Methodology

Evidence, observations, and interpretations

A fact records a source value and its effective time. An observation applies a repeatable condition without assigning investment meaning. An interpretive pattern requires an accepted hypothesis and calibration; it is kept private until its public maturity requirements pass.

Maturity and calibration

Patterns move through draft, provisional, research-backed, and production states. Production visibility also requires explicit public enablement, an approved threshold profile, applicable evidence, and required calibration. Preliminary research-backed output is shown only when separately enabled; no such output is currently enabled.

Evidence strength and attention

Evidence strength describes completeness, freshness, agreement, provenance, and independence. Attention indicates which displayed condition deserves review. Neither is a probability, a prediction, or an aggregate rating. Missing, stale, conflicted, unsupported, and inapplicable evidence remain distinct states.

Timing and historical cutoff

Each finding shows evaluation and effective times. Historical evaluations use only evidence at or before the displayed cutoff. Later evidence cannot be inserted into that historical context. Provider coverage, mappings, definitions, thresholds, and source revisions can change later.

Historical analogue registry and corpus gate

The analogue registry is version 1. The current research corpus is genuine-starter-corpus@2. Its 12 genuine starter cases remain under review: the actual public counts are 0 approved genuine cases and 0 approved pre-outcome snapshots. The gate requires 30 and 60, respectively, so public comparison remains blocked.

Configured profile versions: native_network@1, defi_protocol_token@1, governance_token@1, memecoin@1, evm_utility_token@1, solana_fungible_token@1, exchange_linked_token@1, wrapped_representation@1. Enabled policy-gated family versions (14): durable_native_network@1, durable_protocol_token@1, protocol_with_uncertain_token_value_capture@1, high_dilution_early_lifecycle@1, legitimate_project_severe_token_underperformance@1, legitimate_project_failure@1, economic_death_spiral@1, evm_contract_control_incident@1, evm_liquidity_rug_pull@1, solana_market_manipulation_rug_pull@1, security_exploit_without_confirmed_fraud@1, verified_repository_decline@1, recovery_after_security_or_economic_crisis@1, transparent_speculative_memecoin@1.

Outcome-blind case selection

Matching uses only approved structural evidence that was knowable at each snapshot cutoff. Eventual historical outcomes are physically separate from matching, cannot affect the selected order, and are attached only after that order is persisted and frozen. Only approved outcomes and approved, accessible source references can then be displayed.

Bands, differences, and counterexamples

Public results use named structural-overlap bands and minimum comparable-dimension requirements; internal ranking values and weights are never displayed. Every card emphasizes the strongest structural similarities, prominent differences, incomparable or missing evidence, both cutoffs, lifecycle, and coverage. Durable counterexamples and boundary cases appear separately instead of being blended into the primary analogue list.

Active public production patterns

• evidence.key_facts_conflicted@1 — Reports unresolved conflicts in analytical evidence without selecting a winning source.
• evidence.required_facts_stale@1 — Reports required evidence that is stale in the explicit evaluation context.
• evidence.provider_coverage_missing@1 — Reports applicable provider coverage that is unavailable without penalizing inapplicable categories.
• evidence.low_hypothesis_testability@1 — Reports when applicable approved hypotheses cannot be tested with the available evidence.
• supply.dilution_gap_observation@1 — Reports exact valuation and circulating-supply gaps without assigning investment meaning.
• supply.supply_definition_uncertain@1 — Reports missing, unclear, or materially conflicting supply definitions without a tokenomics verdict.
• market.reported_turnover_observation@1 — Reports turnover arithmetic while stating that reported volume is not executable liquidity.
• market.volume_spike_observation@1 — Reports the exact recent volume comparison without describing it as manipulation.
• market.volatility_observation@1 — Reports historical measured volatility windows without predicting future volatility.
• market.deep_drawdown_observation@1 — Reports drawdown from the reported all-time high without implying value or recovery potential.
• protocol.economic_activity_observed@1 — Reports mapped protocol activity without implying economic value capture by token holders.
• protocol.valuation_context_observation@1 — Reports non-annualized valuation context ratios without cheap or expensive labels.
• security.privileged_mint_capability@1 — Reports provider-observed mint capability and related controls without asserting future use.
• security.transfer_control_capability@1 — Reports each available transfer-control capability without collapsing them into a security score.
• security.upgradeability_dependency@1 — Reports proxy-based upgrade dependency without characterizing upgradeability as malicious.
• security.provider_honeypot_indicator@1 — Reports an explicitly attributed provider honeypot indicator without declaring confirmed fraud.
• security.coverage_unknown@1 — Reports unavailable applicable contract-security coverage instead of claiming no risks were found.
• development.verified_repository_active@1 — Reports recent activity in a verified repository without implying project or code quality.
• development.verified_repository_archived@1 — Reports that the verified repository is archived without declaring the whole project abandoned.

Limits

A condition not being observed does not establish the opposite claim. Public provider indicators are attributed observations, not independent verification. Repository activity is not project quality; protocol activity is not token value capture; reported volume is not executable liquidity; historical variation does not determine future behavior. Findings can be incomplete when evidence or provider coverage is missing.

The analogue engine uses versioned deterministic rules, not AI. It calculates no odds, forecast, investment recommendation, or verdict. A historical outcome is factual context and does not establish how a current asset will behave.