1. Census
Defines the authorized ballots and the electoral capacity that must be organized. Its cuts and movements help explain why the universe changes.
Active reading mode. Shows the judicial explanation in common language.
RESERVED · JUDICIAL DELIVERY · v0.4.1
Autonomous and independent intervention · judicial-delivery version
An election is not proven because a result appears on a screen. It is proven when its complete chain can be reconstructed.
This autonomous intervention presents technical evidence organized so that the Court can follow, verify and request the sources necessary from the polling table, the electoral form and the digital file to the data and the published result.
Main result
The book documents distinct technical facts that converge on the same need: a verifiable reconstruction of the chain connecting polling table, document, file, data and result. It presents documentary differences, arithmetic operations and territorial or sequential regularities with their sources and calculations. The main result is an evidentiary chain requiring the institutional records capable of explaining every link.
Open the main submission Open the pedagogical manual
Prologue
Why was FITE born?
Democracy does not end when the results are published. It also requires that the citizen be able to reconstruct and verify how they were formed.
FITE was born from a simple question: can an independent person verify an election using only the public information that authorities make available to them? Displaying results on the screen is essential, but it does not by itself guarantee that someone can locate the documents, download them, sort them, compare them and reconstruct the relationship between a table and the declared result.
The difficulty may be material: data visible only one item at a time, fragmented files, no complete reusable tables, information that does not reach polling-place and polling-table level, or forms whose reading requires technical extraction and human control. Those conditions affect effective citizen oversight and require reconstruction of whether the publication permits independent verification from beginning to end.
The following chapters begin with concrete technical checks. Each chapter turns part of that question into verifiable evidence: the origin and technical history of files, digital identity across versions, correspondence between document and data, polling-table and territorial patterns, reconstruction of public figures, and relationships among people, documents and polling tables.
An election is materially verifiable when a third party can locate, obtain, process, compare and reconstruct the path that leads from the table to the declared result.
Purpose of this delivery
The book allows the reader to follow each finding from its question and source through its reproducible operation, technical result, evidentiary scope and specific request for information under institutional custody.
Collaborative governance and contribution traceability. Prepared legal delivery. Filing or external distribution only with registered human authorization.
Open the pedagogical manual for judicial reading
Reading path
Index and reading guide to the file
Reading mode
Explain the problem with common words, show the way verification and separates what is proven from what is pending.
Verification mode
Open locators, hashes, provenance levels and sources complete for an expert to repeat the review.
Before entering the files
Five questions, five ways to look at the system
Not all tests say the same thing or use the same source. Tap a card - the book will show you what you can check and take you to the corresponding station. This prevents a figure, a file difference or a territorial pattern from being confused with each other.
Choose a question
The answer will appear here. You can then open the station and do the checking yourself.
The question has a public base
The Constitution recognizes the participation and control of political power, access to public documents and citizen oversight of public management. Law 1712 requires maximum publicity, quality and availability of information; Law 1475 orders the immediate scanning and publication of table scrutiny minutes; and Law 527 recognizes legal and evidentiary value to data messages whose integrity and conservation can be evaluated. The application of each rule to specific claims will be legally reviewed before filing.
Official sources: Constitution, arts. 40, 74 and 270; Law 1712 of 2014, arts. 2, 3 and 7; Law 1475 of 2011, art. 41; Law 527 of 1999, arts. 8, 9, 10 and 11.
Verification mode active. Extracts, controlled files, and integrity data needed to repeat the review now appear in the source sections.
The five step rule
1. Original source
2. Repeatable operation
3. Result
4. Legal scope
5. Condition to complete verification
Reading rule: each station presents the source, verification,
result, legal relevance and institutional information needed to
complete the chain.
Procedural rule: FITE presents the evidence. The
actuated exert contradiction. The court evaluates and decides.
The pending documentary conditions are formulated as
exhibition requests to the authority that guards the
information.
Methodological chronology
How a question produced the discovery
It was not a name search to confirm a conclusion prior. It was a documented sequence of extraction, control, correction and comparison.
Initial question
The jury lists and Do the remnants retain nominal and documentary coherence?
The jury lists and Do the remnants retain nominal and documentary coherence?
Observation
Automated reading noted identifications that required human review; an alert is not treated as a conclusion.
Automated reading noted identifications that required human review; an alert is not treated as a conclusion.
Readability check
They were discarded doubtful readings and visible coincidence of identification, name, position and locator.
They were discarded doubtful readings and visible coincidence of identification, name, position and locator.
Working hypothesis
Some designations They could correspond to identifications that were not valid due to death.
Some designations They could correspond to identifications that were not valid due to death.
Independent test
Each ID prioritized was consulted in an institutional source and recorded the returned state.
Each ID prioritized was consulted in an institutional source and recorded the returned state.
Human matching
Leonilda Orrego Viera reviewed each correspondence. Three cases retain nominal registration structured complete and four visual confirmation preserved.
Leonilda Orrego Viera reviewed each correspondence. Three cases retain nominal registration structured complete and four visual confirmation preserved.
Correction and delimitation
The location of five cases was normalized. Two still require closure against the original list. The conclusion distinguishes appointment, attendance and actual service.
The location of five cases was normalized. Two still require closure against the original list. The conclusion distinguishes appointment, attendance and actual service.
Result
A file was opened nominal/documentary of seven cases and a specific list of information that only the authority can complete.
A file was opened nominal/documentary of seven cases and a specific list of information that only the authority can complete.
Public, contractual and documentary architecture
How an electoral result is constructed and circulated
The final result is the last output of a chain. To understand the findings, it is necessary to follow which universe was created, how it was divided into tables, what document was born in each table, what systems processed it and what records allow each transformation to be audited.
The operational chain
Electoral census→DIVIPOL/DIVIPOLE→Stalls and tables→Juries and materials→E-14→Transmission and precount→MMV/MMS→E-24/E-26 and scrutiny
Each arrow transfers information and responsibility. That is why each link must maintain a source, a date, a version, a custodian, a transformation rule and a control record.
2. Assignment
DIVIPOL/DIVIPOLE converts the census into department, municipality, zone, position and table. The numbering introduces the administrative order that FITE analyzes.
3. Document
E-14 fixes the written result of the table. It must relate to your pages, signatures, digital version, transmission and subsequent forms.
4. Data
Precount, MMV/MMS and counting represent the table in different layers. An image, a structured row, and a consolidated are not the same object.
5. Operators
The contracts identify components, members, supervisors and files. They allow you to request the correct record from someone who had a documented function.
6. Audit
The verification requires correspondence between table, document, data, version, publication event and result, with human and statistical controls.
One polling table produces three E‑14 records and opens three control routes
The Electoral Registry states that, after the count, polling jurors record the result in three identical copies of form E‑14: Transmission, Delegates and Claveros. All three are completed and signed at the polling table and are then separated for different destinations. They are not three names for one PDF or interchangeable publications: they are distinct physical copies with their own functions, custodians, processing times and digital versions.
| Copy | Official route after the polling table | Transformations that must leave a record | Required reconciliation |
|---|---|---|---|
| Transmission E‑14 | It is delivered to the polling-place transmitter. Codes and figures are dictated by telephone to a Telephone Reception Center (CRT); the receiver repeats them through the “echo” control and records one FRT per polling table and contest. Accepted data are validated, processed, consolidated and disclosed in preliminary-count bulletins. The copy may also follow a separate digitization and publication route. | Delivery to transmitter; call; audio or call record; FRT; FRT capture or entry; interpretation; validations; alerts; corrections; processed row; consolidation cutoff; bulletin; scan and published E‑14T version. | E‑14T values and signatures ↔ call ↔ FRT ↔ processed row ↔ bulletin ↔ published image. |
| Delegates E‑14 | It is delivered to Electoral Registry delegates, transferred to a digitization point or center, scanned and published for citizen inspection. | Collection; receipt; scanner and workstation; source file; software; operator; batch; date and time; hash; metadata; upload; URL; replacements and versions. | Values, polling-table identity, pages and signatures ↔ source file ↔ published PDF or image ↔ viewer history. |
| Claveros E‑14 | It travels with the electoral records inside the claveros envelope; the delegate issues the E‑17 receipt, the package is transported under custody, the clavero receives it and issues E‑19, the envelope enters the three-key chest and its movements are recorded on E‑20. The scrutiny commission examines it and consolidates the result through minutes, E‑24 and E‑26. Its image may also be published from the scrutiny route. | Sealing; condition of envelope; delivery and receipt times; transport; entry into and removal from the chest; opening; recounts; complaints; authorized corrections; scrutiny-system entry; scan; publication and versions. | E‑14C ↔ E‑17/E‑19/E‑20 ↔ general scrutiny minutes ↔ recounts and decisions ↔ E‑24/E‑26 ↔ published image. |
Comparison rule: every record must preserve election, department, municipality, zone, polling place, polling table, copy type, page, capture time and publication version. Hashes from Transmission, Delegates and Claveros must not be compared as though they were the same file: each copy and scan is a distinct object. The evidentiary control is to reconcile values, polling-table identity, pages and signatures, and to document every divergence between the three routes or within the history of one route.
Why this locates a divergence: if the Transmission E‑14 agrees with Delegates and Claveros but the FRT or bulletin differs, the review focuses on transmission, reception, capture, validation or consolidation. If a paper copy or witness photograph agrees with a copy but its published PDF changes, the review focuses on scanning, assembly, upload or versioning. If scrutiny changes a value, the decision must appear in the minutes, recounts, E‑24/E‑26 and software records. This identifies whether a difference arose at one relay or propagated through several.
Open the controlled matrix of the three copies and their official sources
Who does what: complete journey to publication
The expression “the electoral system” brings together different people, public agencies, contractors and platforms. This table separates each relay so that the reader can identify the institutional person in charge, the procedure and the output that must be documented.
| Stage | Who intervenes | How to do it | Exit and control that must exist |
|---|---|---|---|
| 1. Identification and news | National Registry, identification and civil registry areas; authorities that report legally relevant news. | Registrations, transfers and new developments are incorporated that enable or exclude IDs according to the law. | Record of the novelty, date of effect, source, user or batch and historical status of the ID. |
| 2. Formation of the electoral roll | Electoral Census Directorate of the Registry and its territorial dependencies. | The general registry of qualified citizens is purged and a cut for the election is closed. | Census with cut-off date; registrations, cancellations, transfers, duplicates resolved and traceability of each movement. |
| 3. DIVIPOL/DIVIPOLE | Electoral Delegate Registrar's Office, Electoral Management Directorate and Electoral Census technical group – DIVIPOL – Logistics; territorial registry offices. | The census is organized by department, municipality, zone, position and table; DIVIPOLE is updated with the electoral sites. | Versioned DIVIPOLE file, unique codes, address, capacity, coordinates, positions created, transferred or retired. |
| 4. Location of positions | Municipal or district registrar, in coordination with the corresponding mayor. | The places where the tables will operate are set by resolution and must correspond to the current electoral division. | Resolution of positions, availability supports, changes due to force majeure and publication date. |
| 5. Potential and number of tables | The National Registry establishes by resolution the maximum number of citizens per table; Registrars apply the potential to each position. | The potential of the position is divided by the authorized capacity and, when appropriate, a “queue table” is created with the remainder. | Census by position and table, formula applied, capacity, queue table and exact numbering rule. |
| 6. Distribution of citizens | Registry and census system/DIVIPOLE. | Each card is associated with a place and a table; from there the lists E-10 and the controls E-11 are produced. | Sorting rule or algorithm, version, parameters, input cut and ID-position-table correspondence table. |
| 7. Appointment of juries | Registry, territorial registrars and designation software; Entities and companies send lists of eligible people. | Jurors are selected and notified, excuses, replacements and remainders are trained and managed. | Source lists, draw or selection, E-1, resolution, attendance, replacement, function and table actually served. |
| 8. Production and distribution of kits | Registry as responsible authority; contracted logistics within the comprehensive electoral solution. | Identity-card lists, forms, envelopes and other materials are printed, packaged, identified, transported, delivered and controlled by polling place and polling table. | Manifests, serial numbers, routes, delivery records, leftovers, contingencies, responsible parties and chain of custody. |
| 9. Installation and voting | Post delegates, jurors, Public Force in security, witnesses and observers in control. | The table is set up, the voter is identified, their participation is recorded and the ballot box and documents are kept. | Installation minutes, E-10, E-11, news, authorizations E-12, signatures and total number of voters. |
| 10. Polling-table count and separation of copies | Polling jurors, under observation by witnesses, observers and authorities present. | At closing, ballots and voters are reconciled, votes are counted and all three E‑14 copies are completed and signed. Witnesses may photograph them before separation; each copy then begins a different route. | Transmission, Delegates and Claveros E‑14s; photograph before separation; polling-table identity, pages, signatures and time. |
| 11. Precount Branch | Post delegate or person in charge delivers the transmission E-14 to the transmitter; transmitter and receivers of the Telephone Reception Centers (CRT). | The transmitter dictates only codes and figures over the phone. The receiver repeats each data through the “echo” control and writes it in a Telephone Reception Format (FRT), one per table and position. | E-14T, audio or call log, transmitter and receiver identity, signed FRT, time, line, incidents and retries. |
| 12. Precount validation | Pre-counting system and verification/processing personnel under institutional and contractual responsibility. | Department, municipality, zone, position, table, transmission code, party and candidate are validated; Alerts are generated for exceeded potential or table at zero. | Validation record, alert, authorized correction, user, time, software version and FRT-processed row relationship. |
| 13. Consolidation and newsletters | Processing centers, consolidation and dissemination software and operation managers. | Accepted FRTs are processed and accumulated; The system produces successive cuts and bulletins for web, applications and media. | Base of each cut, sequence of newsletters, time and time zone, incorporation rules, logs and hash of each publication. |
| 14. Digitization branches | E‑14 collection/digitization personnel and software; Electoral Registry and contractual operator. | Delegates and Transmission copies are converted to images or PDFs and uploaded to their respective viewers. The Claveros copy may be digitized and published from the scrutiny operation. Each route must remain an independently identifiable documentary series. | Copy type; source file; scanner or workstation; operator; time; hash; metadata; batch; version; replacement and public URL. |
| 15. Custody and scrutiny | Claveros, triclave chests, scrutiny commissions made up of judges, notaries or other designated persons and, in the corresponding stage, CNE. | The E-14 of keys and the documents are guarded; The commissions review minutes, address complaints, recount when appropriate and consolidate the official result. | Record of delivery, entry and opening of the ark, AGE, E-24, E-26, claims, counts, decisions, signatures and versions. |
| 16. MMV/MMS and structured files | Results and disclosure systems under the custody of the Registry and the applicable contracted operation. | The information is exported in rows per table and candidate or in table control fields. MMV is not the digitized image of E-14: It is a structured representation that allows you to locate votes by table. | CSV or base, field dictionary, source of each value, version, hash, date, link to E-14/FRT/E-24 and generation record. |
| 17. Citizen publication | Registry, institutional portals, cloud/CDN services and operators of the contracted platforms. | Census and position consultation, pre-counting, scrutiny, E-14 viewers, downloads and structured data are enabled with access regimes that may be different. | Inventory of URLs/endpoints, published files, availability, access roles, change history, certificates, logs and preserved copies. |
| 18. Software audit and custody | Registrar's Office, Attorney's Office, Comptroller's Office, party auditors, observation and external audit; UT ILE participates in the documented procedure. | In 2026, E-14 pre-counting, consolidation, disclosure, scrutiny and digitization source codes and executables were copied, hashed, encrypted and packaged. | Inventory of servers, hashes, storage medium, multi-party password, freeze/thaw record and effectively deployed executable. |
The phone rooms that turn a voice into a bulletin
The pre-count route preserves a decisive human transformation. The data written by the jurors does not initially travel as a database row: the position transmitter read on the phone to a Telephone Reception Center. The receiver listens, repeats and records the figures in the FRT. This is followed by automatic validations, processing and consolidation.
E‑14 transmission→post transmitter→call to CRT→receiver + echo→FRT→validation→prosecution→public bulletin
To audit this branch, comparing the final PDF with the bulletin is not enough: the E‑14T, call, FRT, FRT capture or entry, processed row, alerts, corrections and bulletin that incorporated the polling table must be joined. Scanning the E‑14T is an additional output from that same copy. Both routes differ from digitization of the Delegates copy and from custody of the Claveros E‑14 through official scrutiny.
Official source: this is how the transmission of electoral results works (RNEC, 2026)
Informative path
E-14 broadcast → CRT/FRT → precount → bulletins. It produces quick information that citizens see during the day and leaves a chain of voice, reception, typing, validation and publication.
official path
E-14 of key rings → custody → scrutiny commission → E-24/E-26 → declaration. Reviews documentation, claims and counts and produces the result with legal effects.
Contractors and operators identified in the most recent investigation
| Actor or cape | Documented link | Function or operational question | Closing document |
|---|---|---|---|
| National Registry | Authority, contractor, institutional manager and custodian of the electoral architecture. | Defines rules, approves components, supervises, publishes and is responsible for traceability even if it contracts execution. | Procedures, resolutions, supervision, minutes, systems, logs and official responses. |
| UT ILE 2026 | Awardee of Act 8135/2025 and supplier of contract 049/2025. | Comprehensive solution with thirteen components: registration, candidates, juries, kits, Infovoters, biometrics, position control, pre-counting/communications, scrutiny, E-14, E-11, infrastructure and help desk. | Temporary union agreement, proposal, technical annex and RACI matrix by component. |
| UT ILE members | Thomas Greg & Sons Limited; Thomas Greg & Sons of Colombia; Thomas Processing & Systems; Security; Thomas Greg Express; Transbank; ASD Group; Opesa; Colombia Telecommunications. | Membership is documented; The exact distribution of the thirteen components among members must be proven, not assumed. | Percentages, individual obligations, subcontracts, managers, personnel, SLA and execution reports. |
| Thomas Processing & Systems / TEMIS | FITE preserved the TEMIS path to the production domain prdtpssas.com, AppSync/Cognito and AWS us-east-2; the company integrates UT ILE and declares platform and processing functions. | Prioritizes the identification of accounts, roles, deployments, processing and publishing batches. The cloud is infrastructure; The log must identify the specific execution. | Account ownership, IAM, CloudTrail, S3 versioning, AppSync, Cognito, deployments and tickets. |
| Telecommunications/CRT | “Precounting and communications” is a contractual component and the RNEC protocol uses calls between transmitters and CRT. | Determine who provided lines, PBX/services, recording, routing, redundancy, receiving stations and support in each territory. | Communications annex, CRT inventory, numbers, recordings, CDR, contingencies and service records. |
| DISPROEL and INDRA, 2022 | The official documentation for 2022 separately mentions the computerized scrutiny solution of UT DISPROEL 2021 and the National Scrutiny Software of the RNEC–INDRA contract 105/2021. | Don't treat them as a single box: identify what level of scrutiny, interface, consolidation and output corresponded to each contract. | Contracts 071/2021 and 105/2021, annexes, architecture, deliverables, audit and logs. |
| JAHV McGregor / audit | Official and journalistic sources place it in external audit functions of electoral components. | Know scope, independence, samples, exceptions, findings, improvement plans and operator response. | Contract, audit plan, working papers, complete reports and monitoring. |
| AWS, Akamai and domain registrars | They are observable infrastructure on routes preserved by FITE. | They identify hosting, distribution, or registration, but do not replace the identity of the operator of a task. | Contracting account, owners, roles, logs, certificates, historical DNS and deployment orders. |
What is documented about the 2026 operation
| Record | Documented fact | Why is it incorporated? |
|---|---|---|
| Act 8135 of 2025 | The Registrar's Office awarded UT ILE 2026 a comprehensive solution with thirteen components, including census/registration, juries, biometrics, pre-counting, scrutiny, E-14, E-11 and infrastructure. | Delimits the components and files whose execution must be reconstructed. |
| Contract 049/2025 | SECOP identifies the national contract CO1.PCCNTR.8067244, UT ILE 2026 provider and logistics, technological and IT object for 2025-2026. | It allows you to follow the contract, technical annex, responsible parties, supervision, versions and payments. |
| Software freeze | The RNEC reported the custody in 2026 of E-14 pre-counting, consolidation/disclosure, scrutiny and digitization source and executable codes, with hash and password built by four actors. | It allows requiring correspondence between the stored copy, the deployed executable and each published output. |
| Contract values | The SECOP summary field, contractual PDF, and preserved CDP show different administrative figures that must be reconciled with addendums, commitments, execution, and settlement. | Avoid confusing budget, offer, registered value, component and payment. |
| Tunja 034/2026 | SECOP identifies UT ILE TUNJA 2026 and the contract CO1.PCCNTR.9648423 for the mayoral election. | Opens a documentary comparison of components, interfaces, custody and supervision. |
| Interior MMV, 2026 first round | FITE preserved and inventoried 3,126 CSVs, 1,327,982 rows, 13,488 polling places and 118,340 polling tables; the corpus contains 0 rows from overseas constituency 88. | Establishes which structured layer is available for the interior and which layer is required to reproduce the overseas analysis. |
| E-14 exterior | The FITE collection preserves 3,670 E-14s from abroad, including 987 from the United States. | It distinguishes the existence of minutes from the absence of an equivalent foreign MMV/MMS institutional export in the public corpus. |
Public directory of routes and documentary contacts
These addresses allow you to locate sources or direct requests to the competent channel. The judicial mailboxes do not replace the PQRSDC system and must be used only for the purpose assigned to them by the entity itself.
| Entity or layer | URL or public domain | public mail | Use |
|---|---|---|---|
| RNEC electoral portal 2026 | Presidential portal, second round | PQRSDC channel from the institutional portal | Resolutions, positions, results, minutes, primers and regulations. |
| RNEC · procedures | Electoral debate procedures | — | Census, DIVIPOLE, juries, kits, logistics and specifications. |
| RNEC · transmission | CRT, FRT, echo, validation and bulletins | — | Official protocol of the 2026 pre-count. |
| RNEC · judicial notifications | Judicial mailboxes page | notificacionjudicial@registraduria.gov.conotificaciontutelas@registraduria.gov.co | Only judicial notifications and guardianships, according to RNEC warning. |
| CNE | Citizen attention | cnenotificación@cne.gov.co | Electoral authority channels and notifications. |
| SECOP II · national contract | File 049/2025 | SECOP file channels | Contract, annexes, guarantees, modifications and execution. |
| SECOP open data | SECOP II · Electronic contracts | — | Reproducible query by contract ID. |
| TEMIS / observed production | prdtpssas.com · AWS AppSync/Cognito · region us-east-2 | Only contacts identified in the contractual file | Preserved technical route; requires ownership, roles and logs. |
| Thomas Greg & Sons | Public corporate directory | customerservicegsc@thomasgreg.com | General contact published by the company. |
| Technical Information Management | Corporate directory | customerservice.MTI@thomasgreg.com | General contact of the company in the public directory. |
| Security | Corporate directory | mesadesoporte@securidsas.com.co | Published support desk. |
| Group ethical line | Public ethical channel | transparency@thomasgreg.cometicatps@thomasgreg.com | Ethical reports directed to TGSC or Thomas Processing & Systems. |
Documentary responsibility map: public need → call → award → contract → component → responsible → document or system → version/hash → supervision → publication → audit.
What should be displayed
The specifications, the proposal, the temporary union agreement, the technical annex, the matrix of responsibilities for each of the thirteen components, subcontracts, CRT inventory, recordings and FRT, acceptance tests, supervision reports, incidents, payments, settlement and the logs that relate each published version to its source and authorization. The chain must respond, for each stage: who received, who transformed, who validated, who authorized, who published and who supervised.
Integrated main document
Object of the intervention
Leonilda Orrego Viera appears as autonomous intervener and independent, formulation subject to final legal control, to provide a limited documentary verification and request the incorporation of institutional sources necessary to clarify it.
Context of the intervention
This aid does not present separate documents. Presents a sequence of findings that must be read as evidentiary support for the same question: whether the published electoral chain can reconstructed from the original source to the consolidated data.
The following chapters function as files of verification: each one identifies a source, an operation reproducible, a result, its evidentiary scope and the information that must exhibit the authority to complete the control institutional.
central question
Can the Court completely reconstruct the chain between table, document, published data and electoral result with the information currently available, or must request the authority the sources, versions, records and media that remain under its custody?
evidentiary thesis
The evidence gathered documents digital differences, arithmetic reconstructions, structured transformations, territorial patterns and individualized files that are not They resolve with a general declaration of normality. Every discovery requires concrete verification with complete official records.
methodological unit
The research is organized as a chain: question, source, operation, result, evidentiary scope and request to the Court. This continuity allows us to understand why the findings are not isolated anomalies, but rather verification points within the same documentary and technological reconstruction.
Main document
Autonomous intervention document
The formal writing brings together the thesis, the facts, the requests and the relationship of findings of this intervention. This copy is intended exclusively for the judicial proceeding and preserves its own documentary identity within the package.
Open the main submissionHuman control, corrections and current scope
FITE preserves the error, documents the correction and publishes the current figure
Traceability also applies to the research itself. When a machine reading or methodological formulation is corrected, the previous record is preserved and the current version explains exactly what changed.
| Revised object | Previous formulation | Current correction |
|---|---|---|
| Paired coverage, 2022 | 103,364 was previously used as the number of exact matches between rounds. | Each round contains 103,364 unique keys; the exact DEP-MUN-ZONA-PUESTO-MESA intersection is 103,295. The remaining 69 keys per round contain overseas zone remappings and are kept separate. |
| Confirmatory universe | Not all filtering stages had been distinguished. | 103,291 tables retain the required candidacies; 103,200 have a defined share in both rounds; the final contrast uses 89,718 polling tables in 5,291 valid polling places. |
| Order Footprint direction | Similarity between neighboring polling tables had been described as if it identified favorability by itself. | The current version separates three questions: continuity between neighbors, direction along the sequence and the margin effect produced when candidate share coincides with the amount contributed. |
| 104 E-14 of the United States | Automatic extraction opened 104 apparent differences; A first check left six candidates. | The second visual control closed 104/104: 101 coincidences, 3 documentary discrepancies and 99 errors of automatic extraction versus human reading. |
| Candidacy 2026 | An inherited column was named “Petro.” | The correct identification on the E-14 reviewed is Iván Cepeda Castro; The other candidate compared is Abelardo de la Espriella. |
Three documentary discrepancies that remain open
88_360_40_02_056: E-14 2/10 versus pre-count 2/11. 88_360_40_81_013: E-14 12/400 versus precount 12/480. 88_360_40_85_004: E-14 26/332 versus precount 25/332. The values are expressed as Iván Cepeda / Abelardo de la Espriella.
Control result: The human review removed 101 alerts as differences in extraction or transcription and retained only the three visible differences that must follow the documentary and publication chain.
Individualized file · FITE-JUR-007
Seven cases documented in preserved jury lists
Before the show
What is this file?
This station brings together seven individualized matches found by reviewing preserved public lists of full or remaining jurors. In each case, the list allows a person to be located in a municipality, position or table; A preserved institutional consultation records a status that requires documentary reconstruction of the designation, any replacement, attendance and signatures. It is not a list of opinions or a national figure: there are seven specific files that the Court can review one by one.
Why is it important to understand this finding?
The validity of an election day depends, among other elements, on whether it can be reconstructed who was designated to fulfill a function and what official documentation allows that designation to be verified. When two public sources refer nominally to the same person, the court needs to be able to examine that correspondence with complete documents.
Question investigated: Do the preserved official designations match, by identity and location, with state records that require additional documentary verification?
What the reader will check: how a match stops being an isolated alert when the name, the necessary identification according to the public source, the municipality, the designation and the corresponding institutional source are connected, step by step.
Fidelity of evidence: The public documents that support this module are preserved intact in the controlled source. The interface displays the information necessary to understand and verify each correspondence, without replacing or altering the original document.
7nominal matches
2departments reviewed
5municipalities reviewed
3preserved raw responses
The actual scale of this check
This station credits seven individualized cases within a review that reached five municipalities of two departments: Apartadó, Cañasgordas, Medellín and Turbo in Antioquia, and Ibagué in Tolima. It does not present those seven cases as a measurement of all the municipalities in the country. Precisely because Therefore, the result converts a documented alert into a request for institutional contrast on the entire universe of appointments, replacements, attendance and forms.
History of discovery
- Review of juror and remainder lists generated identification alerts that required human verification.
- Each alert was subjected to a readability check: name, identification, position and locator had to visibly match.
- Prioritized identifications were queried individually in an institutional source and the result was preserved.
- The human matching linked each match to its documentary support and separated the cases with a preserved raw response from those that require completion of that provenance.
- The result was an individualized file of seven cases and a precise request for documentation in the custody of the authority.
Methodological result: The discovery does not come from an isolated figure; It is born from a visible sequence of alert, control, consultation and comparison.
Probation station
Build the conclusion with the evidence
Activate the available parts and see what statement is objectively accredited. The demo uses controlled copies and results submitted to verification.
0 of 7
collated cases
ListName, identification and function
StateRegistered institutional result
OriginRaw body preserved when it exists
Select the cases you want to compare:
Start by activating the list.
The evidentiary scope will change with each piece examined.
Accredited fact
The same ID appears in a list of designation and in a registered institutional result as “Cancelled due to Death”. In three cases it is also raw response preserved.
Condition to complete verification
The authority must display the history of validity, the appointment, notification, attendance, replacements and the forms related to each case.
Expert conclusion of the module
The objective technical finding is an individualized nominal and identifying correspondence between seven preserved designations and seven institutional states registered as “Cancelled by Death.” The evidence proves the need for individual documentary reconstruction. The category of impact is a break in verifiability that requires completing the institutional chain of each case.
Specific request to the court
Require from the authority the individual documentary chain of the seven cases to establish the sequence between state, appointment, eventual replacement and electoral performance.
Also require the national universe of lists of jurors and remnants, with their history of changes, to apply the same documentary contrast in a complete and verifiable manner.
The operation, explained without technicalities
Locate the person on the list
Read ID and title
Check the institutional status
Compare both pieces
Register limits and slopes
Source A: designation list
The extract allows you to see name, identification and function without depend on a transcript.
Open full nominal record of the seven cases
The registry keeps the name, identity card, function, location, date of consultation and locator of each public source. It does not replace the original sources: it allows to locate the exact evidence that must be collated.
Source B: institutional response
For cases C05, C06 and C07, the raw body that returned status Cancelled by Death. To C01–C04 the recorded result and its date are retained, but the equivalent raw body must be incorporated into the package.
C05_respuesta_institucional.json · SHA-256 verifiable on the manifest
Operation: join two pieces by the same identification
No odds are added or behavior inferred. It is check that the visible identification in the list is the itself consulted in the institutional source.
reproducible result in the seven
preserved records, identifying the list
corresponds to the identification consulted. The force
documentary is not identical: three cases have provenance A and
four provenance B.
Documentation to be exhibited by the authority
To complete the individual chain of each case requires the validity history, the act of designation, the notification, replacements, attendance, and forms of the table in institutional custody.
Level A 3 cases with crude responseB Level 4 cases that require completion
provenance
Second station model FITE-CEN-001
How was this figure constructed?
Act 1 · The question
Why did we investigate this calculation?
In the public debate, a figure of 751,156 votes associated with three groups of municipalities: 95 with an electoral list ratio greater than 100%, 150 greater than 95% and 256 greater than 90%. Such a figure can influence the understanding of an electoral debate; therefore, before using it as support for any claim, it must be able to be reconstructed from its list and its operation.
Before comparing: electoral census and population are not the same thing
Population census: It is the DANE estimate or count of the people who inhabit a territory. It is used to describe the population.
Electoral census: It is the database of cards authorized to vote, inside and outside Colombia. It is updated with inscriptions, first expeditions and news of registrations and cancellations from the National Identification Archive.
The essential difference: The electoral roll is not a statistical projection. It is an administrative database of people eligible to vote.
Therefore, a percentage can only be interpreted if it identifies what population figure use, from what date is and with which version of the electoral roll was compared. The comparison may reveal a case worthy of review; it does not convert a projection of inhabitants into a list of voters.
The decisive point is simple: the three groups are not three independent lists. The group of 95 is within the group of 150, and the group of 150 is within the group of 256. If added as separate lists, some municipalities enter two or three times in the calculation. If counted as unique municipalities, each municipality enters only once.
Question investigated: How was the public figure constructed from the municipal groups that accompanied it, which municipalities did it include, which were left out and what counting rule produced the disclosed result?
What the reader will check: reconstruct the operation with the three published groups, see that they are contained within each other and observe how the result changes when each municipality is counted only once. You can also alternate between the initial sweep of 257 municipalities and the public list of 256, where the case of Sanctuary, Risaralda, is identified.
The reproducible reconstruction transforms a disseminated figure into a controllable operation: it allows identifying the universe, the inclusions and exclusions, the counting rule, the sources of population and electoral potential, and the documents that must be exhibited to support it.
Institutional source of the meaning of each census: National Registry of Civil Status, official section “Electoral Census”, consulted on August 8, 2026.
- Question: understand how a public figure was formed.
- Reconstruction: activate the groups and change the counting rule.
- Learning: observe what changes between accumulated sum and single municipalities.
- Verification: identify the official information necessary to complete the origin of the calculation.
Calculation impact: 751.156 It is not the margin of three different municipal universes. It is the result of adding three groups that contain each other. When each municipality is conserved only once, the margin of the published universe of 256 municipalities is 384.215. The difference, 366,941 votes, corresponds to margins already included that were recounted; they are not new votes.
Act 2 · Interactive reconstruction
Rebuild the figure step by step
Activate the groups, see which contains which, and then change the counting rule. You perform the operation that allows you to understand how the figure was born.
0
votes in visible calculation
Universe:
256greater than 90% group
150already included within the group of 256
95already included within the previous groups
Activate at least one group to start visible reconstruction.
95 municipalitiesThey appear three times when the three groups are added together.
55 municipalitiesappear twice when the 90% and 95% groups are added.
106 municipalitiesappear once: they are only in the group greater than 90%.
Municipality excluded from the public list
Sanctuary, Risaralda appeared in the initial sweep with a ratio of 92.3877% and a margin of 2,496 votes. When switching between the two universes, notice how this exclusion modifies the calculation without altering the counting rule.
Act 3 · What did we learn?
Experience shows that the result depends on the rule with which the municipalities are grouped and accounted for. Now you can read the full calculation, not just its final figure.
Proven Result
evidentiary scope
The operation reproduces exactly how the figure of 751,156 is formed from accumulated groups. By counting each municipality only once, the observed margin in the universe of 256 municipalities is 384,215. The difference of 366,941 does not represent additional votes: it is the arithmetic effect of adding municipalities that were already included.
Specific request to the court
Require the original municipal list, the formula used, the version of the population used, the bases of electoral potential and the results supports that allow the formation of any aggregate figure to be verified.
Expert conclusion
The objective technical finding is a reproducible arithmetic inconsistency: the sum of accumulated groups incorporates repeated municipalities, while the count by single municipality preserves a single universe. The category of affectation is loss of verifiability of an aggregated figure until the list, the sources and the calculation rule that produced it are displayed.
What this station includes and excludes
It includes the arithmetic reconstruction of the aggregate margins and the effect of counting municipalities of nested groups once, twice or three times. It excludes municipalities outside the thresholds and does not attribute, by itself, modifications of votes, tables, juries or individual documents. The complete reconstruction depends on the original list, its versions and the requested institutional supports.
Condition to complete verification
The authority must display the coded list of municipalities, the source and date of population, the electoral potential, the municipal results and the formula used. This documentation allows you to verify the exact origin of the published list and any exclusions or updates applied.
Cross control · three figures, three different objects
What are we really counting?
Why is it important? Because a figure can refer to qualified people, to a projection to organize several days of voting, to votes or to a difference between years. If these objects are mixed, the conclusions of USA 81–86, E-11, table capacity, versioning and municipal census are also mixed.
What the reader will check: Select each figure and assign it to the layer you think is correct. The station will show what is reproduced, what corresponds to an institutional explanation and what file is missing to complete the check.
Start with a number.
Then choose the layer to which you consider it belongs.
751.156unclassified
885.400unclassified
2.419.734unclassified
Proven Result
751.156 It is a reiterated sum of margins of municipal groups contained within each other. 2.419.734 is the net growth between the official censuses of 2022 and 2026. 885.400 It is explained by the Registry and the MOE as the sum of quantities projected to organize the outdoor sessions from Monday to Saturday; FITE reproduces the aggregate structure of that layer, but the exact sum still depends on the primary file delivered to auditors.
Why does it affect other tests?
This separation allows you to correctly read the USA codes 81–86, the potential of each table, the E-11, the consular days, the comparison of versions and the municipal figures. A stall code between 81 and 86 is also not sufficient on its own: there are ordinary stalls with these numbers in the interior. The identification of a foreign day requires department 88, country or consulate, zone, day, position and table.
expert conclusion
The evidence proves that the three figures belong to different operations and universes and cannot be added or used as if they described the same phenomenon. It also proves the existence of an external operational layer separate from the nominal census. The independent reproduction of 885,400 remains incomplete because the file and dictionary containing the projected amounts are not available in the public information examined.
Specific request to the court
Require a full copy of Basic auditors president 2026, its dictionary, formula, unit of each row, version, date and hash; and order its reconciliation with the DIVIPOLE, the nominal census of 41,421,973, the foreign potential of 1,414,661, the 1,489 additional tables and the individual documents E-10, E-11 and E-14 from days 81 to 86.
Open controlled file: three figures, three different objects
controlled source
Preserved summary of the reconstruction: groups of 256, 150 and 95 municipalities; single margin and cumulative sum. The archive also records the initial sweep of 257 municipalities and the exclusion of Sanctuary, Risaralda, before reproducing the public list of 256.
12_summary_margen_second_turn_alerta_censo.json · SHA-256 verifiable in the manifest
Repeatable operation
It is checked if each group is independent or if it is contained within the previous one. When the cumulative sum is chosen, the three published margins are added. When single municipalities are chosen, the group of 256 is retained only once.
Result
384.215 + 274.197 + 92.744 = 751.156. The last two components correspond to municipalities that are already within the universe of 256.
Information to be displayed
Codified list of municipalities, population source, date and version of each base, electoral potential, municipal result, formula and any record that allows reconstructing or correcting the aggregation.
Document Station · E13-21
What happens when the same record retains its visible content, but changes its documentary representation?
Before the show
Why does this question matter?
The reliability of a digital record does not depend only on whether a result can be read. It also depends on being able to reconstruct the published representation, its versions and the marks that accompanied the document.
What the reader will check: will switch between a historical page, a current aligned version, and a display of the zones that changed.
Controlled Demonstration Case
Anolaima, Cundinamarca · zone 00 · position 00 · table 002
This case is part of 21 comparable pairs. On the 21, the “Confidencial” visual brand is present in the historical version and absent in the current one; The content of the corresponding QR code is preserved.
Choose a view.
The image will change so that you can check the visual transformation.
Operation available.
Observe each view and activate “View Difference” to visually locate the transformation.
21comparable pairs
42compared pages
21/21distinct hash and corresponding QR
The verified magnitude: the same transformation is repeated in All Pages of the 21 pairs closed, distributed among Cundinamarca (17) and Sucre (4). This station displays one of those pairs so the reader can verify it directly.
Expert conclusion
The evidence accredits a repeated visual transformation: on 42 pages of 21 comparable pairs the “Confidential” mark disappears among the preserved representations, while the corresponding QR identification is preserved. The category of impact is structured transformation of representation and rupture of documentary traceability.
Specific request to the court
Require the institutional history of versions, loads and replacements, along with records identifying the process responsible for the new E14 representation.
Documentation station · E4-Meta
How can the same table appear publicly as “non-digitized” and also as a readable E-14?
Before the show
Why does this question matter?
To verify an election it is not enough that a file exists. It should also be possible to know which file officially represents each table, what was published before and after, and why a query may show different documents for the same reference.
What the reader will check: will activate three preserved representations of table 001. Two display the “Page not digitized” warning; The third matches the expected name published on the official list and shows the E-14 form. The operation allows us to see why they are not equivalent.
Verifiable individual case
Goal · municipality 50_001 · zone 99 · position 20 · table 001
The public preserved list allTransmissionCodes register the position once, with idTransmissionCode 3105717 and an expected file name. The station does not decide what was the cause of the change: it leaves visible the documentary difference that must be explained with the institutional records.
Choose a representation.
The station will show what each file allows to verify.
Operation available.
Open the two copies with notice and then the expected file to check the difference in document type.
1individualized table
2copies with notice
1E-14 visible by expected name
expert conclusion
The preserved evidence accredits an objective ambiguity of public representation: for the same table, two PDFs coexist that report “Page not digitized” and a PDF whose name matches the file expected in the official list, which does allow the E-14 to be seen. The technical category is a break in traceability between the publication reference and the accessible documentary representation.
Information to be displayed
Publication history, versions, timestamps, origin of the expected name, reason for the “Page not digitized” notice, and certified correspondence between the transmission code, the E-14, and the other table documents.
Specific request to the court
Order the display of that individualized documentary history for code 3105717, including the relationship between each URL, file, upload, and replacement.
Documentation station FITE-VERS-587
The same document, several versions: what changed and who should explain its history?
Before the show
Why does this question matter?
An electoral PDF is not just a visible image: it is a digital object whose history must be verifiable. When the same reference appears in two versions, the court needs to know whether the two are identical, when each was published, and what process explains any difference.
What the reader will check: open two preserved copies of the same table, toggle between them, and run a comparison showing their hashes, sizes, geometry, and representation. Knowledge of the general fit method is not required to verify this particular pair.
Controlled Demonstration Case
Atlanta - Consulate · zone 01 · station 02 · table 001
Version A constitutes the preserved early memory of the document. Version B corresponds to a later query of the same reference. The generation date is not recorded in the properties of the PDFs: this absence is part of the institutional information whose display is requested.
Choose a version to get started.
The station will open the selected copy and its preservation tab.
Operation available.
Activate A, B and then “Run Comparison” to transform the visible evidence.
987recovered pairs
400pairs with identical SHA-256
587pairs with different SHA-256
How to read the magnitude: Within the 587 digitally distinct pairs, FITE prioritized 266 for detailed review. In that review, 191 showed relevant loss of quality and 223 affected numerical zones. These numbers describe overlapping controls: they do not add to each other or automatically convert to outcome changes.
Expert conclusion
En el caso demostrado, las dos copias tienen SHA-256 distinto, tamaños distintos y geometría de página distinta, aun cuando ambas se identifican con la misma mesa, conservan dos páginas y no aportan texto extraíble. En el universo preservado, esta diferencia digital se reproduce en 587 de 987 pares. El hecho técnico acreditado es una diferencia digital objetiva entre versiones que exige reconstrucción documental del historial de publicación. La categoría de afectación es ruptura de trazabilidad documental y pérdida de verificabilidad independiente.
Information to be displayed
Version history, institutional hashes, timestamps, generating software or service, publication records and logs that relate each object to its previous and subsequent version.
Specific request to the court
Require the preservation and display of records that allow reconstruction for each pair of the date, process and version responsible for the proven difference.
Open controlled case file · Open controlled summary of the universe
Played Plugin · Los Angeles
The same nominal record changes from three pages to two
FITE preserved two versions identified as United States · Los Angeles · zone 035 · block 81 · table 001. The reader's action allows us to verify that not only the fingerprint changes: it also changes the structure of the document.
Open V1 and V2.
Then run the comparison to see what changes between the two copies.
Operation available.
The demo uses the preserved files, not a simulation.
3 → 2pages
20 → 15objects declared by /Size
98.027 → 53.725bytes
SHA-256 ≠non-identical copies
Supplementary expert conclusion
Preserved versions of the same documentary reference are neither digitally nor structurally equivalent. V1 contains three pages, three images and /Size 20; V2 contains two pages, two images and /Size 15. Both lack sufficient internal documentary metadata to reconstruct from the PDF itself when and by what process the substitution occurred. The category of impact is proven structural versioning and breach of documentary traceability.
Information to be displayed
History of uploads and substitutions, complete institutional versions, hashes, timestamps, responsible user or service and documentary justification for the deletion of the third page.
Specific request to the court
Order the preservation and display of the institutional history that allows identifying the origin, date, process and technical person responsible for each version.
Custodial precision: This case accredits the change between V1 and V2. External claims about exact cloning between acts, nine XObjects, or uniform canvases require incorporating and reproducing their exact source files before attributing them to FITE.
Limited documentary reconstruction · Los Angeles
A table, several public references: what is missing to unite your story?
This station adds a different verification: the same table appears in public records with bulletin references and, on the other hand, in PDFs preserved with different digital structures. The reader reconstructs the available path and sees the exact point at which the history is needed in institutional custody.
Credits from public sources: tabular data preserved by Carlos / onorosaurosrex and for AnZaCa. Limited cross-reading and verified PDF versions are separately credited and maintained by FITE. These sources share coverage in part of the universe; For this reason, they are read as complementary documentary routes, not as two independent confirmations.
Start with the same table.
The station will gather only the public data and preserved copies necessary to show what can be followed and what must be delivered by the authority.
Checking available.
Open the four layers in order: table, bulletin references, PDF representations, and pending institutional record.
expert conclusion
For table 88-360-35-81-001, the compared public records maintain the same numerical result, but display different bulletin references: 2, 0002 and 9999. FITE also preserved non-identical PDF representations of the same nominal reference. The set documents a incomplete public chain: With the available materials it cannot be established alone which bulletin, time, file and PDF version form a single institutional sequence.
Information to be displayed
The institutional definition of bulletins 2, 0002 and 9999; your date, time, time zone and system of origin; and the certified history of table loads, substitutions, hashes, URLs and files.
Specific request to the court
Require from the authority that individualized documentary history for the table, so that the relationship between result, bulletin and PDF version can be verified without depending on inferences.
Technical attribution finding · TEMIS/AWS
From the file that changed to the path that published it: how far can responsibility be traced?
The investigation began with non-identical documents. The next question was inevitable: by which service did these versions circulate and which organization had documented functions over their processing and publication?
Why is it important to understand this route?
A digital file is similar to a package that passes through several stations. The public page shows it; a service queries the data; an infrastructure transports it; and a contract assigns functions. Confusing these layers can attribute responsibility to those who only provide infrastructure or leave unidentified those who did have a documented technical function.
What the reader will check: will build the route link by link. Each action will add a visible connection and show which function is documented, without jumping from an internet address to an unsupported personal attribution.
How do we come to investigate this route?
- Observed difference: FITE preserved copies of the same electoral reference that were not digitally identical.
- Repetition: The differences did not appear as a single case; They were grouped into sets of files and dates, with repeatable structural transformations.
- New question: If multiple copies were processed in a similar way, which service generated or republished them?
- Portal preservation: TEMIS public configuration showed low production service
prdtpssas.com, connected with AppSync and Cognito on AWS. - Contractual crossing: the procurement documents led to contract 049/2025, UT ILE 2026 and the documented technical functions of Thomas Processing & Systems.
- Remaining question: What automated work, technical identity, and authorization produced each particular version?
TEMISofficial electoral portal
→prdtpssas.comproduction service domain
→AppSync/Cognitodata consultation and technical access control
→AWS us-east-2observable public infrastructure
→Contract 049/2025electoral contractual framework
→UT ILE 2026full contractor
→Thomas Processing & Systemsdocumented technical functions
Start with TEMIS.
The route will only be illuminated when each connection has public or contractual support.
Route not yet reconstructed.
Activate all seven steps in order.
Visible addresses and companies do not fulfill the same function
Select each trace. The explanation changes to prevent infrastructure, distribution and domain registration from being confused with the authorship of a transformation.
Choose a technical trail.
You will see what allows you to identify and which document completes its attribution.
Do the transformations follow a pattern?
Activate each dimension. They are not four figures to add: they are four ways of looking at the same phenomenon. The screen changes to show how many files it covers, when the versions appeared, where they are concentrated, and what digital operation is repeated.
Choose a dimension.
Visible evidence will change to show scale, time, territory, or form.
Correct reading of the chain: TEMIS → domain
prdtpssas.com → AppSync/Cognito → observable public infrastructure → contract 049/2025 → UT ILE 2026 → Thomas Processing & Systems → documented responsibilities. The infrastructure identifies the environment; The contract and functional documents allow progress towards technical responsibility.expert conclusion
The public and contractual evidence allows us to reconstruct a coherent technical route from the TEMIS portal to Thomas Processing & Systems, a member of the UT ILE 2026 with documented functions of platform design and implementation, electoral data processing and digitization and publication of minutes. The proven fact is the attribution of the technical and contractual ecosystem; The identity of the automatic job, account, user, or service that produced a particular version must be established with operational records under institutional and contractual custody.
Condition to complete verification
Domain and AWS account ownership required; AppSync and Cognito identifiers; object and version histories; execution, deployment and access logs; batch jobs; technical roles; and the tickets or authorizations associated with processing.
Specific request to the court
Order the Registrar's Office, the UT ILE 2026 and its members to preserve and display these records, with the correspondence between each published PDF, its version, the process that generated it, the technical identity used and the applicable authorization.
Document continuity station · USA-987 / days 81–86
How is the identity of a table preserved when passed from one record to another?
Why is it important to understand this finding?
An electoral investigation cannot stop at a code appearing on a list. You should be able to follow it to the documents that explain the table, the voters and the result. This station clarifies a confusion that affects the reading of the USA case: the 303 codes for days 81 to 86 did not disappear from the preserved census; They are within the universe of 987. The evidentiary question is another: whether public information allows each one to be connected with their individual electoral supports.
Question investigated: Do the 303 codes constitute an absence of census or a gap in documentary correspondence?
What the reader will check: It will first activate the entire universe, then isolate days 81 to 86 and, finally, follow the documentary chain that must exist for each code.
Related institutional background: Before this response, FITE had already preserved an array of 253 operational positions abroad. The official file provided by the RNEC confirmed these 253 positions without numerical differences in potential or number of tables and added a row with zero voters and zero tables. The RNEC also reported that the official potential by table, country, consulate, position, day and day makes up the Electoral Census, but provided only one aggregate table per position. It did not include the details by table, day or day or the requested E-11s, which it indicated were being digitized and processed. It also invoked the legal reservation to not disclose individualized or disaggregated data that would allow registered citizens to be directly or indirectly identified. The sequence confirms the territorial base preserved by FITE and delimits the exact point that remains without public verifiability: individual documentary correspondence.
Reproducible demo
Build the correspondence map
Open the views in order. The screen goes from showing a total figure to showing the specific question that must be answered for each code.
Start with the codes that FITE saved from the public source.
The first action shows in which public file the 303 codes were left.
Operation available.
The station does not seek to create a new figure: it transforms the question from “do they exist?” towards “can they be verified individually?”
expert conclusion
The preserved codes From days 81 to 86 there are 303 and they are part of a public archive that FITE kept with 987 records. “Preserved” means that they were preserved as they appeared at the source, with date and fingerprint, so that they can be reviewed again. The accredited technical fact is not an absence of census, but rather a lack of verifiable connection between each code and its documents: The preserved public information does not allow to completely and verifiably link each code with E10, E11, MMV and MMS. The category of affectation is the loss of the possibility of verifying, with public documents, how a code is connected to its electoral supports.
Information to be displayed
An institutional table by code that relates day, consulate, position, table, E10, E11, MMV, MMS and the corresponding E14 identifier; along with publication date, version and source of each link. When the information contains data subject to confidentiality, its display must be carried out with the safeguards established by the Court, without preventing the verification of the finding.
Specific request to the court
Require from the authority the individualized correspondence matrix of the 303 codes and the necessary documentary support so that each relationship can be verified independently.
What are the 74 cases?
They are a subset of the 303 codes. They were prioritized because, in addition to belonging to days 81–86, they cross with special versioning or hashing states. They are not a number of votes nor are they added to the 303: they indicate where the individual documentary verification should begin.
Documentary continuity file · FITE-USA
Identifiers with numbers and letters, Houston and versions: how do we check if they are still the same file?
Before the show
Why does this question matter?
To know if a document has changed, it is not enough to look at the current copy: you need a previous copy saved in a verifiable way. This station compares an early download with later versions to check if they are still the same file and can still be related to a particular table.
What the reader will check: will activate four different documentary controls. Each answers a different question: what was preserved, how many versions reappeared, whether the visible identifier actually verifies the file, and what institutional information was obtained or remains in custody.
Continuity demonstration
Answer one question at a time
These four cards do not add up. Each answers a different question about the history of the archives: what was saved at the beginning, whether another version appeared, what the visible code identifies, and what information the authority must explain.
Choose a question.
The station will display the file or record compared, the operation performed and the result that can be checked.
Checking available.
Choose a card to see the evidence that answers that question.
Expert conclusion
FITE preserved three lines of evidence that answer different questions. First, it recovered 1,006 digital files or records from an early stage of publication. Second, within the USA-987 set, 134 pairs also displayed a third digital version during a subsequent download. Third, in a separate review of 6,008 comparable pairs, the number and letter code visible next to the document did not match the SHA-256 fingerprint calculated on the downloaded PDF. This shows that there are version histories that need to be rebuilt and that the visible code alone does not allow us to verify what the final published file was.
Information to be displayed
Institutional manifests by version, correspondence between visible identifier, digital object and calculated hash; in addition to dates, URLs, publication history and logs of the service that produced each version.
Specific request to the court
Order the delivery of a verifiable history by object, from the first preserved publication to the currently served version, with its hashes, identifiers, dates and technical managers.
Own station: HNI · hexadecimal identifier.
This station allows you to check whether a code visible next to a document actually serves to identify the PDF being downloaded.
Go directly to the HNI demonstration.
HNI-0 station · visible hexadecimal identifier
Does a visible code actually verify the file being downloaded?
HNI-0 is a FITE work tag. It is not an official category nor does it presuppose what the code means. “Hexadecimal” means a reference written with numbers and letters; “bytes” are all the digital content of the PDF. HNI-0 allows you to compare that visible reference with the SHA-256 fingerprint calculated on the entire file.
Why is it important to understand this finding?
When a portal displays a code next to a document, a citizen may believe that this code confirms the identity of the file. Before using it as a guarantee, you have to check it. This station reconstructs that comparison without converting the visible code into something the source has not defined.
Question investigated: Does the visible reference written with numbers and letters match the fingerprint calculated on the final downloaded PDF?
What the reader will check: will choose a seed reference, activate the matching and check how three things are separated: territorial code, visible identifier and calculated file footprint.
1. Select a seed reference.
Choose a preserved territorial code. Then turn on matching to see what comparison was made.
The comparison has not been run yet.
Select a reference first; The comparison does not modify the source, it only compares the visible identifier with the calculated fingerprint of the preserved PDF.
6.008comparable HNI-0
0matches with final SHA-256
5.622Audited URLs
1.149representation changes with the same text
expert conclusion
The reproduced comparison proves that, in the revised HNI-0 universe, the visible hexadecimal identifier didn't work like the SHA-256 fingerprint of the final downloaded PDF: in 6,008 comparables, 0 matches were recorded. The technical fact is not that the code necessarily lacks function, but rather that its function does not allow itself to verify the digital identity of the final PDF without the corresponding institutional dictionary. The impact category is loss of independent verifiability of the published reference.
Information to be displayed
The authority must provide the identifier dictionary, its generation rule, the exact object it corresponds to, URL, ETag, Last-Modified, official hashes by version and the publication records that connect each code with each PDF.
Specific request to the court
Require the institutional table of correspondence between visible identifier, published object, URL, version, hash and territorial code, along with the technical records necessary to individually reconstruct each reference.
Related documentary controls
Pick a part and see what it makes possible to check
These checks do not repeat the main findings. They explain which documents must remain together so that a table can be traced without confusion.
Select a documentary piece.
The station will show what information must be kept connected to perform the match.
Operation available.
Open a part to see the minimum verification chain.
Orientation station four MMS/MMV families
Before the acronyms: four different questions about the same control records
Before the show
Why is it important to understand this finding?
It is not necessary to memorize MMS or MMV. They are technical names of control records that the system relates to a table. What is important is this human question: when multiple records should describe the same table, do they match each other?
FITE found four ways to review the same question. These are not four accusations or four figures that should be added up. Each one looks at a group of tables and fulfills a different function.
- MMS-65: the small set where a specific difference was located and verified table by table.
- MMS-834: a broader classification of cases that must be separated by type of difference.
- MMS-1226: an even broader family showing the scale of repeated transformations.
- MMS-301: the tables that appear simultaneously in the two large families; It is used to check which part is shared and which part is not.
Question investigated: What does each review contribute, which tables coincide with another and which groups should be kept separate?
What the reader will check: will open the four revisions one by one. The diagram will show how many tables each contains, and by viewing the common tables, it will separate the 301 flagged by both from the 925 only flagged by MMS-1226 and the 533 only flagged by MMS-834.
How to read this page: think of four table lists. A table can be in only one list or appear in two. That is why MMS-65, MMS-834, MMS-1226 and the 301 common tables are related, but not added: counting the same table twice would give a misleading figure. Each station retains its own question, source and conclusion.
Scope Demonstration
Open a review and see what tables it includes
Each choice changes the diagram: it shows which group of tables is being reviewed and avoids confusing one group with another.
Open a review.
The station will show the tables included in it. Open MMS-834 and MMS-1226 to see the 301 tables they both list.
Operation available.
Start with one review and then see which tables the two large reviews share.
Expert conclusion
The evidence is not a scattered list of warnings. Document a repeated control phenomenon on four scales: 65 tables with a localized transformation that can be exactly reconstructed; 834 cases classified by the form and location of the difference; 1,226 tables where the relationship between records appears again; and 301 tables flagged by two independent reviews. The partial coincidence and repetition by classes allow us to affirm that these are differences structured and documentable, not from a single isolated discrepancy. Keeping the groups separate protects the accuracy of that conclusion.
Specific request to the court
Require an official explanation of what each MMS/MMV field means, what rules validate them, when they were created or updated, and what tables allow each table to be related to its records.
Control station MMS-65
What happens when a control value does not match the records it should summarize?
Before the show
Why is it important to understand this finding?
Control records help verify that the figures traveling between stages of scrutiny retain the same reference. When a control field presents a different figure from the other records in the same table, it is necessary to reconstruct precisely where the difference is located.
Question investigated: In what part of the chain is the difference observed in the field located? TOTALVOTOSURNA from MMS?
What the reader will check: will select three demo tables, visually separate the digits of each MMS value, and remove the highlighted digit. You will then check how the reconstructed number matches MMV, field E11, and the total located in E24.
65closed single tables
18departments with cases
351.511commission records reviewed
33available interior territories
The verified magnitude: the pattern plays in 65 of 65 tables of the closed set, distributed in 18 apartments. The subsequent national review contrasted 351,511 records commission on 33 territories; 15 compared territories did not provide cases. This station shows three examples so that the operation is verifiable without replacing the entire record.
Reproducible demo
Find the digit and reconstruct the number
MMS, MMV and E11 are names of official control fields. The reader does not need to know programming: just compare the numbers in the same table. Votes are not added or subtracted here: it is checked whether, when removing a specific position from the MMS number, exactly the same number reappears as shown in the other three records.
Select a table.
The station will display the MMS value and the three contrast records of that same table.
Operation available.
Open a table and remove the indicated digit to repeat the check.
Expert conclusion
The objective technical finding is a reproducible difference located in the MMS field TOTALVOTOSURNA. On the 65 closed tables, removing a single digit from the MMS value exactly reconstructs the MMV value; In those same 65, MMV, field E11 and the total E24 located by complete territorial key coincide. The category of affectation is a systematic transformation located in a field of control and a need for documentary reconstruction of its generation process.
evidentiary scope
The check separates the field where the difference appears from the three matching records in the same table. It does not transform the MMS difference into a vote figure: it documents precisely where it is located in the chain of control.
Specific request to the court
Require from the authority the MMS field schema, its creation and update records, the history by table and the documentation that explains the process that generated the observed values.
Sorting station MMS-834
What does a review reveal when 834 cases do not reflect a single difference?
Before the show
Why is it important to understand this finding?
A large number is only useful if we know what's inside. Separating cases by type of difference avoids treating them as if they all said the same thing.
Question investigated: How are the 834 cases distributed according to the size of the difference, the record where it appears and the existence of a visible event?
What the reader will check: will open each group and see what tables it contains. You will also check that these groups are not added as votes: they are different ways of ordering the same reviewed cases.
834prioritized cases
139with visible event
695no visible event
620plausible queue to close
Reproducible demo
Separate the cases and keep where each one comes from
Select a group. The station shows which tables it contains and keeps the link to your documents. The groups are formed with different questions: the size of the difference, the place where it appears or the existence of an event. That's why they don't add up to each other.
Select a criterion.
The station will show how many cases there are in that group, without mixing groups formed with different rules.
Operation available.
Open a group to see what question it answers.
Expert conclusion
The 834 cases demonstrate that the difference is not concentrated in one table nor does it take a single accidental form. The review identifies repeated classes: 514 differences between 0 and 2, 87 between 3 and 10 and 155 greater than 10; In addition, it locates 65 possible added digits in MMS, 2 in MMV, 11 unusual values and one value with a visible event. These counts answer different questions and are not added. Read together, they prove a structured distribution of differences in control records and justify the individual reconstruction of prioritized cases.
Specific request to the court
Require an official explanation of what each field means, what events and changes were recorded, and documents counts, appeals, exclusions and charges related to the 834 table codes.
Scale station MMS-1226
What changes when the same check is observed in an extended family?
Before the show
Why is it important to understand this finding?
Scale matters: a control difference can be isolated or reappear in a broad set. This station shows what proportion of the extended family retains the same type of relationship between MMV, E11 and MMS.
Question investigated: How are the 1,226 tables distributed according to the type of difference registered and whether or not there is an associated visible event?
What the reader will check: You will be able to see the full 1,226 tables, separate them by whether or not there is a visible event, and see the five recorded ways of the difference. Each selection rearranges the diagram to make its size visible.
1.226family tables
31departments
1.047no visible event
179with visible event
Reproducible demo
See the scale and try a real numerical transformation
The review examined 351,511 MMS records corresponding to 118,350 control tables. The screen does not change the original data: it only arranges it so that its distribution can be seen. It then allows you to open a preserved example of each change form to see which numerical position was modified.
Select a view.
The station will show how the group of 1,226 tables is distributed.
Operation available.
Open a view to see how many tables each group contains.
Expert conclusion
The MMS-1226 family provides a repeated break within a specific control record. In 1,226 tables, MMV, E11 and the MMS field of voters maintain the same reference figure, while TOTALVOTOSURNA contains a different figure. This locates the difference at a specific point in the chain, not the three matching records.
The modification follows five repeated classes: 964 substitutions of a character, 65 insertions, 12 order exchanges, 37 losses and 148 cases of another registered class. The repetition of shapes on the same field proves a structured and reproducible transformation. Its cause, date, and validation process can only be completed with records in institutional custody.
Specific request to the court
Require the schema and validations applied to TOTALVOTOSURNA, the table history of its transformations and the records of the process that built and published the MMS/MMV fields.
Common table station · MMS-301
What does the part that two different families share teach?
Before the show
Why is it important to understand this finding?
Two forms of review can point to some of the same tables or point to different tables. The comparison allows us to clearly see which were indicated by both and which by only one. This separation helps to check whether the two reviews are seeing the same phenomenon or differences of different types.
Question investigated: How many tables were flagged at the same time by MMS-834 and MMS-1226, and what does that comparison show within all the tables reviewed?
What the reader will check: will activate each family and then the view of common tables. The station will show the 301 tables highlighted by both reviews and, separately, the tables that only appear in one of them.
301common tables
925only MMS-1226
533MMS-834 only
118.350control tables
Reproducible demo
See the tables that both reviews point out, without counting them twice
First open each list. Then click “See common tables”. The diagram will show tables that are only in MMS-834, those that both revisions point to, and those that are only in MMS-1226. Each table appears only once.
Open both lists.
The common tables view will be enabled after opening MMS-834 and MMS-1226.
Operation available.
Open both lists before viewing the common tables.
Expert conclusion
The comparison shows that MMS-834 and MMS-1226 are two different revisions: both indicate 301 tables; MMS-1226 also reports 925 that are not in MMS-834, and MMS-834 reports 533 that are not in MMS-1226. In the 301 common tables, the five MMV rows add up exactly to the total MMV and were compared to the 118,350 tables reviewed. This partial coincidence needs to be explained: different control records are pointing to some of the same tables and some to different ones.
Specific request to the court
Require documents that explain how MMS, MMV, E11, E14, E24 and E26 are connected for each table, what rule joins those records, and what controls verify that the figures match.
Comparative station · USA · E11
What changes when the same network of tables is compared in two rounds?
Before the show
Why is it important to understand this finding?
A figure for a single day does not allow us to clearly distinguish what belongs to the reported participation of that day and what changes when observing the same universe in another round. Comparing equivalent codes allows us to reconstruct the behavior of the table network without confusing an isolated case with the whole.
Question investigated: How did the total number of voters reported in E11, by consulate and by day, change between the first and second round in the preserved codes of the United States?
What the reader will check: You will choose a day and see how the comparable totals of the two rounds change. You can then open a preserved individual case to repeat the basic change calculation.
987preserved codes
983comparable in E11
216.288first round
222.532second round
The proven scale: The comparison covers the 99,59 % of the 987 preserved codes. Four were not charged because field E11 was not completed in one of the rounds. The operation keeps those four cases out of the calculation, instead of filling in figures that are not in the source.
Reproducible demo
Change your day and rebuild your balance
Field E11 expresses the total number of voters reported at each table. Only comparable codes are grouped here; The change is calculated by subtracting first round from second round.
Select a day.
The station will display the totals kept for each lap and the arithmetic difference.
Operation available.
Choose a day to compare the same universe of codes between the two rounds.
Expert conclusion
The objective technical finding is a reproducible variation of the total E11 between two rounds in 983 preserved equivalent codes. The entire universe goes from 216,288 to 222,532 reported voters, a net difference of 6,244. The temporal distribution is not uniform: the codes from Monday to Saturday increase by 10,130, while those from Sunday decrease by 3,886. The category of impact is a documented temporal pattern that requires institutional reconstruction of the correspondence between day, position, table and voter lists.
evidentiary scope
The station credits the comparable E11 totals and the result of their subtraction, with explicit coverage and without imputing the four incomplete codes. It also allows you to verify that the aggregate variation changes depending on the selected day.
Specific request to the court
Require the E10 and E11 nominal lists of both rounds, the official table that links country, consulate, day, position and table, and the supports for assignment, opening, closing and consolidation of the compared codes.
Statistical station table order
Does the result change systematically with the table number?
Before the show
Why is it important to understand this finding?
Tables in the same position share a place and electoral context. If their order were irrelevant to the composition of the outcome, rearranging that order within the position should not destroy a repeated national direction. That's why FITE compared the low, medium and high tables within each position.
Question investigated: Does the observed proportion between the two candidates change regularly as the number of tables within the same position increases?
What the reader will check: will alternate between the official order, a control that excludes the last tables and partially used tables, and 999 internal rearrangements. The display will react by showing which part of the pattern remains.
81.769analyzed tables
3.391stalls with 14 or more tables
3.249positions with positive direction
33 of 33departments with positive gradient
The verified magnitude: The analysis does not depend on a single streak or a single city. The direction appears in 95.81% of the polling places examined and in all 33 departments available with polling places containing at least 14 polling tables. This station presents the controlled universe, its denominators and its result.
Reproducible demo
Change the order and see what remains
Each bar represents one-tenth of the journey from the lowest-numbered tables to the highest-numbered tables within each position. The height shows the aggregate proportion of Iván Cepeda between the votes of the two candidates.
Select a test.
The station will transform the visible evidence according to the chosen scenario.
Test available.
Open the official order to get started and then apply the controls.
expert conclusion
The technical phenomenon observed is a reproducible national gradient between the table number and the proportion of votes registered within the same position. In the official order, the first decile registers 41.09% and the last 58.91%, a difference of 17.82 points. The pattern retains 98.02% of that difference by excluding the last table and tables with E11 less than 80% of the position median. In 999 rearrangements that preserve each position and its vote distribution, the average slope is close to zero; the observed reaches the empirical limit of the test, p ≤ 0.001. The behavior is compatible with a systematic rule of composition or arrangement of tables and not with a random permutation of the same results within the position.
evidentiary scope
The evidence accredits the magnitude, territorial frequency and robustness of the gradient. The institutional rule that formed the tables is not contained in the public data examined; Its reconstruction requires documentation in the custody of the authority.
Specific request to the court
Require the official rule or algorithm applied in 2026 to assign qualified citizens to table numbers, the corresponding nominal ranks or links, the census enabled by table and forms E10 and E11 that allow checking the composition of the prioritized positions.
Territorial station · accumulated reconstruction
What changes when the territories are moved from largest to smallest?
Before the show
Why is it important to understand this finding?
A final national result can hide territorial movements in the opposite direction. The continuous shape of a cumulative curve allows you to see where the change in direction is formed; Reconstructing the margin municipality by municipality allows us to verify this route with specific votes and territories, without attributing to time a phenomenon that belongs to the territorial order of the data.
Question investigated: How does the accumulated margin between the two candidates evolve when the territorial units are covered in the order preserved by the source?
What the reader will check: It will move the territorial cut and observe how many units and votes have been incorporated, what the margin is at that point and how much the remaining section contributes. You will be able to directly compare the mathematical minimum with the territorial cut used by the source and verify that they are not the same point on the route.
1.122territorial units
25.731.269represented votes
−574.206cumulative minimum
+73.021final margin
Reading Accuracy: The preserved source declares an order from largest to smallest population and excludes the international vote. Its horizontal axis represents accumulated votes. The population table, its date and the territorial key used to sort must be displayed to institutionally validate that criterion.
Reproducible demo
Move the cut and see where the curve changes
The margin is defined as Espriella votes minus Cepeda votes. A negative value favors Cepeda; a positive one favors Espriella. Activate first the mathematical minimum and then the source cut: the station shows your positions, accumulated votes and margins without treating them as equivalent.
Move the control.
The station will update the territory, accumulated votes and margin.
Tour available.
Compare the mathematical minimum with the Almaguer–Cajamarca cut defined in the source. The curve is not a decorative illustration: each point corresponds to territorial units and verifiable accumulated votes.
expert conclusion
The preserved series reproduces a large accumulated territorial variation. The global minimum is reached in Inzá, position 272, after incorporating 20,810,300 votes, with a margin of −574,206. The territorial cut defined by the source appears later: Bogotá–Almaguer gathers 441 units and 22,687,050 votes, equivalent to 88.17% of the series, and closes at −488,555. From Cajamarca to the smaller unit, 3,044,219 votes are incorporated, the remaining 11.83%, and the margin changes by +561,576 until ending at +73,021. The technical phenomenon is a reproducible accumulated territorial investment; The mathematical minimum and the beginning of the second section are not the same point.
evidentiary scope
The station credits the shape of the series, its objective points and the arithmetic contribution of each section. It also documents that the conclusion depends on a declared territorial order that must be contrasted with the official population table and with the original municipal results.
Specific request to the court
Require the official table of results by municipality used in the series, the population source with date and version, the territorial union key, the treatment of homonyms and special territories, and the documented justification of the Almaguer–Cajamarca cut.
Independent control case · Atypical election of mayor of Tunja 2026
How do you check a streak without confusing different choices?
Before the show
Why is it important to understand this finding?
Tunja corresponds to an atypical mayoral election that is different and independent from the presidential election analyzed in the other modules. It is incorporated only as a control case: it allows us to see, with a closed universe of 362 tables, how a base is refined and how a streak is checked by table order before interpreting it.
Question investigated: Within this independent election, does the grouping of winners by consecutive tables remain after removing duplicates, missing tables, unclosed records, and numerical associations that do not match the official PDFs?
What the reader will check: will switch between received file, strict layer, strict prudent layer and reordering control. Each action modifies the coverage and visible statistical evidence.
Origin and authorship of the work
Documentary contribution and origin of the Tunja line: Frente Digital (@frentedigital2026), tabular files from Tunja received between August 1 and 2, 2026. Independent verification, reconstruction and analysis: FITE, integrity audit, key normalization, prudent exclusion, second manual comparison of individualized cases and formal streak testing, performed on August 2, 2026.
362official polling tables
359unique keys received
37rows with inconsistencies
358polling tables in the formal test
This result has two different parts:
1. The file was reviewed first. It contained 362 rows, but did not represent 362 distinct tables. When each table was identified by its area, position and number, there were 359 unique tables: three appeared repeated and three tables on the official list were not in the file. These differences were corrected before studying the streaks.
2. The order of the corrected tables was then studied. The test found that winners were grouped into consecutive tables more often than expected in the statistical control. That clustering also appeared when comparing the victories of the candidate identified as C6 versus all other candidates. When repeating the same test for C10, the result was not sufficiently different from what a rearranged sequence could produce. That is why C6 and C10 do not receive the same conclusion.
1. The file was reviewed first. It contained 362 rows, but did not represent 362 distinct tables. When each table was identified by its area, position and number, there were 359 unique tables: three appeared repeated and three tables on the official list were not in the file. These differences were corrected before studying the streaks.
2. The order of the corrected tables was then studied. The test found that winners were grouped into consecutive tables more often than expected in the statistical control. That clustering also appeared when comparing the victories of the candidate identified as C6 versus all other candidates. When repeating the same test for C10, the result was not sufficiently different from what a rearranged sequence could produce. That is why C6 and C10 do not receive the same conclusion.
Reproducible demo
Purge the base before interpreting the streak
Select a stage. The station will show what information can be used at any given time and how the pattern reading changes.
Select a stage.
Start with the file received and work your way up to the formal proof.
Control available.
Statistical evidence is enabled only after the integrity check.
expert conclusion
In the independent atypical Tunja election, the technical phenomenon observed is a reproducible clustering of winners at consecutive tables that remains under different layers of control. In the strict prudent layer, 129 of 211 consecutive pairs share a winner, compared to a null mean of 96 pairs; The maximum streak is 18 tables and the empirical contrast registers p = 0.00025. An additional formal test on 358 tables obtained 144 multicategory runs versus 170.92 expected and 70 runs for C6 versus the rest versus 106.06 expected, both at the limit of 99,999 permutations, p = 0.00001. The behavior of C10 compared to the rest did not produce the same result, p = 0.14658. The technical category is a grouping pattern by table order, stable against quality controls and compatible with a non-random logic of composition or ordering.
evidentiary scope
This station does not mix or transfer results between elections. It proves, within Tunja's own universe, that a streak must be reviewed after the integrity control and that the pattern persists after excluding unclosed records and repeated numerical traces. It also separately documents the errors of the collaborative file, so as not to attribute to the official E-14 a difference originating from extraction or association of rows.
Specific request to the court
Require for Tunja the complete official index of the 362 tables, the E-10 and E-11, the rule for assigning voters and jurors by table number, the E-24 and recount records of individualized cases, and the institutional publication history of the E-14.
Verifiable annex of the finding FITE-JUR-007
The seven cases: what these tokens represent
Before checking the registry
What is being seen here?
These seven tokens are not a separate find or a random list. They are the individual record of the previous discovery: seven people who appear on public lists of jury or remainder appointments and who, when compared with a preserved institutional source, appear with the status “Cancelled due to Death.”
Where do they come from? Each file brings together two pieces that refer to the same identification: the designation list and the institutional record consulted. That is why municipality, function, table or group of remnants and identification appear.
What does each card represent? An individual documentary file that allows the court to locate the exact case and require its official chain: designation, date and registration status, replacement, attendance and table forms.
How to read it: first identify the location and function; then look at the ID match and institutional status. The “Provenance A” classification indicates that a crude institutional response was preserved; “Provenance B” identifies cases whose documentary provenance must be completed with the authority.
This reading view displays only the information you need to get your bearings. Verification mode opens the corresponding locators and controlled media.
Return to the step-by-step demonstration of the discovery
Apartadó, Antioquia
| Function | Vocal |
| Table | Zone 1 · position 02 · table 10 |
| ID | ****3778 |
| State | Cancelled by Death |
Cañasgordas, Antioquia
| Function | Chairwoman |
| Table | Zone 99 · position 22 · table 2 |
| ID | ****5749 |
| State | Cancelled by Death |
Medellin, Antioquia
| Function | Remainder |
| Table | Zone 23 · position 01 · remnants |
| ID | ****6863 |
| Earring | Table normalization |
Turbo, Antioquia
| Function | Chairwoman |
| Table | Zone 99 · position 14 · table 18 |
| ID | ****5876 |
| State | Cancelled by Death |
Ibagué, Tolima
| Function | Chairwoman |
| Table | Zone 7 · position 01 · table 2 |
| ID | ****6967 |
| Control | Preserved nominal collation |
Ibagué, Tolima
| Function | Vocal |
| Table | Zone 7 · position 06 · table 1 |
| ID | ****7052 |
| Control | Preserved nominal collation |
Ibagué, Tolima
| Function | Vocal |
| Table | Zone 9 · position 02 · table 12 |
| ID | ****7634 |
| Control | Preserved nominal collation |
Expert conclusion of the annex
The seven files identify the correspondence that gave rise to the discovery: the same identification appears in a public list of jurors or remaindermen and in an institutional consultation preserved with the status “Cancelled due to Death.” The comparison proves the nominal and documentary coincidence that must be reconstructed case by case. It does not replace the appointment, attendance, replacement and signature documents: it identifies exactly the seven people and tables for which these supports are necessary.
Specific request to the court
Require for each file the original designation, its date and validity, the E-10, the E-11, the attendance, the replacements, the signed record and the institutional history of the registration status, preserving the identification necessary to carry out the comparison.
Integrative expert report · FITE joint reading
Integrated Report on the Documentary Chain and the Order Footprint
From polling table to result: the complete chain that must be explained.
The purpose of this report
This report brings together the documentary, census, territorial and statistical stations to show the complete chain: census and position → E-10/E-11 → table → E-14 → versions and representation → MMV/MMS → E12 → E-24/E-26 → order pattern → Order Footprint.
The central question is the architecture that takes a census decision to the numbered composition of each table and, then, to the declared result.
Document
The E-14 set the documentary point of the table; E13, E4 and E6 allow representation and versioning transformations to be followed.
Structured data
MMS-65, MMS-834, MMS-1226 and MMS-301 retain separate universes and crossovers, with identifiable denominators.
Weather and exterior
E12 and USA-987 require connecting consulate, post, table, minutes and publication event.
census architecture
The census and DIVIPOLE define the board; The E-10 concentrates the rule that orders and divides enabled people into numbered tables.
Measured order
Blocks, streaks and the Order Footprint measure continuity while traversing official polling-table numbers within the same polling place.
Reproducible control
Tunja and the national test apply purification, electoral unity and permutations before interpreting the signal.
Executive matrix of the fourteen stations and families
This table serves as a reading map. The rows that share a universe are kept separate so as not to add figures that answer different questions.
| # | Station or family | What was observed | What the authority must complete |
|---|---|---|---|
| 1 | Juries and remainders | 7 individualized nominal correspondences. | Designation, attendance, replacement and signatures per table. |
| 2 | Municipal census | Reconstruction of a public figure from 751,156 to 384,215 by counting unique municipalities in 256 cases. | Original list, date, source, formula and version. |
| 3 | E13-21 | 21 pairs and 42 pages with preserved visual change. | Version history and rendering process. |
| 4 | E4-Goal | A table with two representations of “not digitized” and a readable E-14 file. | Certified correspondence between code, URL, file and upload. |
| 5 | USA-987 version | 587 different prints and 400 identical prints within 987 pairs. | Manifests, dates, versions and reason for replacement. |
| 6 | USA-303 | 303 correspondence codes preserved; Cases require individual comparison. | Official keys between table, bulletin, E-14, E-24 and E-26. |
| 7 | USA/HNI-0 Continuity | Comparison of visible identifiers, bytes and SHA-256 in the preserved universe. | Institutional definition of the identifier and relationship with the final file. |
| 8 | MMS/MMV Families | Four control universes with separate questions and denominators. | Dictionary, validation rules and upload versions. |
| 9 | MMS-65 | On 65 tables, removing a digit from the MMS field reconstructs MMV. | Field layout and generation records. |
| 10 | MMS-834 | 834 cases classified by size, field and event; The classes do not add up. | Complete base, exclusions and update traces. |
| 11 | MMS-1226 | Extended family to check if the shape is repeated outside the initial subset. | Universe, territorial key and inclusion criteria. |
| 12 | MMS-301 | Intersection of two independent reviews on 301 tables. | Identity of the reviews and documentation of each crossing. |
| 13 | Tours, territory and Tunja | Paired controls, territorial order and streaks within comparable units. | Data from both laps, official order and comparison rules. |
| 14 | Order Footprint | Continuity between neighboring tables, direction along the sequence and margin effect measured separately in the paired 2022 panel. | Electoral roll/DIVIPOLE, E-10/E-11, numbering, polling-table loads, remappings, tie-breakers, exceptions and logs. |
Minimum statistical reproduction sheet
Corrected and expanded central result
The Order Footprint
FITE separated similarity between neighboring polling tables, ordinal direction and the arithmetic effect on the margin. This separation resolves the contradiction in the previous report and states precisely what each test measures.
103,295exact keys between rounds
89,718polling tables in confirmatory tests
5,291valid polling places
B=999permutations in each global control
1. Local continuity: neighboring polling tables do not behave like just any order
In the second round, the average lag-1 correlation of the Hernández/(Hernández+Petro) ratio was 0,5606 in 5,294 positions. Under 5,000 permutations, 76.18% of the positions obtained p<0.05 and 72.69% remained significant after FDR of 5%. This result establishes the continuity of the sequence; the direction is determined in the next control.
2. Direction: Hernández's share is concentrated at the beginning of the order
Across 5,291 polling places, the directional measure had a mean of −0.65185 and a median of −0.72308; 96.54% of polling places were negative, with lower-tail and two-sided p=0.001 against the reordered controls. The negative sign means that Hernández's share is higher in early-numbered tables and declines toward late-numbered tables.
3. Margin effect: load and relative change shift toward Hernández
Across 89,718 polling tables in 5,291 polling places, the observed combination equals +252,655.30 margin units toward Hernández, with upper-tail and two-sided p=0.001; the central range of the reordered controls was approximately −4,457 to +4,278. The effect is positive in 89.43% of polling places and remains +251,871.11 after excluding overseas constituency 88.
The observed mechanism combines two gradients: early tables contribute an average of 242.55 Hernández+Petro votes versus 201.38 in late tables —20.44% more— while the relative change between rounds is +11.85 points at the beginning versus +2.55 points at the end. Both controls have p=0.001.
4. Historical persistence: 2018, 2022 and 2026
The cross-sectional control compares the observed sum with a scenario in which polling tables inside each polling place contribute equal weight while retaining their candidate shares. The arithmetic effect recurs in three elections:
| Election | Candidacy | Margin shift | Early/late profile | Control |
|---|---|---|---|---|
| 2026, interior | Abelardo de la Espriella | +432,826.77 | 57.75% / 42.00%; early load is 1.335 times the late load | B=999; p=0.001; Bonferroni=0.006 |
| 2022, second round | Rodolfo Hernández | +439,806.81 | 56.79% / 36.99%; early load is 1.313 times the late load | B=999; p=0.001; Bonferroni=0.006 |
| 2018 | Iván Duque | +382,115.90 | 61.71% / 43.91%; early load is 1.387 times the late load | B=999; p=0.001; Bonferroni=0.006 |
Persistence is not limited to a national total: the effect is positive in the 34 coded departments in 2018, the 34 in 2022, and the 33 interior departments available in 2026. The 2026 overseas constituency remains a separate line because the interior MMV corpus does not contain DEP=88.
5. Evidentiary expansion of August 12: the complete trajectory
The new test followed the cumulative margin table by table across 104,782 polling tables in 6,116 interior polling places in 2026 and compared the official sequence with 9,999 within-place reorderings that preserved the same votes, the same polling tables and the same final result in every polling place.
The observed trajectory remained ahead at every decile and reached +5.3726 percentage points above the linear path to the same final margin at the midpoint. The simultaneous full-curve test yielded familywise p=0.0001. Among 1,181 polling places with exactly two runs, the Abelardo run appeared first in 1,156 cases—97.883%, compared with 50.047% in reordered sequences.
The control assigning equal weight to every polling table retained a midpoint curve of +2.7147 points in 2018, +3.3611 in 2022 and +2.7323 in 2026, always with familywise p=0.0001. The shape also persisted across 4,118 common administrative polling places in all three years. The result therefore does not depend only on larger polling-table loads or on different polling places entering each election.
6. E-11, documentary and consolidation controls
Antioquia supplied 15,801 polling tables. After controlling for load, E-11, last-table status and ordinal shape, and then adding all available documentary-status controls together, 96.47% of the residual signal remained, with familywise p=0.0001. Complete exclusion of the seven audited E-11 alerts did not remove the result.
These controls narrow the next link to the composition and assignment of voters to official polling-table numbers. Verification requires the assignment-system rule and version, E-10 or an anonymized table-level matrix, historical DIVIPOL directories, and MMV/MMS for overseas constituency 88.
7. Verifiable evolution of documentary traceability
The historical review corroborated a traceability change: the official 2024 build constructed deterministic territorial paths and names by department, municipality, zone, polling place, polling table and election; the early 2026 corpus preserves 121,629 territorial/table keys associated with hexadecimal names, and the official catalog archived on June 23 links 122,014 table, territory, transmission-code and expected-name records.
Credit for the historical line and samples: Primera Línea Digital - Frente Digital ANZACA AndreTaker. Independent corroboration, preservation and date delimitation: FITE Laboratory.
Why Hernández did not win in 2022: the calculation documents a margin shift within the analyzed panel and order; This displacement coexisted with the national total that gave victory to Petro. The investigative question is no longer whether the Footprint disappears because Hernández lost, but rather how an arithmetic advantage associated with the order was produced that was not able to reverse the aggregate result.
Controls that strengthen the result
Between 88% and 90% of the effect follows a smooth ordinal position profile. After removing that profile using linear, quadratic, and cubic controls, a residual of approximately 25,933 to 29,008 margin units remains, with p=0.001. The 2018-2022 census growth does not absorb the specific change: when controlling for job size and department, the partial correlation is ρ=0.0083, p=0.550; for the quadratic residual against the census and its growth, ρ is around ±0.002, p≈0.87.
What should be displayed
Historical census and DIVIPOLE cuts, table creation and numbering rules, E-10/E-11 per table, charges and voters, zone remappings, tiebreakers, exceptions, external MMV/MMS, version histories and records that connect table, minutes, transmission, consolidation and result.
Integrative conclusion: FITE documented a chain of documentary, structural and statistical regularities that converges on a specific requirement: disclosure of the complete architecture linking the electoral roll, polling table, electoral form, structured data, versions and results. The Order Footprint makes this requirement visible in the official polling-table sequence.
Integrated expert conclusion
What the research as a whole shows
The conclusion in simple words
The electoral result is constructed as a chain: first it is defined who can vote; Then citizens are distributed in numbered positions and tables; the jurors count and write three E-14s; A copy is dictated by telephone for the pre-count, others are digitized or entered for scrutiny; Finally, systems and operators produce files, bulletins and results.
FITE followed this chain and found documentary differences, version changes, repeated transformations in control fields and a national regularity linked to polling-table number. The most difficult pattern to see—the Order Footprint—can be explained with boxes: early polling tables not only had a more favorable candidate share, but also contained more votes. They therefore contributed more to the final sum.
In the election addressed here, the cross-sectional effect in the 2026 interior corpus shifted the margin by +432,826.77 toward Abelardo de la Espriella. The paired control of the same tables in 2022 opened the mechanism: the coincidence between load and relative change shifted the margin by +252,655.30 toward Hernández; after excluding the overseas constituency it remained +251,871.11. The 2018 precedent toward Duque completes the historical persistence. Table number, candidate share and the number of votes contributed are not acting as independent variables.
To explain how this architecture was built, the census allocation by polling table, E‑10/E‑11 and all three complete E‑14 copies with their pre-separation photographs must be disclosed. The evidence must reconstruct separately Transmission → call → FRT → capture → validation → bulletin, Delegates → receipt → scanner → file → viewer → versions and Claveros → E‑17/E‑19/E‑20 → chest → commission → E‑24/E‑26 → publication, together with MMV/MMS, contracts and every operator's logs.
A chain, not an accumulation of figures
This book does not ask to add different findings or convert a technical difference into a conclusion that the evidence does not contain. It presents a chain of verifications: identity of people and documents; construction of public figures; correspondence between records; preserved versions of files; control fields; order of tables; territorial series and patterns that remain after explicit controls.
The evidentiary force arises from the fact that each station preserves its source, its operation and its result, and that several stations reach the same institutional need: to reconstruct the documentary and technological history that cannot be deduced from the visible publication alone.
1. Identity and designation
Seven individualized correspondences link preserved lists of jurors or holdovers to a recorded institutional status. They demand reconstruction by person of designation, validity, attendance, replacement and signature.
2. Figures and correspondences
The census reconstruction shows that a public figure depends on how the municipalities are grouped. In the USA, 303 codes are within 987 preserved records, but their individualized correspondence with the media is still necessary.
3. Documents and versions
Pairs E13 and PDF document repeated visual and digital differences. The Los Angeles case further demonstrates a structural change preserved from three to two pages in the same documentary reference. The TEMIS rebuild links the portal, production domain, AWS services, and election contract to documented technical responsibilities; The logs must identify the specific execution of each version.
4. Fields and results
The MMS/MMV family is read in four groups that are kept separate: MMS-65 checks for a localized difference; MMS-834 orders prioritized cases; MMS-1226 shows a broader review; and MMS-301 shows the tables that the last two groups have in common. The USA comparison between rounds, the order of tables, the territorial curve and Tunja document operations and patterns that can be repeated and verified.
5. Joint technical scope
The set accredits traceability ruptures, loss of independent verifiability, proven digital and documentary differences, arithmetic reconstructions and patterns that require institutional explanation with primary sources.
6. Information in custody
Complete verification requires institutional versions and hashes, upload records, logs, assignment rules, E10, E11, E14, E24, correspondence tables and individual files that only the authority keeps. The recent responses from RNEC and CNE confirm that the foreign potential by table, country, consulate, position, day and day is part of the institutional census, and delimit the dependencies that maintain the requested technical support; however, individualized detail and background records must still be displayed to allow verification.
FITE general expert conclusion
The evidence presented accredits a verifiable break in documentary and technological continuity in the route that goes from the table to the published result. It is not a single figure or a single file: they are reproducible checks at different points on the same route.
FITE demonstrated that a public figure changes when eliminating repeated municipal counts; that documents preserved with the same reference may have different digital and visual representations; that the publication history cannot be reconstructed from the visible PDF; that an identifier shown to the reader does not by itself confirm the final file; that a control field can differ from three matching records in the same table by repeated ways; that the official order of the tables and certain territorial series preserve patterns after explicit controls; and that seven nominal correspondences require individual documentary reconstruction.
The cause, user, and timing of each transformation must be determined through the institutional records of the corresponding link. The already proven fact is this: The available public material does not allow for the independent, complete and continuous reconstruction of the chain that should connect table, form, digital version, control record, consolidated data and official result.
Therefore, for FITE, table relationship omissions, version histories, load records, assignment rules, E-10, E-11, E-14, E-24, E-26, and designation brackets are not administrative details. They are the documents that would allow the differences already observed to be verified or explained. They remain in institutional custody and must be exhibited.
FITE technical opinion: The accumulation of findings constitutes an objective basis for a comprehensive judicial reconstruction of the documentary and technological chain. Considered with the collective rights, facts and legal rules set out in the autonomous intervention in the popular action, this evidence supports preservation, disclosure, audit and protective measures capable of securing a verifiable electoral chain.
Electoral integrity is not presumed because a result appears on the screen. It is accredited when a competent person can follow, check and explain the complete path from each table to the declared result.
Linked evidentiary request
Information necessary to close or dismiss each case
- Certify the exact historical status of each ID in the date of appointment and on the election date.
- Provide the novelty, act or ANI lot that bases the status “Cancelled due to Death”, with effective date.
- Provide the original version of the jury list, date of issue, author and record of modifications.
- Inform notification, training, excuse, attendance and replacement of each person.
- Submit E10, E11 and the applicable table forms, including replacement and signature records.
- Explain the two locations not yet standardized and certify the position and the exact condition of appointment.
- Provide all three complete E‑14 copies—Transmission, Delegates and Claveros—and the photographs taken before separation for every polling table examined; certify the correspondence of identification, values, pages and signatures.
- Provide the chain E‑14T → call → FRT → capture/entry → validations and corrections → processed row → bulletin, together with the published file and versions of the Transmission copy.
- Provide the chain Delegates E‑14 → receipt → scanner/workstation → source file → batch → upload → viewer → versions, including operator, date, time, hash, metadata and replacements.
- Provide the chain Claveros E‑14 → E‑17 → transport → E‑19 → chest/E‑20 → commission → minutes, recounts and decisions → E‑24/E‑26 → published image, including responsible persons, times, software and versions.
- Build a judicial polling-table matrix that classifies every divergence as: a difference between copies, a change within the history of the same copy, or a difference between a document and processed/consolidated data.
The request is formulated on a case-by-case basis, not as a generic request. Your The purpose is to obtain the official documentary chain necessary to explain or dispute accredited correspondence.
Controlled access
Sources and annexes
Full feeds contain personal data. They open in mode verification and its use is limited to this judicial proceeding.
AnZaCa Credit: the public heritage of AnZaCa provides documentary references, versions and external analyzes used as contrast in E13/E6, versioning, external records and the technical layer. FITE separately accredits its cross-checks, calculations and verifications; sources are read as complementary paths, not duplicate commits.
Activate the mode Verification at the top
to view locators and files.
Preserved official lists
Rated human control
CSV record
·
Metadata
Nominal manual comparison carried out by Leonilda Orrego Viera. It is not presented as a certified digital signature.
Census pilot: preserved summary of the reconstruction
Open census calculation summary
Includes the values to reproduce the sum of the nested groups, the unique universe of 256 municipalities and the documented effect of the exclusion of Santuario, Risaralda from the initial sweep of 257.
Finding 14 · pedagogical explanation and confirmatory result
The Order Footprint: why polling-table position changes the contribution to the result
The Footprint is not a mark drawn within a PDF. It is the statistical trail that appears when the tables are ordered by their official number and it is discovered that the first and the last do not provide the same combination of political percentage and number of votes.
Primary focus of this presentation: the 2026 presidential election. The investigative route began with Tunja and the candidacy of Abelardo de la Espriella. In the 2026 interior corpus, the combination of candidate share and vote load shifted the margin by +432,826.77 toward Abelardo: the observed margin was +121,084, while the equal-weight polling-table comparator yielded −311,743. The 2022 and 2018 analyses were incorporated afterward as historical controls to determine whether the structure was new or already present.
Explained for a child: four boxes of oranges in a store
Imagine a store with four boxes in a row. Boxes 1 and 2 are large; boxes 3 and 4 are small. Some oranges carry a yellow sticker and others a blue one:
| Box/table | Size | Yellow | Blues | Yellow percentage |
|---|---|---|---|---|
| 1 · early | 100 oranges | 70 | 30 | 70 % |
| 2 · early | 100 oranges | 60 | 40 | 60 % |
| 3 · late | 50 oranges | 20 | 30 | 40 % |
| 4 · late | 50 oranges | 15 | 35 | 30 % |
Sum of real boxes
Yellow: 70 + 60 + 20 + 15 = 165.
Blue: 30 + 40 + 30 + 35 = 135.
Margin: yellow +30.
If each box had 75 oranges
Yellow: 52.5 + 45 + 30 + 22.5 = 150.
Blue: 22.5 + 30 + 45 + 52.5 = 150.
Margin: tie.
Now let's do the subtraction: yellow goes from the balanced comparator from 150 to 165, that is, 165 − 150 = +15. Blue goes from 150 to 135, that is, 135 − 150 = −15. Since the total remains fixed at 300 oranges, the 15 that yellow is above the comparator are the same 15 that blue is below. That is why the distance between the two does not change by 15, but 30: (165 − 135) − (150 − 150) = 30.
In electoral language, if yellow represents candidate A and blue represents candidate B, the box architecture leaves A with 15 more votes and B with 15 fewer votes than under the equal-size comparator. No vote already cast was physically transferred from one candidate to the other. The effect arises because the shares most favorable to A coincide with the large polling tables, while the shares most favorable to B coincide with the small ones. Large polling tables multiply their percentage by a larger number of votes when they enter the national sum.
Why do we say that they “weigh more”? Not because an orange—or a vote—has greater value. A box “weighs more in the sum” only because it contributes more units: a table with 100 votes contributes twice as much as a table with 50.
What is inside the actual box? In the 2022 comparison, the box contains the Hernández + Petro votes recorded at each polling table. A table can contribute more because more eligible citizens were assigned to it, because turnout was higher, or because the final “tail table” received a smaller remainder of voters. Separating these explanations requires joining the assigned electoral roll, E-10, E-11, turnout and results for the same table.
The Footprint measures how many votes were inside each box and how they related to their political position and proportion. The census and documentary chain must explain why the initial boxes were systematically fuller.
That is the heart of the Footprint: Early tables concentrate a more favorable proportion and, at the same time, contribute a greater number of votes to the total. The coincidence between position, percentage and amount contributed shifts the final sum.
The three questions that were previously mixed
O · do neighboring tables resemble one another?
It compares each polling table with the next one. If neighboring tables maintain similar shares, there is local continuity.
S · Where does it go down or up?
See if the candidate's ratio increases or decreases as the table number advances.
Who receives the weight?
Checks whether the relatively more favorable tables are also those that contribute the most votes and how much that changes the margin.
The historical control of 2022 that allows the mechanism to be explained
FITE followed 103,295 exact administrative keys across the first and second rounds. After the necessary filters, the confirmatory test used 89,718 polling tables in 5,291 polling places.
At those tables, the early half contributed an average of 242.55 Hernández+Petro votes per table; the late half contributed 201.38. At the same time, Hernández's relative increase between rounds was 11.85 percentage points in the early half and 2.55 points in the late half.
Combining the amount contributed and the relative change produced a margin shift of +252,655.30 toward Hernández. The 999 reorderings within polling places placed the central 95% between approximately −4,457 and +4,278; the observed result lay outside that range. After excluding the overseas constituency completely, the effect remained +251,871.11.
Why didn't Hernández win then?
Because an arithmetic advantage within the table architecture can be great and still not exceed all the votes that the opponent gathered in the country. In the comparable universe of 2022, the transversal effect towards Hernández was approximately +439,807 margin; It was material, but insufficient to reverse Petro's national advantage.
Why the 2026 test is strengthened by history
The cross-sectional effect presented for 2026 shifts toward Abelardo de la Espriella in the interior corpus (+432,826.77). Going back in time, the same architecture shifts toward Rodolfo Hernández in 2022 (+439,806.81) and toward Iván Duque in 2018 (+382,115.90), always with 999 reorderings and p=0.001 before multiple-comparison correction. The historical sequence makes 2018 and 2022 persistence controls for the central 2026 finding.
Serious conclusion: The official numbering of the tables is reproducibly associated with two variables that act together—the political proportion and the number of votes—and this association materially shifts the margin between candidates. The institutional explanation must reconstruct the assignment of citizens, the census by table, E-10/E-11, the three E-14, FRT, MMV/MMS, E-24/E-26, the versions and the intervention of each operator.
Version and custody
Integrity Manifesto
Version: 0.4.1 · State: ready for judicial delivery · Class: JUDICIAL DELIVERY · Canonical author: Leonilda Orrego Viera.
The SHA-256 hashes of all files are located in MANIFEST_SHA256.csv. a hash It works like a fingerprint: if the file changes, its fingerprint changes.
Actual technical limit: the system registers
authorization, version and content, but cannot prevent a
authorized person photograph or copy what you already received. That's why
combine minimal exposure, attribution, registration and
usage restrictions.
Job ID: FITE-COAD-JURADOS-007 · Canonical entry: OPEN_HERE_FITE.html
General conclusion of the book
The integrated conclusion of the book is this: the modules document different technical facts, but they all require being able to follow and verify the same chain, from the table and the minutes to the published file and the declared result. The tour shows a public figure whose construction was reproduced; versions of documents that are not digitally identical; technical information that no longer accompanied the file; a visible identifier that does not itself verify the final PDF; localized differences in records that should describe the same table; territorial and table sequences that persist after controls; and seven cases that require official documentation of appointment, attendance, replacement and signature. Recent institutional responses also locate where several of the pieces necessary to complete this verification are preserved. For this reason, original sources, lists, calculation and assignment rules, versions, historical states, designations, E-11 and technical records are requested: these are the elements that allow the documentary and technological chain to be reconstructed step by step.
Executive conclusion · The question that the Court must resolve
An electoral chain is only reliable when it can be verified
The contribution presents documentary, arithmetic and territorial findings that require reconstructing the chain between table, document, data and result. FITE provides preserved evidence and requests from the Court the institutional records necessary to complete each specific verification.
Evidence Integrity Supplement
Custody, version and traceability of this delivery
Before reading a discovery, one must know who is responsible for it, what version is being read, what sources support it, and how its traceability is maintained.
Author and responsible custodian
Leonilda Orrego Viera. FITE retains authorship, custody and the change history of the documents that make up this intervention.
Current version
v0.4.1 · Correction, update and evidentiary expansion. The initial autonomous intervention was submitted on August 11, 2026; this version incorporates the 104/104 control, the 103,364/103,295 correction, the separation between continuity, direction and margin effect, the 104,782-table trajectory with 9,999 reorderings, equal-weight, common-place, E-11 and documentary-state controls, the operational-contractual chapter, the glossary and the Integrated Report on the Documentary Chain and the Order Footprint. Its transmission must be entered in the custody ledger.
What the reader receives
The supporting document, the pedagogical manual, the sources necessary to verify the specific findings and the interactive judicial reading stations.
Custody trail
Authorship, sources, versions, hashes and the change log allow each statement to be traced back to its documentary support.
Attribution rule
Every figure, table, conclusion or source must retain its identifier, version, date and link to its original support.
Access and use
This delivery is reserved. It is intended for judicial action and the competent persons to verify it. It does not authorize editing, transformation or redistribution without express authorization from the author.
Custody principle: FITE retains the authorship, versions and traceability of each finding. The integrity of the evidence is maintained through hashes, change logging, and accurate references to its sources.