Typed input
Numbers, Hebrew text, dates, times, identities, roles and sources.
RTM LAB · RESEARCH PLATFORM
From documented records to a network that can be reconstructed and tested.
RTM Lab takes records, text, identities, dates and times; reconstructs Evidence, finds structures in Discovery, builds a dependency-aware network, and runs null-model tests to compare the observed network with alternative worlds under the same search freedom.
This page is public and indexable. The compute environment and private API remain access-controlled.
RTM Lab
● API ONLINE
01 Input02 Evidence03 Discovery04 Network05 QA06 Statistics07 Export08 AI Analysis
INPUT MATRIX
One finding, modeled as one information system
Text / contentDateTime
Identity / roleCore anchorSource
FROM RAW INPUT TO AUDIT
Eight stages — from structured input to an audited research report
The engine separates reconstruction, topology, statistical calibration and interpretation. You can see what was found, where it came from, what it depends on, how it behaves in Null worlds, and how an advanced AI analysis layer reads the frozen audit package after computation is complete.
Reconstruction
Legal paths are checked against frozen targets with an audit trail per closure.
Search
New structural relations are searched under the same locked search rules.
Topology
Closures become a network of nodes, families, bridges, convergence and dependency.
External QA
Manifest and validation tools separate source, representation, computation and interpretation.
Null & audit
Alternative worlds run through the same search contract and are compared structurally.
Export
Inputs, network, provenance and test settings can be retained for reproducibility.
Advanced whole-finding analysis
The frozen finding, Event Dossier and statistics are read together after computation, without changing gematria values, Null results or p-values.
RTM AI · analysis layerWhat the engine actually measures
The unit of analysis is the architecture — not a single numerical match
A network-level result asks what remains after duplicates and dependencies are consolidated, how many structural families connect, how much of the frozen input is covered, and how the pieces form one reproducible system.
≡
Dependencies are priced
Equivalent paths are not repeatedly counted as new evidence. The system consolidates dependence and keeps genuinely distinct evidence units separate.
Dependency-aware⌘
Roles are preserved
Identity and role are not freely interchangeable. Sender, recipient, date, time, text and core-anchor slots may obey different event-model rules.
Role-aware◇
The network comes first
The engine evaluates the system as a whole: anchor coverage, structural diversity, connection complexity and convergence. It can continue from initial closures into derived nodes, loops and representation convergence while retaining path provenance.
Network-first↻
Same search freedom
A Null world is not handed the observed equation. It receives the same search contract and may discover its own strongest eligible structure.
Matched search freedom✓
Complete audit trail
The system records what was fixed, what was resampled, which transforms were allowed, the seed, world count and observed network signature.
Audit trailT
Separate Tanakh lookup
A separate tool can search gematria values at verse, word or sequence level without automatically promoting external matches into evidence.
Tanakh lookupDERIVED STRUCTURES WITH PROVENANCE
The search continues beyond the initial closures
The engine remembers which anchors and paths produced each derived value. A structure is promoted only when explicit conditions of independence, recurrence and convergence are satisfied.Technical layer: Post-R41 · Provenance Memory
Search profiles
Three search profiles, three levels of freedom
RTM Lab keeps Standard, Strict and Super-Core profiles separate, including at the statistical stage.
STANDARD
Standard
Full search under the active search contract.
STRICT
Strict
A fixed profile with reduced search freedom.
SUPER-CORE
Super-Core
The hardest core of the locked search rules.
Important: profile-score p-values and the structural-network p-value answer different questions and are reported separately rather than multiplied.
Event and network testing against alternative worlds
The test freezes the rules of the game, not the observed equation. It determines what remains fixed and what may vary, generates an alternative event world, runs the same search process, and compares that world's strongest network with the observed network.
Time onlyExhaustive scan of all 1,440 legal minutes while the rest of the event stays fixed.Time-only
Date + timeDate and time are resampled according to their declared types and legal domains.Time + Date
Primary modelContent + date + time vary; identities and the core anchor may remain fixed according to the declared schema.Primary
Population identitiesA separate test requiring an explicit population model; it is never silently assumed.Population identities
WHOLE-NETWORK TEST
All structural layers are tested together inside the same Null world
Each alternative world reconstructs initial closures, derived nodes, recursive loops and representation convergence, then receives one joint statistic for the complete architecture. This produces one calibrated whole-network endpoint without multiplying separate p-values or assuming independence between layers.
Technical: Whole-Network Joint · V86 · R41 + R33/R34 + R35Why the Null matters
The question is not “did a closure exist?” but “how much freedom existed to find one?”
When many values, representations and operations are available, chance can produce attractive relations. RTM Lab attempts to price that search freedom inside the null model.
1
Same Search ContractObserved and null worlds use the same search rules.
2
Max-over-searchEvery null world may return the strongest network it finds.
3
Dependency preservedRepresentations derived from the same source do not become artificial independent evidence.
4
Exhaustive when feasibleSmall finite spaces are enumerated rather than approximated by Monte Carlo.
What a small p-value can say
- Under the specified null, networks at least this strong were rare.
- With frozen rules and anchors, it quantifies extremeness relative to that model.
- Sensitivity modes show which parts of the event carry the anomaly.
What it cannot say by itself
- It does not independently prove a physical mechanism or causal theory.
- It cannot erase undocumented retrospective search freedom.
- Zero hits in N simulations is not p=0; resolution and uncertainty remain.
Designed for auditability
Built so a finding can be inspected, not merely admired
RTM Lab keeps raw inputs, computation, topology and statistical inference conceptually separate, making it easier to reconstruct what actually contributed to a result.
A
Typed variables
Each anchor can be fixed, variable, dependent or derived, with an explicit role.
B
Provenance
Seed, ranges, null model, world count, search contract and observed signature are recorded.
C
Export
Exports support reconstruction outside the session in which the result was discovered.
Advanced AI after the audit
After computation: an advanced AI system analyzes the complete finding
Once the deterministic engine and statistics finish, the AI system receives a frozen package containing the anchors, closures, network, chronology, data provenance and Null results. It interprets and critiques what has already been computed — without recalculating gematria, changing p-values or replacing the statistics engine as the numerical authority.
Frozen finding→Statistics+Event dossier→Research AI report
FAQ
Frequently asked questions
Is RTM Lab a gematria calculator?
No. Gematria is one representation type. The central object is an information network made of anchors, roles, paths, dependencies, topology and null-model tests.
Why is the engine private?
This overview is public. The full compute environment and its API are access-controlled, separating public methodology from private research compute.
Does RTM Lab automatically “prove” RTM?
No. It provides infrastructure for reconstruction, audit and extremeness testing under declared models. Moving from statistical anomaly to a theoretical claim requires additional evidence, confirmatory prediction and external replication.
What is the difference between a profile score and the whole-network test?
A profile score summarizes search characteristics under one level of search freedom. The whole-network test reruns both initial closures and derived structures inside the same alternative worlds and computes one joint statistic. They answer different questions and are therefore reported separately.
Private compute · Public methodology
Ready to enter RTM Lab?
The public page explains the methodology; the full engine remains a private, authenticated research environment.