Autonomous and independent intervention with a technical and documentary contribution
Leonilda Orrego Viera
FITE research on documentary traceability, electoral data and publication patterns
Package version: 0.4.1 · State: current judicial-reading version · Initial filing: August 11, 2026.
Executive summary for the office
This intervention maintains that the election should be reviewed because FITE found a material breach of integrity and traceability in the documentary and technological chain supporting the electoral result.
An electoral result is not born as a final figure. It begins at a polling table, passes through a physical form, is digitized, published, converted into data, consolidated and finally declared. When original data disappear, digital fingerprints change, parts of a form are separated, structured fields are transformed and repeated territorial patterns appear, the Court faces an evidentiary chain that requires judicial reconstruction.
What FITE found: massive loss of verifiable metadata; E-14 documents reprocessed or published with less traceability; versioning of files abroad; families of repeated transformations in MMS/MMV data; changes in documentary representation; patterns of streaks and territorial blocks; inconsistencies that require E-11, E-14, E-24, E-26, logs and original databases; and a census case where a public figure was reconstructed and reduced from 751,156 to 384,215 votes by eliminating double and triple municipal counting.
Central request: protect the collective rights and interests invoked in the popular action through the preservation, exhibition and comprehensive contrast of the documentary and technological chain of the election, table by table, document by document and data by data.
Visual map of the findings
Open the interactive presentation to walk through each file in simple language, with figures, comparisons and diagrams.
1. Appearance
I, Leonilda Orrego Viera, appear as an intervener under Article 24 of Law 472 of 1998 to contribute to the protection of the collective rights and interests invoked in the action and to submit a technical, documentary and citizen investigation into the traceability of the electoral process.
Law 472 allows any natural or legal person to intervene in a popular action before the first-instance judgment and establishes that the intervention has prospective procedural effect. This filing joins the proceeding so that the Court can understand, preserve and verify the facts linked to the requested collective protection.
This intervention is submitted autonomously. It delivers an orderly, understandable and verifiable evidentiary record built from public information, preserved sources, human comparisons and technical analysis assisted by digital tools.
Trace of institutional obtaining of evidence
FITE's investigation does not start from an isolated statement. In the face of each gap that prevented following the documentary and technological chain, the available public sources were preserved and the corresponding institutional requests were responded to. This route is part of the traceability of this test: it shows why certain pieces could be directly verified and why others remain in the custody of the authority.
In particular, the applications submitted to MinTIC on July 6 and 21, 2026 asked about infrastructure, interoperability, cybersecurity, monitoring, connectivity and technical support related to publication, transmission and electoral consultation. The institutional response based 262124517, of July 27, 2026, reported that MinTIC carried out actions to articulate the PMU Ciber and control communications during election days, and attached reports from the National Electoral Guarantees Plan. The established scope 262126721, of July 29, 2026, expressly referred to that response and its annexes.
These responses delimit a verifiable part of the institutional architecture: monitoring mechanisms and follow-up documents existed. At the same time, they confirm that individualized histories of versions, uploads, objects, identifiers, URLs and technical managers must be required from the custodians that manage them. That is why the evidentiary requests in this intervention are concrete: they identify documents and records whose necessity emerged during FITE's progressive reconstruction.
The registry status responses preserved for individualized juror cases are linked to their respective records. Its content is not used to affirm attendance or effective performance without the official designation, replacement, attendance and signature documents requested from the Court.
Origin and continuity of the research
The findings, evidence matrices, reproducible operations, control sheets and pedagogical explanations presented in this contribution come from FITE's own research directed by Leonilda Orrego Viera. Each statement is kept linked to its source, cut-off date and verifiable support. External collaborations, when incorporated, are expressly identified and retain their own origin.
2. What FITE found
FITE found a sequence of traceability breaks that connect to each other. The central finding is that an election is not verified by looking only at the final result. It is verified following the complete chain: physical record, scan, digital file, fingerprint, published data, consolidated forms, territory and result.
FITE found that this chain was weakened at points where the law requires integrity, authenticity, publicity, conservation and possibility of verification. The consequence is serious: if the public document loses its technical history, if the file changes its footprint, if the pieces are separated and if the data is transformed by families, the system is exposed to manipulation without the citizen being able to notice it from the public viewer.
| No. | Finding | What was found | Why it is serious |
|---|---|---|---|
| 1 | Jurors and alternates | Seven individualized nominal correspondences. | Designation, attendance, replacement and signatures must be reconstructed table by table. |
| 2 | Municipal census calculation | A public figure of 751,156 is reconstructed as 384,215 when unique municipalities are counted across 256 cases. | The list, source, formula, date and version must be disclosed. |
| 3 | E13-21 | Twenty-one pairs and 42 pages preserve a repeated visual change. | The version and representation history must be reconstructed. |
| 4 | E4-Meta | One table appears in two “not digitized” representations and in one readable E-14. | The authority must certify the link among code, URL, file and upload. |
| 5 | USA-987 versioning | Across 987 pairs, 587 hashes differ and 400 are identical. | A different hash proves that the digital files are not identical and requires their version history. |
| 6 | USA-303 | Three hundred and three preserved codes require individualized linkage across layers. | Official keys linking table, bulletin, E-14, E-24 and E-26 must be provided. |
| 7 | USA continuity / HNI-0 | Visible identifiers, bytes and SHA-256 were compared across the preserved universe. | The institution must define the identifier and its relationship to the final file. |
| 8 | MMS/MMV families | Four control universes answer different questions and retain separate denominators. | Data dictionaries, validation rules and load versions are required. |
| 9 | MMS-65 | In 65 tables, removing one digit from the MMS field reconstructs MMV exactly. | The field schema and generation process must be explained. |
| 10 | MMS-834 | Eight hundred and thirty-four cases were classified by size, field and event without adding overlapping classes. | The complete dataset, exclusions and update traces are required. |
| 11 | MMS-1226 | An expanded family shows repeated forms of difference in a specific control field. | The universe, territorial key and inclusion criterion must be provided. |
| 12 | MMS-301 | Two independent reviews converge on 301 tables. | Each review and every intersection must be documented. |
| 13 | Rounds, territory and Tunja | Paired controls, territorial order and streaks appear within comparable units. | Both-round data, official order and comparison rules are required. |
| 14 | Order Footprint | Continuity between neighboring polling tables, direction along the sequence and margin effect were measured separately; in 2026 the cross-sectional effect shifts the margin by +432,826.77 toward Abelardo. | Census/DIVIPOLE, E-10/E-11, numbering, loads, remappings, exceptions and logs must be disclosed. |
3. Core of this intervention: verifiability
FITE's contribution does not consist of repeating a political narrative or discussing electoral preferences. It consists of bringing a verifiable question to the process:
The preliminary answer from the research is that this path appears incomplete in several layers. The problem is not limited to an altered file, a dubious table or a striking figure. The problem is the connected sum of ruptures: the document loses source data, the file changes its footprint, the page is separated, the structured data is transformed, the territory shows streaks and the census requires an explanation of the source and formula.
That is why this document asks that the complete chain of verifiability be reconstructed and that the break in that chain be assessed as a substantial impact on the documentary integrity of the electoral process.
The chain of verifiability
| Stage | judicial question | Risk found |
|---|---|---|
| Table | Who voted and who acted as jurors? | E-11, jurors, replacements, signatures and attendance remain to be crossed. |
| E-14 physical | What was really written on the table? | It must be compared with a complete image and corresponding signatures. |
| Digitization | When and how did paper become archival? | Absent or reduced metadata prevents verification of origin from the public PDF. |
| Published file | Is it the same file or change between versions? | Different hashes and versioning require technical history. |
| Structured data | Does the system reflect what the minutes say? | MMS/MMV families show repeated transformations. |
| Consolidation | Does the final result correspond to the set of tables? | Streaks, blocks and territorial concentrations require statistical control. |
4. Central thesis
The declaration of election must be examined in light of a documentary and technological chain that presents traceability ruptures, loss of technical information, repeated transformations in data and documents, and territorial patterns that cannot be understood as isolated facts.
The investigation is not based on a single file or a single anomaly. It is based on a sequence: first the loss of traceability appears; then file and representation changes appear; Afterwards, structured data transformations emerge; Finally, territorial patterns show why those layers should be reviewed together.
5. Expected legal string and found string
The law does not protect a loose image. Protects an electoral document capable of proving its origin, its integrity, its authenticity and its route. In a modern choice, the digital E-14 is not a decorative photograph: it is a data message with evidentiary value. That is why it must allow us to know where it came from, when it was produced, if it changed, who published it and how it is connected to the real table.
The legal purpose of scanning and publishing the minutes at the conclusion of the table scrutiny is to preserve immediate, complete and verifiable evidence. If the published file loses its metadata, does not retain a verifiable fingerprint, separates votes and signatures, changes between versions or appears reprocessed without public traces, we are not facing a simple aesthetic failure: we are facing the breaking of the locks that make the electoral document reliable.
Law 527 of 1999, articles 8, 9, 10, 16 and 19 - original, integrity, proof and attribution: The original requirement can be satisfied by a data message when there is a reliable guarantee of integrity and the information can be displayed; integrity requires that it remain complete and unaltered; the data message is admissible as evidence; Its attribution may depend on a person or a programmed system; and duplicates must be identifiable. Applied to digital E-14, visual appearance is not enough: identity, integrity, origin and version must be able to be verified.
Law 270 of 1996, article 95 - technology at the service of the administration of justice: The administration of justice can and must use technical, electronic, computer and telematic means, but this technology must serve the public function, not destroy verifiability. If an electronic document enters the judicial debate, it must arrive with conditions that allow control, consultation, conservation and evidentiary value.
Law 1475 of 2011, article 41 - scanning and immediate publication: At the end of the table scrutiny and after reading the results, authorized persons must scan the minutes to publish them immediately. Ruling C-490 of 2011 explained that speed, complete publication and delivery of copies reduce opportunities for manipulation and alteration between the initial accounting and the final proclamation. That is why the continuity between minutes, scanning, publication and versions is not a technical detail: it fulfills a constitutional control function.
Law 1712 of 2014, articles 2, 7, 11 and 28 - maximum advertising and usable data: The information in the custody of the obligated subject is public except legal reservation; must be available by electronic means; information of public interest must be complete, reusable and actionable; and it is up to the entity to justify a restriction. In electoral matters, publishing does not just mean showing an image: it means allowing access, downloading, comparison and effective citizen control.
ISO 32000 and ISO 19005 PDF standards: The PDF is not a simple image pasted on a page. It is a technical structure of an electronic document. In archival and testing contexts, the PDF/A family of ISO 19005 seeks long-term preservation, and the ISO 32000 family defines the structure of the PDF format. This matters because metadata, structure, pages, objects, compression, producer and layers of the file are part of the technical life of the document. If they are deleted or degraded without explanation, traceability is affected.
Legal consequence: If the digital electoral document loses the elements that allow verification of origin, integrity, authenticity and inviolability, the minimum trust standard of the electronic document is violated. Manipulation is not just about changing a visible number. It also consists of disassembling the mechanisms that allow us to discover if a number, a page, a signature or a file were changed.
| What must be verifiable | Human equivalent | What was found by FITE |
|---|---|---|
| File origin | Who produced the document and with what tool. | In numerous E-14 of 2026 the original technical information appears empty or reduced. |
| Moment of creation | Date and time when the paper was converted into a digital document. | The absence of metadata prevents verification of the generation time from the public PDF. |
| Integrity | That the document has not changed without leaving a trace. | In the USA, 587 files with different fingerprints were documented out of 987 analyzed. |
| documentary unit | That vote, table and signatures belong to the same form. | Separations and changes in representation were observed that require technical correspondence. |
| Data correspondence | That what is visible in the record matches the fields in the system. | The MMS/MMV families show repeated transformations that require explanation of origin. |
| Territorial behavior | That the results per table behave like a human choice, not like a repeated rule. | The streaks and blocks require statistical contrast and table-by-table documentation. |
Raised violation: The publication of electoral documents without sufficient traceability, with changing fingerprints, separate pieces and data transformations compromises the guarantees of integrity, inviolability, authenticity and effective publicity of the electoral document. The relevant manipulation does not have to appear as a visible erasure; It may consist of breaking the controls that allow the change to be noticed.
6. Compromised rights
The electoral process does not only produce a result. It also produces public information. This information makes it possible to exercise citizen control, verify the flow of documents and participate in an informed manner in political life.
The Transparency Law establishes that publicity is the rule and that secrecy is the exception. Its article 28 establishes that the burden of proving a reservation or restriction falls on the obligated subject. In electoral matters, this means that the authority cannot replace traceability with a general declaration of normality: it must provide the information that allows the process to be verified.
In this case, electoral transparency protects three specific rights:
- Right of access to public information: know the documents, databases, codes, forms and data that explain the result.
- Right of political participation: not only vote, but understand how the citizen will was counted, published and consolidated.
- Right to citizen control: be able to download, machine read, compare, audit and query for irregular patterns without relying on a fragmented visual version.
When information is published without complete formats, without sufficient metadata, without version histories, without downloadable databases and without clear correspondence between document and data, these three rights are materially weakened.
Applicable framework: Law 1712 of 2014 on maximum publicity, availability, actionable data and burden of justifying reservations; Law 527 of 1999 on originality, integrity, attribution and evidentiary force of the data message; Article 41 of Law 1475 of 2011 on scanning and immediate publication; Law 270 of 1996 on technology at the service of justice; and Ruling C-490 of 2011 on the function of these guarantees to reduce opportunities for manipulation and alteration during the scrutiny.
7. How to read the test without technical language
A digital document can be explained as a sealed product. It is not enough to see it from the outside. You must know if it retains its packaging, its exact weight and its factory seal.
| Element | Simple example | Application to E-14 |
|---|---|---|
| Metadata | They are the packaging: they say who did it, when, with what tool and under what conditions. | If the packaging is blank, the origin of E-14 cannot be easily reconstructed. |
| Hash | It is the exact factory weight: if the product changes, the weight changes. | If the hash changes, the PDF is no longer exactly the same file. |
| time stamp | It is the external seal that confirms when the document existed. | It allows you to know if the E-14 was created and preserved at the correct time. |
| documentary unit | The product comes complete, not with loose parts. | Voting, table, pages and signatures must correspond to each other. |
| Structured data | It is the table where the system saves the document values. | It must match what is seen in the minutes. |
| Streak | It is drawing many cards of the same color one after another. | If a sequence is repeated too much, we test whether it could be random or if there is a rule. |
The historical comparison between elections was used as documentary control: allows you to see if the way of publishing, identifying and preserving forms changed between periods. Candidates from different elections are not compared; The architecture of the electoral document is compared.
8. Continuity of research
Continuity is the center of this contribution. FITE did not first find a conclusion and then look for data to support it. The path was the opposite: each piece of information led to the next control.
The thread of the investigation is this:
- Document: First, we reviewed whether the published E-14 preserves its origin, date, tool and route.
- Fingerprint: Then it was compared whether the same file maintained the same fingerprint or had changed.
- Image: Afterwards, it was checked if the change was visible: quality, brands, page, signatures or structure.
- Fact: then it was crossed whether the structured fields of the system repeated transformations.
- Territory: Finally, polling tables, polling places and municipalities were organized to detect streaks, blocks and concentrations.
- Institution: The route ends in what only the authority can deliver: original bases, logs, versions and official correspondences.
The methodological unity is an essential part of the evidentiary value: separating the findings as if they were loose documents prevents us from seeing the relationship between source, data, document, territory and result.
9. Development of the findings
9.0. Three E‑14 copies and three documentary chains
Official fact: after the count, polling jurors record the result in three identical E‑14 copies—Transmission, Delegates and Claveros—sign them at the polling table and then separate them. Transmission feeds a voice route to the CRT and FRT, in addition to any digitization; Delegates follows collection, scanning and citizen-publication routes; Claveros travels with the electoral records through E‑17, E‑19, E‑20, the three-key chest and the scrutiny commission to minutes, E‑24 and E‑26.
Why it matters: after separation there are three physical objects and several digital transformations. An audit cannot treat them as one PDF or confuse the Transmission copy with MMV/MMS. Every comparison must retain election, department, municipality, zone, polling place, polling table, copy type, page, capture and version.
Evidentiary conclusion: the three routes make it possible to locate a divergence. If all three copies agree but the FRT or bulletin differs, review is directed to transmission, reception, capture, validation or consolidation. If the copy or pre-separation witness photograph agrees with the data but the published PDF changes, review is directed to digitization, assembly, upload or versioning. If a value changes during scrutiny, its decision and support must appear in minutes, recounts, E‑24/E‑26 and software records. A difference may arise at one relay or propagate through several; the institutional evidence must identify which occurred.
Linked request: provide all three complete copies and pre-separation photographs for each polling table; certify the correspondence of values, pages and signatures; and separately disclose (i) E‑14T, call, FRT, capture, alerts, corrections, processed row and bulletin; (ii) Delegates E‑14, source file, scanner, workstation, operator, batch, hash, upload and versions; and (iii) Claveros E‑14, E‑17, E‑19, E‑20, custody, minutes, recounts, E‑24/E‑26, system entry, image and versions.
9.1. Loss of technical traceability in E-14 documents
Observed fact: In documents E-14 of 2026, the absence or reduction of technical information of origin is observed. In comparisons with previous documents, old files retain more data about their generation, while new ones appear as reprocessed images with less verifiable capacity.
Why it matters: A digital document is not just what you see on the screen. It also has origin information. If this information disappears, the citizen and the judge lose the ability to know when the file was produced, with what tool, by what route it passed and if it was modified after its generation.
Evidential conclusion: The public E-14 stopped offering sufficient information to reconstruct its origin and route on its own. This loss constitutes an objective breach of traceability: the published file does not allow to fully determine whether it was reprocessed or replaced, nor to identify from the public viewer the date, the user and the responsible process.
Linked request: display of original files, native metadata, scan logs, upload logs, publishing user/process and substitution traces.
9.2. Separation of parts of the electoral form
Observed fact: In the material analyzed, forms or components appear published in separate blocks, including cases where the voting information and signatures do not appear integrated as a single visible documentary unit.
Why it matters: If the vote and the signature are separated, the integrity of the document depends on a technical relationship that must be demonstrable. It is not enough to see a signature on one side and a vote on the other; It must be proven that both pieces belong to the same table, at the same time and to the same document.
Evidential conclusion: When voting and signatures do not appear as a verifiable documentary unit, the material possibility of recombining pieces opens up: a vote with some signatures, or some signatures with another vote. This possibility is precisely the risk of manipulation that the documentary chain must prevent.
Linked request: Technical relationship between voting page, signature page, table code, source file, hash and consolidated form.
9.3. Versioning and changing fingerprint
Observed fact: In the United States universe, 987 documents were analyzed. Of them, 400 kept the same fingerprint and 587 had a different fingerprint. Within that universe, 266 cases were prioritized for visual review; 191 presented a relevant loss of quality and 223 showed affected numerical areas. These categories are layers of analysis, not numbers to add.
Why it matters: when the fingerprint changes between two captures of the same polling-table reference, the file is no longer exactly the same. The comparison remains within the same route and reference; it does not treat the three distinct physical copies as versions of one another. The fingerprint does not by itself identify the transformed content or process, but it proves an objective digital difference.
Evidentiary conclusion: 587 pairs with different fingerprints within a universe of 987 establish digitally non-identical captures of the same polling-table reference. The change requires the history of the same copy or route: what changed, when, through which process and whether it affected readability, quality, pages or numerical zones. Correspondence among Transmission, Delegates and Claveros is tested separately through values, identification, pages and signatures.
Linked request: full US version history, hash manifests, release dates and before/after files.
9.3.1. Institutional background on foreign affairs, E-11 and technical records
Fact documented by the authority: Through official letter RNEC-S-2026-0119086, the Electoral Census Directorate reported that the official potential by table, country, consulate, position, day and day of the United States, Puerto Rico and abroad is an integral part of the Electoral Census. The same official letter indicated that the E-11 forms for the 2026 presidential elections were being digitized and processed by the IT Management.
Fact documented by the CNE: The documents CNE-I-2026-006215 and CNE-I-2026-006216 sent the hashes, manifests, logs and integrity controls to the Information Technology Office of the CNE; and the matters of auditors, witnesses, accesses, downloads and channels authorized to its Electoral Surveillance and Inspection Directorate. In turn, the CNE transferred to the Registry the request for correspondence by table between census, voters, TEMIS, E-14, JSON, E-24 and E-26 abroad.
Scope of delivery and reservation invoked: Before receiving the official file, FITE had already preserved 253 operational polling places overseas. Comparison with the institutional spreadsheet confirmed those 253 polling places with no numerical differences in women, men, electoral potential or number of polling tables; the spreadsheet added one zero-value row. The RNEC provided DIVIPOL, country or municipality, polling place, potential by sex, total and number of polling tables, but not the detail by polling table, date or voting day, nor the requested E-11 forms. The entity stated that it would not provide individualized or disaggregated information capable of directly or indirectly identifying registered citizens, invoking Article 213 of Decree 2241 of 1986. That reservation defines the publicly available layer; verification of the documentary chain still requires the strictly necessary records under safeguards established by the Court.
Evidential scope: These documents identify competent agencies and institutional sources whose delivery or explanation was requested: census by table, E-11, correspondence tables, manifests, hashes, logs and integrity controls. The substantive responses on several of these points remain pending. The missing information is not an abstract requirement, but rather the concrete support necessary to complete the external and documentary verification table by table.
Linked request: directly require the identified agencies to preserve and control the records sent by jurisdiction, with their dates, versions, persons responsible, and integrity support. When it is appropriate to apply confidentiality, the exhibition must allow the Court and the authorized experts to verify the discovery without disclosing more information than is strictly necessary.
9.3.2. Human control of 104 United States E-14s
Control origin: an automatic extraction opened 104 apparent differences between the preserved E-14s and the precount. FITE did not convert these alerts into documentary conclusions: it visually reviewed the first page of the 104 PDFs and carried out a second check on the cases that were still open.
Current result: 104/104 minutes reviewed; 101 matches with the precount; 3 documentary discrepancies; 99 cases in which automatic extraction did not match visual reading. Candidates correctly identify themselves as Ivan Cepeda Castro and Abelardo de la Espriella.
| Table | Visual reading E-14 | Reference precount |
|---|---|---|
88_360_40_02_056 | 2 / 10 | 2 / 11 |
88_360_40_81_013 | 12 / 400 | 12 / 480 |
88_360_40_85_004 | 26 / 332 | 25 / 332 |
Linked request: incorporate the three tables into the collation of bulletins, versions, E-14 of delegates, keys and transmission, E-24/E-26, publication events and correction records; also preserve the 104/104 record to demonstrate control of false positives.
9.3.3. Technical and contractual route of the TEMIS environment
Origin of the research: The preservation of non-identical copies of the same reference and the appearance of repeated structural transformations led to the question of which technical service processed and published these versions.
Documented route: the preserved public configuration allows following TEMIS towards a production domain under prdtpssas.com, AppSync and Cognito services and an observable AWS region. The intersection with electoral contracting leads to contract 049/2025, the UT ILE 2026 and the documented technical functions of Thomas Processing & Systems in platforms, electoral data processing and digitization and publication of minutes.
Evidential scope: The chain attributes the technical and contractual ecosystem within which the documents are served. Identification of the automatic job, account, user, or service that produced a particular version depends on object histories, execution and deployment records, technical identities, batch jobs, tickets, and authorizations under institutional and contractual custody.
Linked request: preserve and display domain and AWS account ownership, AppSync and Cognito identifiers, object and version histories, execution and access logs, automatic jobs, technical roles and authorizations, with individual correspondence to each PDF and published version.
9.4. Repeated transformations on MMS/MMV structured data
Observed fact: Families of repeated transformations were identified in structured fields related to table information. MMS-65, MMS-834 and MMS-1226 are connected but different universes; The MMS-301 intersection shows 301 common tables between MMS-834 and MMS-1226, with exclusive areas that prevent them from being treated as a single total.
Why it matters: When the same type of transformation appears in many territories, the question stops being whether there is an isolated error and becomes what process produced that repetition: capture, conversion, validation, loading, transmission or publication.
Evidential conclusion: The repetition of transformations in structured fields shows an operational rule or repeated process. This is not an isolated table. When a field is transformed in a recognizable way into different territories, the manipulation stops being just visual and enters the data layer.
Linked request: original structured bases, field dictionary, validation rules, load versions and update traces.
9.5. Changes in visual and documentary representation
Observed fact: Changes in representation between versions or families of documents were documented: changes in tone, structure, marks, visible page, non-digitized page and alterations in how the document is presented to the citizen.
Why it matters: Not all change is of the same type. An image change, a metadata change, and a structured data change belong to different layers. When they appear together or close to the same universe, research must separate them and then relate them.
Evidential conclusion: The changes in representation demonstrate that the public document does not always maintain the same way of reading. This allows manipulation of the way in which the citizen, expert or judge sees the document, even when the system retains an appearance of normality.
Linked request: visual versions, source files, conversion parameters, reprocessing logs and correspondence with the table.
9.6. Territorial patterns, streaks and blocks
Observed fact: When ordering the results by department, municipality, zone, polling place and polling table, sequences appear that must be analyzed statistically. The Tunja case—whose line and tabular files were contributed by Frente Digital (@frentedigital2026), with independent verification, reconstruction and analysis by FITE—serves as a microcase because its universe is limited: 362 polling tables, identifiable manual corrections and six documents initially separated due to mathematical inconsistencies.
Why it matters: a table can have a strange result by chance. Many consecutive tables with repeated behaviors require a chance test. Statistics allow us to ask whether the frequency and magnitude of the streak are compatible with an ordinary election or whether they suggest a rule of behavior.
Evidential conclusion: streaks are not valued by intuition. They are assessed by frequency, length, location and comparison with controls. When a territorial concentration persists after correcting readings and separating doubtful tables, a sign of systematic behavior appears incompatible with a purely casual explanation.
Linked request: Table by table results in downloadable format, official order of tables, E-11, E-14, E-24, E-26, jurors, witnesses, biometrics and logs per hour.
9.7. Electoral census and aggregate calculations
Observed fact: The public figure of 751,156 votes associated with municipalities with electoral rolls higher than certain population thresholds was reconstructed. The reconstruction showed that accumulated groups were added as if they were independent. When counting each municipality only once, the margin observed in the 256 municipalities analyzed was 384,215 votes.
Source control: The DANE population census describes inhabitants; The electoral census of the Registry is a database of cards authorized to vote. The electoral roll is not a statistical projection. The ratio between the two is only interpretable if the date, version, territorial coverage and rule of each source are identified.
Necessary separation of universes: The investigation also identified two figures that do not describe that municipal margin. The figure of 2,419,734 corresponds to the difference between the complete censuses of 2022 and 2026. The figure of 885,400 was explained institutionally as a sum of projected quantities to organize the outdoor days from Monday to Saturday. They are different operations and cannot be added or used as if they were votes, new people or the same population.
Reproduced overseas control: The preserved documentation verifies an additional operational layer of 1,489 polling tables: 3,670 overseas TEMIS polling tables minus 2,181 polling tables in the overseas electoral-roll file. Independent row-by-row reconstruction of the 885,400 projected quantities requires the primary file Basic auditors president 2026, together with its dictionary, formula, version and hash.
Additional fact: The calculation compared electoral potential in 2026 with the DANE 2023 projected population. When updating the denominator, 94 municipalities changed category and 207 remained as robust cases.
Evidential conclusion: The census must be treated as its own line of evidence. The reconstruction of the figure of 751,156 votes demonstrates that an aggregate statement changes substantially when double and triple counting is eliminated and the date of the population base is controlled. The universe of qualified persons is a verifiable electoral layer in its own right and requires documented source, cut, formula and movements.
Linked request: electoral roll by municipality, registration movements, cut-off date, DANE source used, formula, database version and complete list of classified municipalities; together with a full copy of Basic auditors president 2026 and its reconciliation with DIVIPOLE, the nominal electoral roll, overseas voting days and the corresponding E-10, E-11 and E-14 documents.
9.8. Juries, remnants and human control census
Observed fact: Lists of jurors and remainder were reviewed with the support of automatic reading and subsequent human control. The review made it possible to detect cases that require institutional verification: names, IDs, designation, attendance, replacements, marital status and correspondence with the table.
Why it matters: The jury is part of the table's chain of trust. The appointment, attendance and eventual replacement must be supported by complete institutional records.
Evidential conclusion: A table's chain of trust includes authorized citizens, jurors, attendance, replacements, signatures, and witnesses. If this information is not provided in full, it is impossible to establish whether the table was guarded by the correct people and effective citizen control is broken.
Linked request: E-10, E-11, lists of jurors and remainder, appointment minutes, attendance, replacements and signature support per table.
10. Relationship between the findings
The fourteen findings are not separate stories. They function as a chain of control:
| Layer | Question that answers | Data that must be required |
|---|---|---|
| Document | Does the PDF retain its origin, date, tool and path? | Metadata, original file, upload history. |
| Fingerprint | Is the file identical between versions? | SHA-256, preserved versions, reason for replacement. |
| Visible content | What you see corresponds to the table? | Original image, signatures, voting, full pages. |
| Structured data | Does the system reflect what the form says? | Bases, codes, fields, validation rules. |
| Territory | Is the employer concentrated by area, position or municipality? | Table by table order, E-11, E-24, E-26, results. |
| Time | When was each data created, transmitted, changed or published? | Logs, events, logs, hashes per cut. |
10.1. Integrated report on the chain and the Order Footprint
Documentary incorporation: The integrative report is part of this contribution like document DOC-012 and develops the joint reading of the documentary, census, territorial and statistical stations. Its navigable version is found in the FITE Digital Judicial Book, chapter #integrador.
The chain that needs to be rebuilt is: census and DIVIPOLE → E-10/E-11 → table → E-14 → versions and representation → transmission and pre-count → MMV/MMS → E-24/E-26 → table order → result.
Explanation for a reader without statistical knowledge
The main controversy concerns 2026. FITE reached the Order Footprint while studying Tunja and the candidacy of Abelardo de la Espriella. In the 2026 interior corpus, the combined effect of polling-table load and candidate share shifted the margin by +432,826.77 toward Abelardo. The subsequent review of 2022 and 2018 serves as a historical control showing that the architecture recurs.
Imagine a store with four boxes of oranges. The first two contain 100 oranges each and the last two contain 50. If the first have 70% and 60% yellow stickers, and the last 40% and 30%, the sum gives 165 yellow versus 135 blue. If all the boxes had the same amount and kept those percentages, the result would be 150 against 150.
The actual sum is: yellow 70+60+20+15=165; blue 30+40+30+35=135. With four boxes of 75, yellow would be 52,5+45+30+22,5=150 and blue 22,5+30+45+52,5=150. The observed architecture leaves yellow 15 above and blue 15 below the comparator; the margin shifts 30.
The complete subtraction is: yellow 165−150=+15; blue 135−150=−15. Both comparisons keep the same 300 oranges. Therefore, if one color is 15 above the balanced distribution, the other necessarily remains 15 below; the distance between candidates changes 30: (165−135)−(150−150)=30.
If yellow represents candidate A and blue represents candidate B, the architecture leaves A with 15 more votes and B with 15 fewer votes than in the equal boxes comparator. An orange—nor a vote already cast—was not physically moved between colors. The displacement appears because the percentages most favorable to A were within the large boxes and the most favorable to B within the small ones.
Large boxes “weigh more” only because they provide more units; not because an orange has greater value. The same thing happens in an election: a vote maintains the same value, but a table with 250 votes contributes more to the sum than a table with 180.
Within each polling place, polling tables are numbered: table 1, table 2, table 3 and so on. FITE found that low-numbered polling tables tended to combine two conditions: a greater number of votes and a higher share for Hernández. As the official polling-table number increased, both variables decreased.
When those two slopes are combined, early polling tables contribute more to the final sum. The paired comparison of the same polling tables in the first and second rounds shifted the margin by +252,655.30 units toward Hernández. Petro's national victory means that the shift did not overcome the aggregate total; the measured arithmetic movement remains present within the polling-table structure.
The same general form—greater candidate share and greater load in early polling tables—appears first in the object defended here: 2026 toward Abelardo de la Espriella; and, as historical controls, in 2022 toward Hernández and in 2018 toward Duque. The institutional explanation must therefore cover how polling tables were created, filled, numbered and consolidated.
What's inside the actual box: votes actually recorded at the table. The box may be fuller due to the assigned potential, the participation or the treatment of the queue table. To distinguish these causes, the census must be joined by table, E-10, E-11, voters and results with the same key.
| Control | Documented result | Linked request |
|---|---|---|
| Exact coverage | Each round contains 103,364 unique keys; 103,295 match exactly. The remaining 69 keys in each round contain overseas zone remappings and are examined separately. | Deliver remapping records, DIVIPOLE history and transfer or renumbering acts. |
| Confirmatory universe | 89,718 polling tables in 5,291 polling places after requiring exact matching, both candidates present, a defined denominator and sufficient variation. | Deliver the official polling-table database, filters and numbering rule. |
| Continuity OR | Average lag-1 0.5606; 76.18% with p<0.05 under 5,000 permutations; 72.69% after FDR. | Deliver order, loads, totals and composition of each table. |
| S · Direction | Mean −0.65185; 96.54% of polling places are negative. Hernández's share is highest at the beginning of the official order and decreases toward the end. | Deliver the rule assigning voters and jurors to each polling-table number. |
| Paired margin effect | +252,655.30 toward Hernández; 999 reorderings, p=0.001; positive in 89.43% of polling places. Excluding the overseas constituency: +251,871.11, p=0.001. | Deliver the electoral roll, E-10/E-11, voters, MMV/MMS, E-24/E-26 and consolidation logs. |
| Mechanism | Early tables: 242.55 H+P votes; late: 201.38. Relative change: +11.85 versus +2.55 points. Both controls p=0.001. | Display capacity, effective allocation and treatment of final or partial tables. |
| Electoral-roll control | Growth from 2018 to 2022 does not absorb the paired change after controlling for polling-place size and department: ρ=0.0083, p=0.550. | Deliver historical electoral-roll cuts and movements by polling place and polling table. |
Legal function: The result identifies the stages, documents and records that must be preserved and exhibited so that the Court can verify the complete chain and decide the necessary measures to protect the collective rights invoked.
Open the complete integrative report in the FITE Digital Judicial Book.
11. Evidence provided and supports whose exhibition is requested
| Annex/support | State | Content | Purpose |
|---|---|---|---|
| FITE Digital Judicial Book and main filing | Submitted | Autonomous intervention, verification stations, navigable index and links to controlled sources. | Present a single path from each claim to its evidence and concrete request. |
| FITE pedagogical manual and glossary | It is contributed | Human explanation of fingerprint, metadata, traceability, MMS/MMV fields, streaks, blocks and matching. | Allow a verifiable judicial reading without advanced technical knowledge. |
| Matrix and index of findings | It is contributed | Relate each finding to source, operation, result, evidentiary scope and required institutional data. | Avoid loose documents and preserve evidentiary continuity. |
| Document representation and versioning files | It is contributed | E13 pairs, E4-Meta case, examples of preserved versions, hashes, visual comparisons and USA-987 universe. | Document digital differences, representation changes, and the need to reconstruct publication history. |
| TEMIS/AWS technical attribution sheet | It is contributed | Controlled route from TEMIS and the production domain to contract 049/2025, UT ILE 2026 and documented technical responsibilities. | Specify the responsible ecosystem and delimit the operational records necessary to identify the specific execution of each version. |
| MMS/MMV structured data records | It is contributed | MMS-65, MMS-834, MMS-1226 and MMS-301 crossover families, with their universe controls. | Require bases, validation rules and records that explain repeated transformations in control fields. |
| Territorial and human control files | It is contributed | Streaks and blocks, Tunja microcase, comparison between rounds and seven individualized cases of jurors or remnants. | Identify sequences that require table-by-table audit and designation, attendance, replacement and signature documents. |
| census file | It is contributed | Reproducible reconstruction of the public figure 751,156 and count by unique municipalities of 384,215, with control of population date and nested groups. | Require list, source, formula, version and official census movements used in the calculation. |
| Human control 104 E-14 USA | It is contributed | Row-by-row record of 104/104 minutes: 101 coincidences, three documentary discrepancies and 99 automatic extraction differences corrected by visual reading. | Individualize the three tables and demonstrate the quality control applied before formulating a documentary difference. |
| Confirmatory Order Footprint record | Submitted | Separation between continuity, direction and margin effect; 103,295 exact keys, 89,718 polling tables, 5,291 polling places and 999 reorderings, with controls for the electoral roll, overseas constituency and ordinal position. | Show in plain language how polling-table load and candidate share shift the margin and which institutional documents are needed to explain the mechanism. |
| Trajectory and control evidentiary expansion | Submitted | 104,782 polling tables in 6,116 polling places; 9,999 reorderings; cumulative trajectory; equal table weights; 4,118 common polling places; E-11, documentary and consolidation controls. | Show that the ordinal shape remains after removing explanations based on load, changing polling places and isolated alerts, and identify the need to disclose voter-to-table assignment. |
| Three-copy E‑14 matrix | Submitted | Official separation of Transmission, Delegates and Claveros; voice/FRT, digitization/viewer and custody/scrutiny routes; transformation points and required reconciliations. | Locate each difference at a concrete relay and determine whether it arose in one route or propagated through several. |
| Operational and contractual chapter | It is contributed | String census-DIVIPOLE-table-E-14-transmission-MMV/MMS-counting; Act 8135, contract 049/2025, Tunja contract 034/2026 and preserved SECOP files. | Relate each component to the person responsible, support, version, supervision and record that must be displayed. |
| Related institutional background | It is provided when they link a finding | Requests for information, responses, technical annexes and preserved actions that are directly linked to a source or a specific evidentiary request. | Specify which records remain in institutional custody and why they are required to complete specific verifications. |
| Originals, histories and records in institutional custody | Your exhibition is requested | E-10, E-11, E-14, E-24, E-26, original bases, transmission codes, logs, logs, hashes by cutting and loading, replacement and publication histories. | Complete the reconstruction of date, process, person responsible and scope of each documented difference. |
12. Evidentiary requests
- Order the immediate preservation of all E-14, E-11, E-24, E-26, structured bases, transmission codes, logs, hashes, logs and published versions.
- Require delivery of the original files and all subsequent versions, with date, time, user, process, server, system and reason for each upload or replacement.
- Sort table-by-table correspondence between visible document, structured data, consolidated form and disclosed result.
- Request records that explain the separation of pages, signatures, images and data, and the technical way to guarantee that they belong to the same table.
- Demand the complete base of the electoral census used for the process, its movements, cut-off dates, population sources and classification criteria.
- Order evaluation of territorial patterns using complete data, including polling tables, polling places, juries, witnesses, biometrics, schedules and publication records.
- Arrange for an independent judicial verification of the specific objects documented in this aid.
- Order the Registry, the UT ILE 2026 and its members to preserve and display the domain records, AWS account, AppSync, Cognito, object versions, deployments, accesses, automatic jobs, technical identities, tickets and authorizations necessary to attribute each specific processing.
- Require the rule, algorithm, parameters and versions used to allocate eligible identity cards among polling places and numbered polling tables, including electoral potential, the final tail table, exceptions, transfers and remappings.
- Require national and foreign inventory of Telephone Reception Centers, transmitters, receivers, lines, PBX or equivalent services, along with FRT, recordings, call logs, retries, contingencies, alerts and corrections by table.
- Require complete correspondence E-14 transmission → call → FRT → row processed → validation → bulletin, with date, time, user, station, software version and hash of each cut.
- Require all three complete E‑14 copies and the photograph taken before separation for each polling table; certify values, identification, pages and signatures; and maintain separate hash inventories and histories by copy type.
- Require the correspondence Delegates E‑14 → receipt → scanner/workstation → source file → batch → upload → viewer → versions, including operator, date, time, hash and replacement record.
- Require the correspondence Claveros E‑14 → E‑17 → transport → E‑19 → chest/E‑20 → commission → minutes and recounts → E‑24/E‑26 → published image, including envelope condition, responsible persons, dates and decisions explaining every variation.
- Require the UT ILE 2026 agreement, the proposal, the technical annex and the responsibilities matrix that identify which of its members executed each of the thirteen components, their subcontracts, personnel, systems, accounts, deliverables, SLA and supervisor.
- Require the minutes, inventories, hashes and supports of the freezing and unfreezing of source codes and executables, and their correspondence with the version actually deployed for pre-counting, consolidation, disclosure, scrutiny and digitization E-14.
- Require the inventory of portals, domains, URLs, APIs, buckets, CDNs, functional email accounts, certificates, endpoints and access regimes used to receive, process, custody and publish electoral information.
- Require by 2022 the contracts and annexes of UT DISPROEL 2021 and RNEC–INDRA 105/2021, with the exact assignment of levels of scrutiny, interfaces, consolidation, files and records produced.
13. Claims of the intervener
Based on articles 2, 24, 25, 28, 34, 74 and 75 of Law 472 of 1998, and within the framework of the collective rights and claims raised in the lawsuit, I request the office:
- Recognize this filing as an autonomous and independent intervention in popular action 25000-23-41-000-2026-00821-00 and admit the technical and pedagogical annexes submitted with it.
- Assess the findings as technical and documentary evidence of traceability breaks, digital differences, structured transformations and reproducible patterns within the chain of documents and electoral data.
- Order the testing necessary to reconstruct the complete chain of electoral documents and data.
- Determine, based on the complete evidence, how traceability ruptures, documentary separations, footprint changes, structured transformations and order patterns affect the collective rights and interests invoked.
- Order the precautionary measures and affirmative or prohibitory relief necessary to preserve the evidence, prevent its loss, complete independent verification and effectively protect collective rights.
- Provide in the sentence the specific orders of preservation, exhibition, reconstruction, correction, audit and guarantee of access that correspond in accordance with article 34 of Law 472 of 1998.
Section 14. Final conclusion
Technical conclusion of the intervener: The evidence provided demonstrates that the public chain that should allow each result to be followed from the table to its publication cannot be reconstructed completely and independently with the materials available to the citizen.
This conclusion does not rest on a general impression. It rests on operations that can be repeated: a public figure whose result changes when counting unique municipalities; files with the same reference that are no longer digitally identical; documentary representations that change; a control field that deviates from three concordant records by means of repeated shapes; sequences of tables and territory that remain after controls; and individualized cases whose verification requires designation, attendance, replacement and signature documents.
The seriousness of finding 14—the Order Footprint—requires an express response concerning the election defended here: 2026. In the 2026 interior corpus, low-numbered polling tables simultaneously concentrated a greater vote load and a share more favorable to Abelardo de la Espriella; that coincidence shifted the margin by +432,826.77 toward Abelardo. The backward review established that the same architecture was already present toward Hernández in 2022—including a paired margin shift of +252,655.30 units between rounds—and toward Duque in 2018. The years 2018 and 2022 are historical continuity controls; the main controversy and requested judicial protection concern 2026. The Court should order preservation, disclosure and explanation of the allocation rules, electoral-roll and voting records, and the complete documentary sequence.
Explanation for technical assessment: the verification separates whether neighboring polling tables resemble one another, whether candidate share changes as the numbering advances, and whether relatively more favorable polling tables also contribute more votes to the total. The contrast preserves the same polling tables and the same votes and compares the observed sequence with 999 within-place reorderings.
Subsequent evidentiary expansion: a second control family followed the complete trajectory of 104,782 polling tables in 6,116 polling places and performed 9,999 within-place reorderings. The curve reached the midpoint 5.3726 points ahead; the result remained after giving every table equal weight, across 4,118 common places in 2018, 2022 and 2026, and after E-11, documentary and consolidation controls. These tests narrow the pending institutional explanation to the rule assigning voters to numbered polling tables.
Decisive documentary clarification: the E‑14 does not follow one route after signature. Its three copies are separated and produce independent chains. The conclusion is therefore subjected to a stricter reconciliation: values, polling-table identity, pages and signatures across Transmission, Delegates and Claveros; version history within each copy; and linkage of each copy to its corresponding call/FRT/bulletin, scanner/viewer or custody/scrutiny route. This matrix locates whether a divergence arose at one point or propagated through several.
The consequence is concrete: to establish what happened between the table, the form, the file, the structured data and the consolidation, the Court needs to order the preservation and exhibition of the originals, versions, metadata, E-10, E-11, E-14, E-24, E-26, correspondence tables, validation rules and requested technical records. Without these supports, the authority retains decisive information and citizen control remains materially incomplete.
For this reason, I request that the technical evidence be evaluated comprehensively with the facts, collective rights and claims of the popular action, and that the necessary tests be carried out to reconstruct the electoral chain. The decision must be translated into effective preservation, display, audit and correction orders that restore the public verifiability of the process.
Democracy is not protected with a final screen. It is protected when the entire journey can be verified table by table, document by document and data by data.
Sincerely,
Leonilda Orrego Viera
Autonomous intervener
Identification: filed with the submission
Email: leonilda.viera@gmail.com
Normative reference sources
Law 472 of 1998, especially articles 2, 24, 25, 28, 34, 74 and 75; Law 1712 of 2014; Law 527 of 1999; Law 1475 of 2011, article 41; Law 270 of 1996; Sentence C-490 of 2011; PDF technical standards ISO 32000 and ISO 19005.