TEXAS ELECTIONS UNDER THE MICROSCOPE: MBT EXPANDS FORENSIC INVESTIGATION INTO 2026 VOTE COUNTS, VOTER REGISTRATION FAILURES, AND THE 2020 AUDIT
Mind Bend Theory has identified numerical discrepancies, examined historical election-audit findings, and submitted formal public-records requests to Texas agencies. The investigation is expanding to determine what happened, whether any votes were affected, and whether the evidence reveals recurring failures.
MIND BEND THEORY | INDEPENDENT RESEARCH & FORENSIC ANALYSIS
Investigation Update — November 8, 2026
THE INVESTIGATION IS GETTING BIGGER
What began as an examination of unusual election-night reporting figures has developed into a broader investigation of Texas election administration.
Mind Bend Theory (MBT) is examining the November 3, 2026 election alongside documented voter-registration software failures, historical election-audit findings, citizenship-verification procedures, and allegations submitted for independent review.
Our investigation now reaches across multiple systems: voter-registration applications, Texas Department of Public Safety transactions, the state's TEAM registration platform, county election reports, provisional ballots, and the original records used to reconcile votes.
We are not simply asking who won. We are asking whether the numbers can be independently reproduced, whether the systems functioned correctly, and whether every discrepancy has a documented explanation.
Some of what we have identified is confirmed. Other findings remain unresolved. We will not present an unexplained discrepancy as proof of election fraud.
FINDING 01: A 1,079-RECORD DIFFERENCE IN TARRANT COUNTY
One of the clearest discrepancies identified in our review involves two different representations of Tarrant County turnout.
The county's November 4 cumulative election report lists 733,287 ballots.
A separate analysis using county voter-history roster records reports 732,208 voters.
Difference: 1,079 records.
The discrepancy is not evenly distributed across voting methods:
Voting method
Difference, county report minus rosterAbsentee/mail−573
Early in person +1,357
Election Day +295
Net difference +079
These figures do not demonstrate 1,079 illegal or missing ballots. Roster records and cumulative ballot reports can differ because of timing, classifications, or reporting procedures.
But the difference is measurable, and MBT is treating it as an open reconciliation question.
FINDING 02: THE 417-BALLOT DIFFERENCE
Our examination of Tarrant County's election-night reports identified another numerical pattern.
The final unofficial U.S. Senate report accounts for 733,287 ballots, while several Texas statewide contests account for 732,870.
Difference: 417 ballots.
Further examination localized that difference to the absentee reporting category.
One possible explanation is that certain absentee voters were eligible to participate in federal contests but not the state contests.
That explanation has not yet been independently verified.
MBT is seeking the underlying ballot-style and reconciliation records needed to determine whether the difference is legitimate and fully accounted for.
FINDING 03: ELECTION-NIGHT TOTALS CHANGED AFTER THE BALLOT COUNT STOPPED INCREASING
We compared two timestamped Tarrant County reports issued early on November 4.
Both reported 733,287 countywide ballots.
However, some candidate totals, write-in categories, and undervote figures changed between those reports.
Across four examined statewide contests, the reported major-party changes totaled:
Republican candidates: +7 vote entries.
Democratic candidates: +33 vote entries.
Named write-in candidates: +746 vote entries.
These figures represent changes across multiple contests, not separate individual ballots.
The pattern appears substantially associated with write-in and ballot-classification reporting. It does not currently establish an improper transfer of votes between parties.
Our next objective is to compare every available reporting interval, not just the final two, to determine whether any unexplained candidate-level corrections consistently favored one party.
FINDING 04: A MAJOR VOTER-REGISTRATION SOFTWARE FAILURE
This investigation extends beyond ballot counting.
Reporting by Votebeat and The Texas Tribune documented a Texas DPS software problem that delayed more than 500,000 voter-registration applications and updates.
A separate programming issue affected nearly 1,000 earlier registration transactions.
Officials subsequently reported that counties had cleared the large backlog, but clearing a backlog is not the same as independently verifying every transaction from beginning to end.
MBT is examining whether applications were properly saved, transmitted, acknowledged, processed, and reconciled.
We are also interested in whether delayed records could have conflicted with more recent voter-registration updates.
The figure of 500,000 represents delayed records and updates—not 500,000 illegal voters or ballots.
FINDING 05: CITIZENSHIP VERIFICATION AND IMMIGRATION-RELATED ALLEGATIONS
MBT is also reviewing allegations concerning noncitizen voter registration, immigration-related identification, DPS transactions, and claims involving Affordable Care Act enrollment.
Our investigation asks whether any actual registration pathway improperly accepted ineligible applicants, and whether any confirmed ineligible registrations resulted in ballots being cast.
We are examining federal border-encounter statistics, Texas citizenship-verification procedures, and available state investigative records.
A border encounter is not proof that someone remained in Texas. A driver's license or health-insurance identification card does not automatically establish U.S. citizenship or confer voting rights.
Likewise, an initial citizenship database flag is not proof of an illegal vote.
The investigation requires a documented connection between eligibility, registration, and actual voting before any numerical impact can be established.
Our goal is to determine what can be proven—not to convert allegations into conclusions.
FINDING 06: WE HAVE ALSO RETURNED TO THE 2020 ELECTION
Our investigation is not limited to 2026.
MBT has reopened the Texas Secretary of State's historical forensic audit of the November 2020 general election.
That state audit examined Collin, Dallas, Harris, and Tarrant counties, including polling-place reconciliation, electronic and paper records, election equipment security, mail voting, provisional ballots, and county election procedures.
The state reported reviewing at least 369 gigabytes of election-related data.
The historical audit documented administrative and reconciliation issues requiring evaluation. Those findings do not, by themselves, establish that the 2020 election outcome was manipulated.
Our updated investigation asks three important questions:
Which problems were identified in 2020?
Which corrective actions were implemented and independently verified?
Do comparable problems appear in the 2026 election records?
If the same type of discrepancy appears across election cycles, we want to know whether it results from an unresolved administrative weakness, reporting definitions, equipment procedures, or another identifiable cause.
The 2020 audit gives us a historical baseline against which to evaluate 2026.
MBT HAS SENT FORMAL PUBLIC-INFORMATION REQUESTS
We are no longer relying exclusively on public dashboards and media reports.
Mind Bend Theory has sent formal Texas Public Information Act requests seeking original records and documentation.
Texas Department of Public Safety (DPS)
We requested voter-registration transaction records, software incident reports, failed transmissions, recovery documentation, and evidence showing whether registration applications reached downstream systems.
Texas Secretary of State
We requested records concerning the DPS-to-TEAM backlog, missing transaction fields, rejection and recovery procedures, safeguards against outdated records overwriting newer information, and independent reconciliation evidence.
Tarrant County Elections
We requested 2020 audit corrective-action documentation, 2026 polling-location reconciliation reports, original tally exports, election-night reporting revisions, machine-result records, and provisional-ballot summaries.
Bee County Elections
We requested final provisional-ballot dispositions, reconciliation records, and documentation concerning the county's preliminary ballot-accounting differences.
These requests have been sent. We are awaiting responses and any releasable supporting records.
We have not received evidence establishing that the discrepancies were caused by deliberate vote manipulation.
WHAT HAPPENS NEXT?
MBT is organizing the investigation around five connected areas:
Election-night reporting: Reconstruct the sequence of vote-count changes and distinguish newly reported ballots from corrections to existing totals.
Ballot reconciliation: Compare voter check-ins, accepted ballots, rejected ballots, provisional ballots, machine results, and certified contest totals.
Voter eligibility: Examine citizenship-verification outcomes and determine whether any confirmed ineligible individuals actually voted.
Historical comparisons: Evaluate 2020, 2022, 2024, and 2026 data for recurring discrepancies, using comparable precinct boundaries and consistent reporting definitions.
Software and system integrity: Examine whether failures in registration processing, transaction recovery, or election reporting created measurable errors.
We are also watching for the release of additional county audit documents and the independent TEAM system audit expected to begin in December.
WHAT WE CAN SAY TODAY
Our investigation has identified:
A 1,079-record difference between Tarrant County's cumulative ballot report and a roster-derived turnout analysis.
A 417-ballot difference between federal and state contest accounting, isolated to absentee voting.
Election-night reporting revisions affecting candidate totals, write-ins, and undervotes.
Documented statewide voter-registration software failures involving more than 500,000 delayed applications and updates.
Historical election-audit findings from 2020 that warrant comparison with current procedures.
Unresolved citizenship-verification and registration questions requiring official evidence.
These findings are important, but none currently establishes a verified number of unlawfully cast ballots or an improper partisan change in the election result.
That is why the original records matter.
THE MBT POSITION: FOLLOW THE EVIDENCE ALL THE WAY THROUGH
An election system should be capable of explaining its numbers.
When two reports disagree, the difference should be traceable.
When software fails, the recovery should be documented.
When a ballot is rejected, there should be a lawful, reviewable reason.
And when an election result changes, the underlying records should explain precisely what changed and why.
Mind Bend Theory will continue examining these questions without treating political affiliation as evidence of wrongdoing.
We are not declaring fraud. We are documenting the questions, testing the numbers, and seeking the records necessary to establish what happened.
The requests are out. The historical comparison is underway. The investigation continues.
MIND BEND THEORY
Independent Research. Forensic Analysis. Documented Evidence.
Formal requests submitted to DPS, the Texas Secretary of State, Tarrant County Elections, and Bee County Elections. Responses and additional evidence will be evaluated as received.