Rare Index
Methodology
How Rare Index measures the collectibles market — and why institutional investors can trust it.
Our Approach
Rare Index provides institutional-grade market intelligence for the collectibles industry, treating trading cards, luxury watches, toys, video games, and other alternative asset categories as benchmarkable markets.
Our indices track price movements, relative performance, and market signals using a rigorous, transparent methodology built on the same principles that underpin traditional financial benchmarks.
We employ the Laspeyres Price Index method — a fixed-weight framework used by major economic indices worldwide — to measure changes in asset values over time. This approach prioritizes consistency, comparability, and resistance to manipulation over reactive trend-chasing.
This page outlines the key components of our methodology. We publish it because transparency is not optional when institutional capital is at stake.
What We Measure
Rare Index quantifies the performance of curated asset universes within the collectibles market. Each category — Trading Cards, Luxury Watches, Toys, Video Games, and Other — is represented by a defined basket of assets selected for market relevance and data reliability.
We measure
- Price appreciation and depreciation — Changes in verified asset values relative to a fixed base period.
- Relative performance — Direct comparisons against traditional benchmarks including the S&P 500, NASDAQ, Bitcoin, and Gold.
- Volatility — Standard deviation of price changes to quantify risk exposure.
- Signal momentum — Short-term trend and momentum indicators derived from pricing data.
- Cross-category capital rotation — Flows of value between collectible categories, highlighting where capital is moving.
We do not measure
- Private or unreported transactions.
- Subjective condition adjustments outside of verified third-party grading.
- Hypothetical, modeled, or estimated pricing.
- Illiquid one-off sales that lack sufficient data verification.
This scope ensures our indices reflect observable, verifiable market dynamics without introducing speculation.
Index Construction
Rare Index constructs proprietary market indices using a Laspeyres Price Index methodology — a time-tested approach that fixes asset weights at a base period and measures pure price change over time.
What that means in practice: We lock in the relative importance of each asset at the base period. From that point forward, we measure only how prices change — not how popularity, trading volume, or market attention shifts. This isolates the price signal from the noise.
Proprietary market index levels are calculated daily from aggregated market signals derived from licensed datasets, public transaction records, and approved partner inputs, where permitted by source licensing.
The Formula
Variable Definitions
| Variable | Definition |
|---|---|
| Pt | Price of the asset at time t |
| P0 | Price of the asset at the base period |
| Q0 | Weight assigned at the base period |
| RIt | Rare Index value at time t |
All indices are normalized to a base value of 100 for intuitive interpretation. A reading of 112 means the basket has appreciated 12% from the base period. A reading of 94 means it has declined 6%.
Why Laspeyres?
The Laspeyres method is the standard framework for financial benchmarking. We chose it for four specific reasons:
It mirrors how institutional indices operate.
The methodology parallels the construction principles behind major equity and economic benchmarks, making Rare Index immediately legible to professional investors and analysts.
It prevents distortion from shifting weights.
Fixed base-period weights ensure that a suddenly trending asset cannot artificially inflate the index. The basket measures market health, not market hype.
It enables consistent long-term benchmarking.
Because the weighting structure remains stable, index values from different time periods are directly comparable without retroactive adjustment.
It mitigates survivorship bias.
By defining the universe at the base period, the index avoids the trap of favoring only assets that performed well after the fact.
For collectibles specifically — where hype cycles, social media attention, and rarity narratives can dramatically skew short-term perception — Laspeyres provides the stability required for credible market measurement. The index reflects what the market is doing, not what it is talking about.
Universe Selection Criteria
Each Rare Index universe is curated to represent the core of its category while maintaining strict data integrity. An asset must meet defined thresholds to be included.
Category coverage depth is determined by verified transaction density and licensing permissions. Assets lacking sufficient observable market depth are excluded from index construction.
Inclusion criteria
- Verified transaction history — A minimum number of observable, arms-length sales in the base period to establish a reliable price baseline.
- Minimum liquidity threshold — Sufficient trading activity to ensure pricing reflects genuine market consensus rather than isolated transactions.
- Sales frequency — Ongoing transaction activity to maintain price reliability over time.
- Brand and category relevance — Assets from recognized issuers, manufacturers, or grading standards with established cultural or economic significance.
- Data reliability — Pricing sourced from verified platforms with cross-validation across multiple data partners and market inputs.
- Outlier exclusion — Removal of statistical anomalies (transactions exceeding three standard deviations from the mean) to prevent individual sales from distorting the index.
Universes are category-specific. The RI Cards Index, RI Watches Index, and RI Toys Index each maintain independent selection criteria calibrated to their market's characteristics. Video Games is currently tracked through brand-level coverage while the dedicated RI Video Games index is deferred to a later phase.
Universe composition is reviewed semi-annually. However, changes are only implemented during formal reconstitution events to preserve index continuity and prevent arbitrary mid-cycle adjustments.
Confidence Scoring Framework
Confidence Score Methodology
Each tracked asset is assigned a Confidence Score (0–100) reflecting the reliability and robustness of its observable market pricing.
The Confidence Score evaluates data quality and liquidity integrity — not investment merit.
Scores are calculated daily in batch following the completion of the pricing ingestion pipeline.
Composite Model Structure
The Confidence Score is a weighted composite of four measurable dimensions:
| Component | Weight | What It Measures |
|---|---|---|
| Transaction Density | 40% | Number of verified historical price observations within a rolling 90-day window. |
| Recency | 30% | Days since the most recent qualifying price observation. |
| Source Quality | 20% | Tiered reliability ranking of contributing data feeds. |
| Price Consistency | 10% | Standard deviation to mean ratio of price observations over a 90-day window. |
Scroll horizontally to view all confidence model columns.
Scores are computed on a normalized 0.00–1.00 scale and displayed as 0–100.
Component Definitions
Transaction Density (40%)
Higher counts of verified historical price observations within the past 90 days increase score weighting. Sparse transaction activity reduces confidence proportionally.
Recency (30%)
More recent verified observations increase score weighting. As time since last observable transaction increases, score contribution declines.
Source Quality (20%)
Pricing feeds are internally tiered based on historical reliability and data integrity. Higher-tier sources contribute proportionally stronger weighting to the composite score.
Price Consistency (10%)
Measured as the ratio of standard deviation to mean price over a 90-day window. Lower dispersion increases score weighting. Higher volatility reduces score contribution.
Interpretation Guide
| Score Range | Interpretation |
|---|---|
| 90–100 | Deep liquidity, recent activity, strong source integrity. |
| 75–89 | Moderate liquidity with stable pricing. |
| 60–74 | Limited transaction depth or elevated variance. |
| Below 60 | Sparse activity or inconsistent pricing. |
Confidence Scores update daily following data ingestion. They do not predict future performance and should not be interpreted as risk ratings, credit assessments, or investment recommendations.
Data Governance
Rare Index applies a formal source-separation policy across ingestion and publication workflows. Auction comparable data, including WatchAnalytics hammer records, is institutionally segregated from market pricing data and is never promoted as a current market price. Each incoming feed is classified by source type and processed through isolated pipeline stages so comp intelligence and pricing intelligence remain distinct by design.
Every price record written to the Rare Index database carries mandatory source attribution as a non-optional provenance field. Records that cannot be tied to a verified source are excluded from all public-facing outputs, including index calculations, asset views, and downstream analytics. This control has been enforced continuously since platform launch across all 17 active data sources.
Each item is assigned a machine-computed confidence score on a normalized 0.00-1.00 scale, derived from recency, source breadth, comparative data depth, and price consistency. Scores are recalculated daily after pipeline completion. Items that fall below institutional confidence thresholds are explicitly flagged rather than removed, preserving transparency on data quality while avoiding false completeness.
Rare Index also maintains an institutional data floor defined as a verified minimum count of matched auction records. That floor is monotonic across releases: any pipeline or logic change that would reduce matched-record coverage below the floor is automatically rejected in release controls. The result is a governance constraint that requires data density to improve over time rather than regress.
Price Selection Protocol
Canonical Price Selection
For each asset and snapshot date, Rare Index publishes a single canonical price observation.
When multiple qualifying price observations exist for the same asset and date, the following deterministic selection protocol applies:
- 01The most recent verified timestamp is selected.
- 02If timestamps are equivalent, higher-tier data sources take precedence.
- 03If no timestamp differentiation exists, the highest verified qualifying price is selected.
This protocol ensures one consistent daily market observation per asset. Outlier filtering is applied in specific categories where transaction dispersion exceeds defined statistical thresholds.
Source Deviation
The Deviation metric displayed in asset detail views measures the percentage difference between an individual data source's reported price and the composite current price for that asset.
Formula
((Source Reported Price − Composite Current Price) / Composite Current Price) × 100
This metric enables users to evaluate pricing agreement across data sources. Lower absolute deviation indicates stronger source convergence. Higher deviation may indicate stale data, different market conditions, or pricing methodology differences between sources.
Visual thresholds
- < 1% absolute deviation — Strong convergence
- 1–3% absolute deviation — Moderate variance
- > 3% absolute deviation — Significant divergence
Source Deviation is distinct from the 30-Day Change metric, which measures an asset's price movement relative to its own historical price 30 days prior.
30-Day Change
The 30-Day Change metric reflects the percentage price movement of an individual asset relative to its own observed price 30 days prior.
Formula
((Latest Price − Price 30 Days Ago) / Price 30 Days Ago) × 100
This metric measures asset-level short-term price momentum. It is calculated using the closest verified historical price observation within a ±5 day window of the 30-day lookback date.
30-Day Change is gated by a minimum confidence threshold to prevent unreliable calculations on assets with sparse pricing data.
Weighting Method
Weights are assigned at the base period and remain fixed in accordance with the Laspeyres methodology. Rare Index uses a price-weighted approach, where each asset's influence on the index is determined by its verified market price at the base period.
Weighting approaches by category maturity
- Price weighting — Default methodology for established categories where reliable base-period pricing is available across the universe.
- Equal weighting — Applied in emerging categories during early data phases where transaction history is still being established. Equal weighting ensures no single asset dominates the index before sufficient market data exists to justify differentiated weights.
Per-category weighting methodology is documented transparently. As categories mature and data depth increases, weighting methods may be upgraded during formal reconstitution — with advance notice and full documentation of changes.
Weights are normalized to sum to 100% within each index.
Rebalancing & Reconstitution
Rare Index balances long-term stability with market relevance through controlled, scheduled updates.
Rebalancing refers to restoring asset weights to their base-period levels when price drift causes significant deviation from the original structure. Under normal market conditions, rebalancing is not required due to the fixed-weight structure of the Laspeyres methodology. Rebalancing is only triggered under defined structural distortion thresholds.
Reconstitution refers to the formal process of adding or removing assets from the universe based on evolving inclusion criteria — for example, incorporating newly liquid assets or removing those that have fallen below minimum trading thresholds.
Schedule
- Semi-annual review — Internal assessment of universe health, data quality, and weighting drift. No changes are implemented.
- Annual reconstitution — Formal updates to universe composition, announced 30 days in advance. Changes take effect at the start of the calendar year.
This cadence ensures indices remain representative of their markets without introducing the instability that comes from frequent or arbitrary adjustments.
Data Sources
Rare Index maintains a diversified data architecture designed to ensure no single source dependency. Market inputs are classified by usage rights and integrated according to their specific licensing frameworks.
Rare Index integrates licensed APIs, verified marketplace transaction feeds, and formal data partnerships. Third-party benchmarks are displayed for contextual comparison only and do not influence proprietary index calculations unless explicitly licensed for such use.
- Primary market signals — Licensed APIs, marketplace transaction feeds, verified auction records, and category-specific data partnerships that directly power proprietary market index construction.
- Reference benchmarks — Third-party indices and analytics displayed alongside proprietary Rare Index indices for contextual comparison, with full attribution.
- Validation inputs — Official manufacturer and grading reference data, plus additional internal-use data sources applied to confidence scoring, cross-validation, and quality assurance workflows and not necessarily displayed publicly.
All pricing data carries full source provenance — every price in our system is traceable to its originating platform and timestamp. Data is cross-validated across multiple classes of market inputs where available to ensure accuracy.
Rare Index does not engage in unauthorized scraping or violate platform terms of service. All data integrations are established through official APIs, licensed feeds, or formal partnership agreements.
For a complete list of our current data partners, visit our Data Sources page.
Data Pipeline & Update Cadence
| Process | Timing (UTC) |
|---|---|
| Daily price snapshot ingestion | ~06:00 UTC |
| Full pricing pipeline completion (all sources) | ~09:00 UTC |
| Confidence Score recalculation | Immediately following daily sync |
| Index level recalculation | Daily, after canonical pricing is finalized |
All market data is considered operationally complete for the current trading day after 09:00 UTC. No retroactive price restatements occur unless explicitly documented and disclosed.
Limitations & Disclosures
No methodology is without constraints. We disclose ours because transparency builds trust — and because informed users make better decisions.
Laspeyres-specific limitations
- Potential overstatement in rising markets — Fixed weights may not fully account for substitution effects when buyers shift toward lower-priced alternatives during bull markets.
- No adjustment for quality changes — The methodology assumes base-period asset relevance persists. Significant shifts in grading standards or authentication practices are addressed through reconstitution, not mid-cycle adjustment.
- Substitution effects not reflected — The index does not dynamically account for changes in buyer preferences between assets within the same category.
Collectibles-specific limitations
- Illiquidity risk — Sparse trading in niche items can produce volatile price signals. Minimum liquidity thresholds mitigate but do not eliminate this risk.
- Condition variability — Pricing assumes standardized grading from recognized third-party services. Ungraded or inconsistently graded items are excluded where possible.
- Authentication risk — Rare Index relies on third-party authentication and grading. Fraud and misrepresentation risks persist in the broader collectibles market and are outside our control.
- Data gaps — Emerging categories may have incomplete transaction histories. These gaps are addressed through phased weighting and conservative inclusion criteria.
Important: Rare Index is a market intelligence tool. It is not investment advice and should not be used as the sole basis for any financial decision. Users should conduct independent due diligence and consult qualified professionals before making investment decisions.
How It Works
Define the Universe
Select a representative basket of assets for each category based on strict inclusion criteria.
Assign Fixed Weights
Weight each asset at the base period. Weights remain locked to isolate pure price movement.
Track Daily Prices
Aggregate verified market signals derived from multiple classes of market inputs with full provenance tracking.
Calculate the Index
Apply the Laspeyres formula to derive the daily index level, normalized to a base of 100.
Benchmark and Compare
Plot index performance against traditional benchmarks — S&P 500, NASDAQ, Bitcoin, Gold — to contextualize collectibles within the broader investment landscape.
Methodology Roadmap
Rare Index is committed to continuous refinement. As data depth increases and market coverage expands, we anticipate the following enhancements:
- Hedonic regression adjustments — Modeling the impact of condition, grading tier, and provenance on pricing to isolate true market movement from quality variation.
- Liquidity-adjusted pricing — Incorporating bid-ask spread proxies and volume metrics to weight prices by market confidence.
- Complementary index frameworks — Introducing Paasche and Fisher Ideal index variants to provide alternative perspectives alongside our Laspeyres baseline.
- Expanded reconstitution governance — Formalizing an independent methodology committee as the platform scales.
All methodology changes will be documented, announced in advance, and published with full rationale.
Index & Category Governance
Rare Index does not add categories lightly. Every new asset class goes through a formal governance process before inclusion on the platform.
How We Add New Asset Categories
Our most recent addition, Video Games (February 2026), followed the framework below.
- Market Depth — Does the category have sufficient trading activity and price discovery to produce meaningful signals? We require a minimum threshold of actively priced items across multiple source providers before a category is considered.
- Source Availability — Can we obtain daily pricing from at least two independent data providers? Single-source categories carry unacceptable concentration risk for institutional-grade analytics.
- Taxonomic Clarity — Can items be cleanly classified without ambiguity? Our classification system must distinguish the new category from existing categories with minimal overlap. Items that span categories are classified by primary market identity.
- Historical Data — Is there sufficient pricing history to compute meaningful risk metrics (volatility, drawdown, Sharpe ratio)? Categories without at least 90 days of backfill data enter the platform with limited analytics until history accumulates.
- Institutional Relevance — Does the category represent a meaningful segment of the collectibles market that wealth managers, family offices, or alternative asset allocators would expect to see tracked?
Video Games met all five criteria with over 100,000 trackable items across two independent pricing sources, clear taxonomic boundaries separating video game items from trading cards and toys, and deep historical pricing data spanning years of daily observations. The category was reclassified from a general holding pool and validated through a multi-step migration process that preserved full pricing history and data provenance.
Current categories: Cards, Video Games, Watches, Toys, Other.
Source Evaluation & Monitoring
Every data source on Rare Index is evaluated continuously across coverage, update quality, and provenance integrity.
Onboarding standards
- Coverage — How many items does the source price, and how does that compare to the addressable universe for its category?
- Update Frequency — Does the source provide daily price updates? Sources that update weekly or less frequently are flagged and disclosed.
- Data Format — Can the source be integrated via API, structured feed, or reliable automated collection? Manual or inconsistent delivery formats are not accepted.
- Independence — Does the source represent an independent price signal, or does it aggregate or replicate another source already on the platform?
Ongoing monitoring
- Freshness — What percentage of a source's items show actual price changes day over day? A source that writes records but delivers identical prices is flagged as stale, regardless of record volume.
- Completeness — Are all expected items present in each daily snapshot?
- Consistency — Are prices within expected statistical bounds? Anomalous price spikes or drops trigger review.
- Provenance — Every price in our system traces back to its original source. We never overwrite source attribution, and our audit procedures verify provenance integrity daily.
When a source fails freshness or quality thresholds, we contact the provider directly. Sources that remain stale for extended periods are disclosed through our confidence scoring system, which reduces the confidence rating for items that rely on outdated pricing.
Current source count: 17 independent data providers across all categories.
Confidence Scoring
Every item receives a Confidence Score (0-100) that measures pricing-data reliability. The score is recalculated daily.
| Factor | Weight | What It Measures |
|---|---|---|
| Transaction Density | 40% | How many recent price observations exist. |
| Recency | 30% | How recently the item was priced. |
| Source Tier | 20% | The reliability classification of the data provider. |
| Consistency | 10% | How stable the pricing signal is over time. |
Scroll horizontally to view all confidence factor columns.
Confidence reflects data quality, not investment quality. A high score indicates stronger pricing reliability; it does not indicate a favorable investment outcome.
Data Quality Standards
Daily audit protocol
- All sources are actively writing to pricing tables.
- No records exist without source attribution.
- Source names conform to standardized naming conventions.
- No duplicate records exist for single-source items.
- Price freshness thresholds are met for monitored sources.
- Historical pricing maintains full provenance back to origin.
The 0.0 != Null principle
We distinguish between a price of zero and the absence of a price. Zero is a claim; null is an explicit indication of insufficient data. We do not substitute zeros where data is missing.
Source provenance
Every displayed price is traceable to its originating source and recorded date. We do not blend sources without disclosure, and provenance is verified through each stage of ingestion, daily snapshots, and historical archives.
Index Construction & Rebalancing Governance
RI Global (RI_GLOBAL) and category indices (RI_CARDS, RI_WATCHES, RI_TOYS) are calculated daily under rules-based controls:
- Universe — All items in the relevant category with a confidence score meeting the inclusion threshold.
- Weighting — Price-weighted with concentration caps to prevent single-item dominance. For emerging categories, equal weighting may be used until depth and history meet production thresholds.
- Rebalancing — Daily, automated, and rules-based with no discretionary overrides.
- Additions — New items enter indices automatically when confidence and history thresholds are met.
- Removals — Items are removed when they fall below confidence thresholds for a sustained period.
When new categories are added, eligible items enter RI_GLOBAL immediately. Category-specific indices are launched only after depth and history are sufficient to support standalone tracking.
Methodology Updates & Disclosure
This methodology page is a living document. Material changes to scoring, source evaluation, or index construction are published with an update date. Historical methodology versions are maintained internally for audit purposes.
Last updated: March 2026 — Added governance framework, category evaluation criteria, source monitoring disclosure, and Data Governance language. Added Video Games as the fifth tracked asset category.
Questions about our methodology? Contact us.