putting a candidate on the ballot after the Democrat incumbent died in office just a few months before the election, but he was going to allow the Democrats to put up a candidate to replace the man who had died – guaranteeing the seat went to a Democrat. The write-in process allowed Skeen to bypass this corruption and give New Mexicans a choice in 1980. Joe Skeen ended up staying in Congress until he retired in 2003.

A more recent write-in candidate success story can be found in the 2022 Primaries when a candidate running for Public Education Commission, Stewart Ingham, was able to run for an unopposed seat as a write-in candidate. He received enough votes during the primary to be listed on the November 2022 ballot as a regular candidate. Ingram is now the Public Education Commissioner for District 6.

As a political outsider, Ingham could not afford a traditional campaign. He had been fired from Albuquerque Public Schools for refusing to submit to weekly COVID tests. This won him significant grassroots support from thousands of parents who are upset with public school COVID policies they viewed as causing harm to their children. These parents wanted to see the Public Education Commission exert more authority over public schools to prevent tyrannical overreach, and now they have a representative in the room where decisions are made.

The SOS, who is not quiet about her contempt for the public, will likely argue that grossly undercounting the number of votes received by the two write-in US Representative candidates in November 2022 doesn’t matter since it didn’t affect the outcome of the election. But the people of New Mexico who have used this tool in the past to overcome the corrupt establishment would disagree.


The Otero County Audit of the November 2020 election published in August of 2022, noted many problems with the Dominion tabulators – one of which was they were not programed to deal with write-in votes correctly and ended up counting many votes that were properly cast for a single candidate as “overvotes.” Overvotes are where a voter selects too many choices in a race, their intent can’t be determined, so their vote is not counted at all. This same error could be what plagued the November 2022 election.

Also, under pressure from the public, the Torrance County Commission audited their own 2022 Primary election. They found that their Dominion tabulators made errors up to 25 percent in some Republican races.

Instead of doing further investigation into the cause of the error – which may have had to do with write-ins – the Torrance County Commission attempted to placate the public with a promise to do a thorough check of the canvass for the future November 2022 election before they certified it. As the Estancia News reported previously, this “thorough check” ended up being an indifferent observation by an SOS employee of election results paperwork getting printed from internet-connected software called SERVIS. It turns out that, at least in some counties, the clerks do very little cross checking of election documents, and instead rely almost entirely on output from SERVIS, which is under the control of the SOS and a shady company called KnowInk (formerly BPro).

In addition to these two county-wide audits by two different teams pointing to Dominion tabulators’ inability to properly count write-in votes, a write-in candidate running for State Land Commissioner named Larry “Lead” Marker sued the SOS before the November 2022 election. Among several issues cited in Marker’s lawsuit was an allegation that the Dominion tabulators were set up to prohibit a write-in candidate from winning an election. Marker alleged that the SOS was essentially preventing grassroots, write-in candidates from winning because of the equipment she is forcing counties to use. It looks like Marker was exactly right.


This story highlights that even when the data gathered during an RLA finds a significant error, that error will be covered up and will never be dealt with. But the fact is that RLAs themselves are deeply flawed tools the way they are being applied.

The RLA is frequently touted by the SOS as proof that her election machines are accurate. The trouble with that claim is that the audit itself relies heavily on machines and is no more transparent than the election itself.

Possibly the biggest red flag aside from the very limited number of races that are ever checked is how the precincts are chosen for the RLA. After the 2022 Midterms, a 4 ½ hour dog and pony show was conducted using dice and a laptop. The process can be watched here and here.

The way it works are seven dice are rolled to come up with a seven-digit number. This number is then punched into a computer and the software declares which County and Precinct that seven-digit number goes with. This is called this “normalization.”

Using seven, six-sided dice results in about 6,700,000 different numbers that could possibly be rolled. But New Mexico only has about 2,200 precincts. This means that only four dice are really needed to come up with enough possible rolls to match the number of precincts in the state. So why use seven dice? Is it to make the process so cumbersome that it forces the use of software to “normalize” the dice roll into an actual county and precinct?

Based on the number of possible dice rolls and the number of precincts, approximately 3,000 possible dice rolls would be assigned to each precinct. When the dice are rolled, the laptop operator plugs the number in, and the computer program finds that number and reports which county and precinct it was supposedly assigned to ahead of time.

How precincts are matched by the SOS’s secret laptop program to the possible dice rolls is never published, so the public cannot confirm that precincts are, in fact, really being chosen by random dice rolls. In fact, the evidence suggests that they are not really being randomly chosen at all.


The chances of selecting any one of the 2,200 precincts in any given dice roll in New Mexico should be 2,200 divided by 6,700,000 or 0.0003%. In other words, the likelihood any precinct is chosen is extremely small and we would not expect any precinct to be chosen more than once during the precinct-choosing exercise.

However, Bernalillo County Precinct 307 was chosen a total of 3 times, Bernalillo County Precinct 620 was chosen 7 times, and three precincts in a row were chosen from Otero County – which has only 66 precincts. Some counties were never called at all.

In all – it took over 500 rolls of the dice to come up with 137 precincts because there were so many doubles. This suggests that a huge number of precincts were either removed from the pool or the precincts were dishonestly weighted inside the secret program on the SOS’s laptop – which is exactly what a bad actor would do if they wanted to avoid precincts where they knew there was election fraud.

At the 1 hour and 4-minute mark in the video linked above, the computer running the normalization software “froze up” and they decided to shut down the live feed of the video, while they got it working.

A program that is simply supposed to return one of 2,200 values based on a seven-digit number could be written in an hour on an excel spreadsheet. The idea that the program made the computer crash is absurd. The need to turn off the live feed while a laptop got rebooted is even more absurd.

Clearly, whatever was going on inside that computer program was not random, not honest, and not transparent.


The RLA was invented by Philip Stark and Mark Lindeman who intended them to be performed within very specific parameters and with limited promises of their ability to find inaccuracies in elections as described in this paper. Stark worked with a leftist group called Verified Voting, who partnered with election machine vendors to promoted RLA’s to jurisdictions across the country.

However, Stark was too honest for Verified Voting and the election machine vendors, and he resigned from his position in 2019 after submitting a scathing public letter. Stark accused Verified Voting and the election machine vendors of improperly applying risk limiting audits, failing to establish that the paper trail used as a basis for RLA’s was trustworthy to start with, and lying to voters and election officials that RLAs proved they could have confidence in their elections:

Whitewashing inherently untrustworthy elections by overclaiming what applying RLA procedures to an untrustworthy paper trail can accomplish sets back election integrity. This is security theater…[Verified Voting] should be demanding convincing evidence that the paper trail is trustworthy…that requires a compliance audit that checks the chain of custody of the paper, ballot accounting, eligibility determinations, signature verification, among many other things. Verified Voting is providing cover for bad actors (election officials and vendors) and inherently untrustworthy voting systems…


Setting aside the serious issue of grossly incorrect tabulation of write-in votes, the maximum overall error reported in the SOS’s RLA for the November 2022 election was 0.051 percent. The maximum error rate allowed by state and federal law is 0.0008 percent, meaning tabulation errors in New Mexico are 64 times higher than allowed by law, even using the SOS’s whitewashed, self-directed, tightly controlled “audit.” According to state statute, the machines must be decertified as a matter of law if they do not meet accuracy requirements.

The Estancia News reached out to the SOS’s office to inquire why the RLA from the November 2022 election was hidden from the public for eight months and what the SOS’s office has done to investigate the gross errors in the races with write-ins. We have received no response.

Written by The Editor