Statistical Anomalies in LA Mayoral Election: A Deeper Analysis June 7, 2026
Posted by Chris Mark in Industry News, Laws and Leglslation, News, Politics, Uncategorized.Tags: music, News, poetry, politics, writing
add a comment
DISCLAIMER: This article presents a statistical analysis of publicly available election data. It does not allege fraud, illegal conduct, or wrongdoing by any candidate, election official, or government entity. The statistical anomalies documented below demand transparent explanation. That is the appropriate standard in a functioning democracy. Nothing more is claimed here.
Introduction
Elections in the United States are decided by votes. The integrity of those votes depends not only on the honesty of those casting them but on the transparency and consistency of how they are counted. When the statistical profile of mail-in ballot counting diverges from election day results by a margin that falls outside any reasonable probability model, the public interest demands a clear and documented explanation.
This article presents a statistical analysis of Spencer Pratt’s performance in the 2026 Los Angeles mayoral primary election. The analysis compares his election day vote share to his performance in subsequently counted mail-in ballot batches. The divergence between these two data sets is not a matter of opinion or political interpretation. It is a mathematical fact that warrants examination.
This is not an endorsement of any candidate. It is an application of basic statistical principles to publicly available election data.
Background: The Race
The 2026 Los Angeles mayoral primary featured fourteen candidates, with incumbent Mayor Karen Bass seeking a second term against a field that included former reality television personality Spencer Pratt, a registered Republican whose Palisades home was destroyed in the devastating 2025 wildfires, and Los Angeles City Councilwoman Nithya Raman, a Democratic Socialists of America member challenging Bass from the left. [1]
Under California’s election rules, if no candidate receives more than fifty percent of votes in the primary, the top two candidates advance to a November runoff election. Mayor Bass secured enough votes to advance. The race for second place — and the November runoff slot — became a contest between Pratt and Raman. [2]
A pre-election UC Berkeley-LA Times poll conducted in May 2026 showed Bass with twenty-six percent support, Raman at twenty-five percent, and Pratt at twenty-two percent among likely voters — a margin of error of approximately three percent. [3]
Figure 1: Election Night vs. Mail-In Ballot Performance — LA Mayoral Race 2026
Election Night Results
Pratt significantly outperformed his pre-election polling. With sixty-six percent of the expected vote counted on election night, results showed:
Karen Bass: 35% Projected to advance to November runoff
Spencer Pratt: 29.4% Comfortably in second place
Nithya Raman: 23.4% Trailing Pratt by approximately six percentage points
Pratt held what appeared to be a comfortable lead over Raman. By Thursday, with additional votes counted, the gap remained near six percentage points. [4]
With 163,549 votes in Los Angeles’ latest tabulation, Pratt maintains a near 6% lead on Raman, who has 130,473 votes. — Fox News, Thursday June 5, 2026 [4]
The Mail-In Ballot Divergence
As mail-in ballot batches were counted and released in the days following the election, a striking divergence from election night results emerged. Rather than tracking the established proportions, the mail-in batches showed a dramatic and statistically extraordinary shift.
The Zero-Vote Batch
The initial anomaly identified was a batch of approximately 24,000 mail-in ballots in which Pratt received zero votes. At his election night rate of 29.4 percent, the expected number of Pratt votes in such a batch would be approximately 7,056.
Probability of zero Pratt votes in 24,000 ballots at 29.4%: 1 in 10^3,629 Effectively impossible by random chance
For context: the total number of atoms in the observable universe is estimated at approximately 10^80. The probability of Pratt receiving zero votes in that batch, if his actual support rate was 29.4 percent, is incomparably smaller than randomly selecting one specific atom from the entire universe on the first attempt.
The Subsequent Batch Analysis
Examining the larger batch of mail-in votes reported since Thursday — totaling 54,245 votes across Pratt, Raman, and Bass — the divergence becomes statistically quantifiable. [5]
Pratt mail-in share: 19.7% vs. 29.4% election night — deficit of 9.7 percentage points
Raman mail-in share: 42.6% vs. 23.4% election night — gain of 19.2 percentage points
Pratt vote deficit: 5,237 votes Below statistically expected count in this batch alone
In concrete terms: if mail-in ballots had simply reflected election night proportions, Pratt would have received approximately 15,948 votes in the analyzed batch. He received 10,711 — a shortfall of 5,237 votes in a single counting batch.
Statistical Analysis
The Chi-Square Test
The chi-square test measures whether an observed distribution of votes differs significantly from what would be expected based on a reference distribution — in this case, election night proportions. Applying this test to the mail-in batch:
Chi-square statistic: 10,376.18 Extraordinarily high — any value above 6 is statistically significant at the 95% confidence level
Degrees of freedom: 2 Three candidates minus one
P-value: Effectively zero The probability this divergence occurred by random chance
A p-value of zero means the observed distribution of mail-in votes cannot be explained by random sampling variation from the election night population. Under standard statistical thresholds, a p-value below 0.05 is considered statistically significant. A p-value below 0.001 is considered highly significant. This result is not in that range — it is below any threshold that statistical science has developed to describe.
The Z-Score Analysis
The z-score measures how many standard deviations an observed result falls from its expected value. In normal human affairs, results beyond three standard deviations are considered extraordinary and warrant investigation. Results beyond five standard deviations are considered essentially impossible by random chance.
Z-score for Pratt’s mail-in performance: -49.35 Forty-nine standard deviations below his election night rate
A z-score of negative forty-nine does not exist in the normal range of human statistical experience. To find a naturally occurring phenomenon with a z-score of this magnitude would require examining astronomical datasets, not election results. This number is not an anomaly. It is a mathematical impossibility under any standard probability model that assumes mail-in voters come from the same population as election day voters.
In statistics, anything beyond three standard deviations is considered extraordinary. Forty-nine standard deviations is not a number that occurs in nature through random variation.
The Current State of the Race
The cumulative effect of these mail-in batches has been dramatic. [6][7]
Pratt current share (78% counted): 27.3% Down from 29.4% election night
Raman current share (78% counted): 26.2% Up from 23.4% election night
Current Pratt lead: Approximately 7,500 votes Narrowing with each batch
Raman received forty percent of votes counted on Saturday — a figure that, if sustained, would be sufficient to overtake Pratt before all ballots are counted. [7]
The race remains uncalled. California law allows counties up to thirty days to complete the official canvass. Millions of mail-in and provisional ballots remain to be processed in Los Angeles County alone — the largest voting jurisdiction in the United States, with 5.8 million registered voters. [8]
Three Possible Explanations
Statistical analysis identifies the anomaly. It does not, by itself, determine the cause. There are three explanations that must be considered:
Explanation One: Population Differences
California leads the nation in mail-in voting, with eighty-one percent of voters sending their choices by post in 2024 — nearly double the national average. [9] It is theoretically possible that Pratt’s support is concentrated among voters who specifically chose to vote in person on election day, and that mail-in voters skew heavily toward Raman and Bass.
However: even accepting significant population differences, a forty-nine standard deviation divergence cannot be explained by population variation alone. The pre-election poll showing Pratt at twenty-two percent among likely voters — not a dramatically different figure from his election night performance — did not distinguish between mail-in and in-person likely voters in a manner that would predict a divergence of this magnitude.
Explanation Two: Counting Methodology or Batch Composition
It is possible that specific batches of mail-in ballots being counted represent geographically concentrated areas where Raman has disproportionate support — council districts she represents, for example — and that the batches are not representative of the overall mail-in population.
If this is the explanation, the Los Angeles County Registrar-Recorder should be able to document precisely which geographic areas each batch represents and demonstrate that the composition explains the divergence. That documentation should be made public.
Explanation Three: Something Requiring Investigation
The third possibility is that something in the counting or reporting process is producing results that do not accurately reflect the votes cast. This article does not allege this is the case. However, the statistical evidence is sufficiently extreme that it cannot be dismissed without documented, transparent explanation of the first or second type.
What Transparency Requires
In a functioning democracy, election results that produce statistical anomalies of this magnitude demand documented explanation — not reassurance, not dismissal, but transparent accounting of the counting process. Specifically:
The Los Angeles County Registrar-Recorder should publicly document the geographic composition of each mail-in batch released since election day — demonstrating which precincts or council districts each batch represents and how that composition accounts for the observed divergence.
The methodology for selecting, processing, and releasing mail-in ballot batches should be made publicly available.
Any candidate or party requesting observation of the counting process should be granted that access consistent with California election law.
The zero-vote batch — 24,000 ballots producing zero votes for a candidate receiving approximately 29.4 percent of all other votes — requires specific and documented explanation.
The appropriate response to a statistical anomaly in a democracy is transparency and documentation — not political dismissal or reassurance. The numbers are what they are. They deserve a clear answer.
Conclusion
Spencer Pratt received approximately 29.4 percent of votes cast on election day in the Los Angeles mayoral primary. In subsequently counted mail-in ballot batches, he has received approximately 19.7 percent — a divergence of 9.7 percentage points that produces a z-score of negative forty-nine and a chi-square statistic of over 10,000.
These numbers are not consistent with random sampling variation from the same voter population. They are not explained by normal statistical fluctuation. They demand a documented, transparent, and geographically specific explanation from Los Angeles County election officials.
The question is not whether Spencer Pratt should be the next mayor of Los Angeles. The question is whether the vote count accurately reflects the votes that were cast. In a democracy, that question is never inappropriate to ask — and it is always appropriate to demand a clear answer.
Chris Mark is an Enterprise Security and Risk Strategist, published author, co-author of PCI DSS, named patent holder, and United States Marine Corps combat veteran. He writes on security, risk, and emerging threats at GlobalRiskInfo.com.
References
[1] NBC News. (2026, June 2). Los Angeles Mayor Primary 2026 Live Results. nbcnews.com/politics/2026-primary-elections/los-angeles-mayor-results
[2] ABC7 Los Angeles. (2026, June 4). Los Angeles mayor race: Live election results and updates on front runners Karen Bass, Nithya Raman, Spencer Pratt. abc7.com
[3] CBS Los Angeles. (2026, June 7). Pratt’s lead over Raman slims in new L.A. mayoral election results. [Citing UC Berkeley-LA Times poll, May 28, 2026, margin of error approximately 3%.] cbsnews.com/losangeles
[4] Fox News. (2026, June 5). Spencer Pratt loses ground to Democrat while Hilton maintains lead in latest California ballot batch drop. foxnews.com
[5] Fox 11 Los Angeles. (2026, June 6). LA mayor’s race: Nithya Raman surges, closes gap on Spencer Pratt for runoff spot. foxla.com. [Reporting Raman: 23,115 votes (38%), Bass: 20,419 votes (34%), Pratt: 10,711 votes (18%) in mail-in batch since Thursday.]
[6] CBS Los Angeles. (2026, June 7). Pratt’s lead over Raman slims in new L.A. mayoral election results. cbsnews.com/losangeles. [Citing 78% of votes counted, Pratt 27.3%, Raman 26.2%.]
[7] The Wrap. (2026, June 7). Nithya Raman Inches Within 1% of Spencer Pratt After Winning 40% of Saturday Tally in LA Mayor’s Race. thewrap.com
[8] NBC Los Angeles. (2026, June 6). Gap between Pratt and Raman gets tighter in LA mayoral race. nbclosangeles.com. [Noting 5.8 million registered voters in Los Angeles County.]
[9] Fox News. (2026, June 5). Spencer Pratt loses ground to Democrat. [Citing California leads nation in mail-in voting at 81% of voters in 2024, nearly double national average of 43%.]
[10] Statistical methodology: Binomial probability calculation P(X=0) = (1-p)^n. Chi-square test comparing observed mail-in distribution to election night baseline. Z-test for proportions: z = (p_observed – p_expected) / sqrt(p_expected*(1-p_expected)/n). All calculations performed using Python scipy.stats library.© 2026 Chris Mark / GlobalRiskInfo.com. All rights reserved. Reproduction with attribution
MY LATEST BOOK RELEASED! “The Science of Security” May 16, 2026
Posted by Chris Mark in cyberespionage, cybersecurity, Industry News, InfoSec & Privacy, Laws and Leglslation, Piracy & Maritime Security, Risk & Risk Management, security, security theater.Tags: ai, artificial-intelligence, cybersecurity, data breach, History, InfoSec, Maritime Security, philosophy, Piracy & Maritime Security, risk management, security, technology
add a comment
Announcing Scientia Securitatis: The Science of Security
After 34 years across nearly every security domain that exists — armed physical security at an overseas critical installation, combat force protection, security in a regional hospital’s psychiatric ward, payment-card industry compliance, armed maritime contracting off the East African coast, and a return to enterprise cybersecurity that has occupied the past decade — I have written the book I wish someone had written when I started.
Scientia Securitatis: The Science of Security — Theory, Frameworks, and Practice is available now.
The gap this book is intended to fill
The security profession does not lack books. Walk into any bookstore, scan any conference vendor floor, search any retailer’s security category, and you will find more material on cybersecurity, physical security, risk management, military theory, criminology, intelligence analysis, and organizational resilience than any single practitioner could read in a career. The field is overwhelmed with information.
What it lacks is integration.
Each security domain has developed its own vocabulary, its own frameworks, its own bestsellers, its own consultants. Each domain — when traced carefully to its analytical roots — is reaching for the same underlying concepts the next domain over named differently. Practitioners in physical and cybersecurity are working on the same analytical problems and rarely speak to one another. When they do, they discover that they have been duplicating each other’s work for decades.
Scientia Securitatis is an attempt to make that recognition the starting point of professional practice rather than an accident a few practitioners stumble into late in their careers.
What’s in the book
The book runs to 525 pages across 11 chapters and three appendices. It develops four original analytical frameworks:
- The Mark Heptad — a taxonomy of seven adversary motivations (financial, espionage, war/defense, facilitation, hacktivism, revenge, nuisance) that maps directly to deterrence strategy
- The IMCM Framework — Ignorance, Mistake, Complacency, Malice — for classifying human-induced vulnerabilities and matching them to specific interventions
- The DIVE Framework — Direction, Intensity, Vulnerability, Exposure — for assessing specific exposure surfaces
- The Multiplicative Security Model — the mathematical basis for defense-in-depth, with implications for how security architecture should actually combine
These original frameworks sit within a broader analytical apparatus drawn from criminology (Cohen and Felson’s Routine Activity Theory, Cornish and Clarke’s Twenty-Five Techniques of Situational Crime Prevention), cognitive science (Kahneman and Tversky on judgment under uncertainty), military theory (Sun Tzu, Clausewitz, contemporary unrestricted warfare doctrine), and systems-safety scholarship (James Reason’s Swiss Cheese Model, Charles Perrow’s normal-accident theory).
The book also examines — and critically engages — the victim-blaming reflex that dominates post-incident analysis, drawing on the foundational criminological literature on victim precipitation and contemporary case studies including Equifax, OPM, Target, and Snowflake.
A note on the Latin title
Scientia Securitatis translates as “the science of security,” and the choice was deliberate. The Latin signals that the book engages security as a serious analytical discipline whose intellectual roots long predate the cybersecurity industry’s tendency to treat its problems as historically unprecedented. The phenomena security examines are ancient; the framework for studying them rigorously has been available since at least the mid-20th century. The book argues that practitioners have, with rare exceptions, declined to use it.
Who this book is for
This book is for the practitioner who has noticed that decades of escalating security investment have not produced proportional security gains, and who wants to understand why. It is for the security executive building defensible programs across multiple domains. The policy professional confronting unrestricted warfare doctrine. The risk and compliance leader who suspects that frameworks alone are not stopping sophisticated adversaries. The graduate student approaching security as an analytical discipline rather than a job category.
It is not a tactical handbook. It is not a configuration guide. It is the analytical apparatus that determines whether tactical choices are well-made — the apparatus the field has been operating without.
Where to get it
Scientia Securitatis: The Science of Security is available now on Amazon in eBook, paperback, and hardcover formats:
If you find the book useful, please consider leaving a review. Self-published analytical nonfiction lives and dies by word-of-mouth among the practitioners it was written for — and a thoughtful Amazon review from a working professional is worth more to other professionals than any amount of marketing.
— Chris Mark
Autocracy, Anocracy, & Democracy – “Verbal Masterba(bleep!)…” January 25, 2016
Posted by Chris Mark in Laws and Leglslation, Politics.Tags: anocracy, autocracy, Chris Mark, democracy, Dr. Heather Mark, facebook, mark consulting group, politics
3 comments
Election season in the US is always interesting. Passions run high and people are quick to proclaim their positions on government and politics. Unfortunately, as many will likely agree, election season also gives voice to many who should probably remain silent.
Recently I was taken to task on Facebook and lectured on the concept of governance and democracy by a particularly obtuse and offensive individual. When I attempted to explain that democracy should NOT be considered a strictly binary proposition and that the US was indeed a democracy, his attacks became personal and I was accused of (among other things) “verbal masturbation”. According to this master of the English language: “Most folks like me would call your ideas verbal masturbation. They sound good from the outside but are really kinda stupid”…he actually wrote: “Kinda”…somehow this person drew a line between my comments on democracy and his belief that the federal government would force parents to stand by while their 12 year old daughters got abortions without consent. I am at a loss as to the logic… But…I digress. Back to democracy! (more…)
Mobile Privacy October 12, 2012
Posted by Heather Mark in InfoSec & Privacy, Laws and Leglslation, privacy.Tags: Dr. Heather Mark, Heather Mark, Location Privacy Protection Act, mark consulting group, mobile privacy, privacy
add a comment
Smartphones have changed the way we interact with our world. They’ve introduced a new level of convenience, but they’ve also introduced a new potential threat to our privacy. As consumers, we should be informed about the choices that we make on our smartphones and how they might impact us. For example, I upgraded my iPhone to iOS 6 this afternoon. (I know. I’m a little late on that one.) Anyway, when I was done I got two prompts. The first asked if I wanted to enable location services. I said yes, knowing that meant that 1) I could use the “find my phone” app, as well as many other apps that come in handy for a frequent traveler, and; 2) that it meant that Apple would have access to my location data. The next prompt suggested that Apple could improve its products and services if I just allowed my phone to send occasional reports to headquarters. That one I declined. I don’t necessarily want Apple to have access to all of my activities on my smartphone.
Now, I’m not naive enough to believe that my simple selection means that I have safely secured my data and mobile behavior entirely. There are companies that are taking advantage of the fact that privacy laws have not kept pace with technology. We know for example, that there are companies that offer device fingerprinting services for fraud prevention that also happen to sell mobile device behavior analytics to marketers. Consumers don’t have any way of knowing that their behavior is being tracked and they have no way to opt out.
This week, Sen. Franken (D-Minn) and Sen. Blumenthal (D-Conn) introduced a bill designed to protect mobile privacy. The Location Privacy Protection Act of 2011 is meant to protect consumer privacy by informing users of how and with whom their location data is shared. There are four primary requirements of the bill. Distilled to their basics, those requirements are:
1) Gain consumer consent before collecting location data
2) Get consumer consent before sharing that data
3) Assist in understanding and investigating crimes that involve the misuse of location data
and
4) create criminal penalties for those that abuse location services or use so-called “stalking apps.”
While I applaud the move to ensure that mobile users are protected from entities divulging their location without the knowledge or consent of the consumer, I wonder if the law goes far enough in protecting consumer privacy. What about those device fingerprinting activities? Do you think the proposed bill goes far enough? Too far? What would you like to see in terms of mobile privacy protection?
Privacy, Social Media, and Legislation September 29, 2012
Posted by Heather Mark in InfoSec & Privacy, Laws and Leglslation.Tags: AB1844, California, Dr. Heather Mark, Heather Mark, mark consulting group, markconsultinggroup.com, privacy, SB1349, Social media
add a comment
This week marks the opening of a new chapter in the rocky marriage of privacy and social media. California has passed two laws related to the protection of privacy on social media platform.
In SB1349, the state prohibits public or private post-secondary educational institutions from requiring students to provide the organization with access to the student (or student groups) social media sites. Nor can the student or group be forced to divulge information contained on those sites.
AB 1844 is similar in nature, but applies to employers. Specifically, the bill “would prohibit an employer from requiring or requesting an employee or applicant for employment to disclose a username or password for the purpose of accessing personal social media, to access personal social media in the presence of the employer, or to divulge any personal social media. This bill would also prohibit an employer from discharging, disciplining, threatening to discharge or discipline, or otherwise retaliating against an employee or applicant for not complying with a request or demand by the employer that violates these provisions.”
These bills are interesting in that they address a core concern around privacy and labor laws as they relate to social media. Employers (and potential lenders) are prohibited from making decisions based upon race, gender, religion, politics, sexual orientation. Most of this information, though, is available on individuals’ private social media profiles. Amid increasing reports of employers requiring prospective employees to turn over credentials or access their sites in view of the employer, privacy advocates were becoming increasingly, and rightly, concerned that the rights of individuals to protect their personal lives from employers were being diluted. These actions on the part of California serve to protect those rights. Frankly, these actions can also protect employers and schools from being accused of discriminatory behavior by not providing them access to this information, which would otherwise be unavailable to them.
It will be interesting to see how quickly other states follow the lead that California has set. Recall that California was the first state to pass a breach notification law and we now have 46 such laws nationwide. So the question, to me, is when, not if, we are going to see the trend take shape.
Global Security, Privacy, & Risk Management 