Original CryptoSignalsReview dataset research

Crypto Signal Provider Name Collision Report 2026

A repeated display name is a queue for entity resolution, not proof that records describe one operator. This report measures exact and normalized collisions while keeping that boundary explicit.

Analysis by CryptoSignalsReview Evidence Desk of 3,200 candidate records frozen 2026-07-05. Published 2026-07-10. Source commit 5dcd30b2b1a0. Dataset release. No provider paid for inclusion or status.

Review the collision findings

No signup, payment, wallet connection, or credentials are required.

Exact collision groups57

124 records share an exact display name.

Normalized groups77

172 records collide after conservative text normalization.

Largest group6

Maximum records sharing one normalized name.

Direct answer

The frozen inventory contains 57 exact-name collision groups covering 124 records, or 3.9% of the inventory. The versioned Unicode comparison key produces 77 groups covering 172 records, or 5.4%. 2 records contain no letter or number after normalization and remain unkeyable rather than being grouped together. A repeated name is not proof of a duplicate record, common ownership, impersonation, misconduct, or even the same service. It means the display name cannot serve as the identity key by itself.

Exact names and normalized names answer different questions

MeasureExact display nameUnicode-normalized nameInterpretation
Unique names3,1333,103The normalized method merges cosmetic variants, so it is deliberately more sensitive.
Collision groups5777Groups with at least two provider slugs.
Records in groups124172Records requiring another identity key before consolidation.
Excess records6795Records beyond the first record in each collision group; not a duplicate count.
Largest group66Maximum group size in this snapshot.

Exact matching preserves every visible character after trimming outer whitespace. Comparison version nfkd-mark-lower-separator-v2 applies Unicode NFKD decomposition, removes combining marks, lowercases without choosing a national locale, and folds non-letter and non-number runs into separators. It is not full Unicode case-folding, homoglyph or confusable detection, transliteration, fuzzy spelling, or semantic similarity. Original names remain the evidence; the key exists only to find records that need another join field.

Collision group sizes

Records sharing a nameExact groupsNormalized groupsReview implication
25167Pairwise ambiguity; compare route, source, operator, and capture date.
344Multi-record ambiguity; build an evidence graph before deciding whether records should merge.
415Multi-record ambiguity; build an evidence graph before deciding whether records should merge.
611Multi-record ambiguity; build an evidence graph before deciding whether records should merge.

Most collisions are pairs: 51 of the exact groups and 67 of the normalized groups. The largest normalized group contains 6 records. CSR publishes these aggregates instead of listing ambiguous names as allegations. Provider-level resolution belongs in a sourced evidence file where corrections and changing routes can be recorded.

What changed after separator folding

Sensitivity measureGroups or recordsMeaning
Cross-spelling normalized groups27Groups that contain more than one exact stored display string.
Records in cross-spelling groups642% of the inventory; these are the records added to a broader comparison queue.
Normalized groups with one exact spelling50The exact spelling already repeated; normalization did not create the ambiguity.
Unkeyable records2No letters or numbers remain, so the records are kept out of collision groups.

27 of the 77 normalized groups contain multiple exact strings, covering 64 records. That is the measurable sensitivity added by case, mark, and separator folding. It is not evidence that the spellings are interchangeable in branding, law, language, or operator identity.

Source context does not settle identity

ContextExact groupsNormalized groupsWhat it can show
One source family4456The records entered through the same kind of catalog or seed workflow.
Multiple source families1321The label appeared across different discovery methods; independence still must be checked.
One source URL4456Records may share a category page, so repetition can originate in one list.
Multiple source URLs1321More than one source route exists, but those routes may copy each other.

Among exact collision groups, 44 share one source URL and 13 span multiple source URLs. A shared category route can explain a repeated label without proving duplication. Multiple directories can also repeat the same underlying listing, so source count must not be promoted into independent corroboration without tracing provenance.

Direct-route context narrows only part of the queue

Route context within exact collision groupGroupsWhat remains unresolved
No direct route captured40Name and discovery source are insufficient to compare destinations.
One route for every record2The shared route is a strong join clue, but ownership and service continuity still need proof.
One route with missing records9Some records can be linked to a route; blank routes must not be silently imputed.
Multiple direct routes6The group may contain separate operators, tiers, migrated handles, resellers, or stale records.

40 of the 57 exact-name groups have no direct route stored. Only 2 groups have one normalized route populated across every record. The collision queue therefore cannot be solved by a bulk name merge.

Why the name page stops at a queue

This report deliberately does not choose a canonical record or publish a bulk merge list. A name key lacks route ownership, operator identity, product tier, source independence, and continuity over time. The separate entity-resolution checklist defines the record-level evidence needed to move from a collision group to a dated resolution status. Keeping that workflow separate preserves this page’s job: quantify name-key sensitivity and expose its limits.

What the collision count can and cannot support

The count supports a narrow factual claim: display-name uniqueness is weaker than slug uniqueness in this inventory, and 172 records need additional keys after conservative normalization. It cannot identify the current owner of a route, prove that one provider copied another, establish which record came first, or determine whether a similarly named service is legitimate. Those conclusions require provider-level evidence beyond this aggregate dataset.

140 records in normalized collision groups remain listed for review and 32 are quarantined. Workflow status does not resolve the collision; it only describes how CSR currently frames the record.

Dataset boundary

This is a census of records in the CryptoSignalsReview candidate inventory at snapshot 2026-07-05, not a representative survey of every crypto signal provider or every messaging channel. Directory inclusion is discovery evidence only. Missing fields stay missing, platform labels can overlap, and no count proves provider quality, legality, profitability, or safety.

Frozen source: 5dcd30b2b1a0da9bacbaeec08244190dae19a49f. Unit: one unique CSR provider slug. Position: coverage is not endorsement.

Official context sources

These sources explain why identity, disclosure, complete performance evidence, and resistance to urgency matter. They do not validate any record in the CSR dataset.

Continue the transparency research