← back

tcgvibes Index Family Methodology

v0.5 draft, published 2026-05-05

tcgvibes publishes substrate-cohort indices for trading card markets. This page discloses the construction, maintenance, calculation, and governance rules so that users, reviewers, and reproducers can understand what the indices mean and what they do not. The methodology is informed by IOSCO Principles for Financial Benchmarks (P1 to P19), the MSCI Global Investable Market Indexes Methodology template, the Mei-Moses Art Index repeat-sales precedent, the Liv-ex Fine Wine Index dealer mid-price precedent, the NCREIF Property Index appraisal-smoothing acknowledgment, the Brown-Warner 1985 survivorship discipline, and the AEA Data and Code Availability Policy.

tcgvibes is the sole administrator of this index family. Indices are forward-looking benchmarks; they are not price predictions, trading signals, or investment advice.

1. Universe and cohort partitioning

The universe covers six trading card games: Pokémon, Magic: The Gathering, Yu-Gi-Oh!, Lorcana, One Piece, and Flesh and Blood. Cards are partitioned along five independent axes, forming cohort cells:

  • Game (six values)
  • Locale (en-US, ja-JP, ko-KR, multi)
  • Grade (raw, PSA9, PSA10, BGS9.5+)
  • Substrate quadrant at cohort inception (filling, over-structured, emptying, chaos)
  • Vintage class (vintage, mature secondary, modern rotation, nascent)

Each cohort cell carries a deterministic identifier. A minimum-population gate suppresses emit when sample size is insufficient; cells with sample size below thirty are annotated with a wider confidence band per Brown-Warner 1985 and Renneboog-Spaenjers 2013.

2. Index maintenance

Constituents are rebalanced quarterly. Additions occur when a new set or card meets the minimum-population gate. Deletions occur when a card is delisted or sample size falls below the survivorship threshold. Reconstitution events (set rotation, format change, mass reprint) are processed per the Corporate Actions rules below and announced in the version history.

3. Three-tier price hybrid

The cohort price uses a three-tier hybrid drawn from alt-asset publishing precedent:

  • Tier 1 repeat-sales (Mei-Moses 1996 isomorphic). Hedonic regression with repeat-sales filter; pair-match same PSA cert at distinct sale dates; weighted average over pairs within cohort cell and period. Premium and vintage and high-grade only.
  • Tier 2 dealer mid-price (Liv-ex 1988 isomorphic). Cross-source weighted average from TCGPlayer market price, JustTCG aggregated market price, and Cardmarket EU market price; outlier filter; smoothing acknowledgment per NCREIF 1978.
  • Tier 3 listing-side proxy (sentiment indicator). Active-listing distribution and listing count and watcher count from eBay Browse and Cardmarket EU. Used as real-time sentiment indicator alongside Tier 1 and Tier 2; not a return measurement on its own.

4. Per-tier maturity disclosure

Honest disclosure precedes computation. Users and reviewers must know which tier is currently active per game, locale, and grade.

  • Tier 3 is producible today across all six games using eBay Browse alone. The Year 1 first-emit consists of nine Tier 3 listing-side indices, each labeled as a sentiment indicator and not a return measurement.
  • Tier 2 is producible today for Pokémon (TCGPlayer plus JustTCG). Other games are partial; Cardmarket EU integration unblocks expansion. ja-JP and ko-KR locale Tier 2 is null until partner upgrade; capillary fiber sentiment is observed for these locales but no price is emitted.
  • Tier 1 is not yet producible at index-publication scale for any game. The bootstrap path includes PSA cert wiring, auction house data partnerships (Heritage, Fanatics Collect, Goldin), and eBay Marketplace Insights partner-tier API. Until those mature, cohort cells fall back to Tier 2 weighted average and the cell carries an honest tier-disclosure flag.

Bootstrap is forward-only. Inception date for each index is the methodology publication date; no historical reconstruction (Mei-Moses 1996 convention).

5. Calculation

Cohort cumulative-return series uses log-returns over horizons of thirty, sixty, ninety, and three hundred sixty-five days. Index value computation is base 100 at inception with per-period multiplicative growth. Confidence bands at 95 percent are emitted alongside index values; cells with sparse sample emit wider bands per Renneboog-Spaenjers 2013. Index level is published weekly with a smoothing-lag acknowledgment of one to two weeks per NCREIF 1978.

6. Survivorship and cross-source agreement

Per Brown-Warner 1985, delisted cards (no listings or sales for ninety days) are flagged in the cohort ledger with decay weights: thirty-day-stale half-weighted, ninety-day excluded from Tier 2, three hundred sixty-five-day flagged low-quality period. Each index carries a survivorship quality flag. Cross-source disagreement above twenty percent coefficient of variation widens the published confidence band per Liv-ex 500-plus dealer aggregation practice.

7. TCG corporate actions

MSCI Section 4 isomorphic with TCG-specific events: set rotation reclassifies vintage class; reprint events create new cohort cells per print variant rather than merging with originals; format discontinuation triggers cohort archive; publisher acquisition or merger and tournament format rule changes are processed per IOSCO P12 procedure.

8. Special events

Trading suspension, counterfeit and authenticity issues, grading service operational issues, locale market disruption, and force majeure are processed per IOSCO P13 transition plan. Affected cohort cells are flagged and published with reduced confidence.

9. Internal controls and governance

Per IOSCO P1 to P5: tcgvibes is the sole administrator; third-party oversight is via the audit trail (every cohort-period emit carries a cryptographic provenance entry per the eight-entity attribution schema); conflict of interest is disclosed (affiliate revenue from TCGPlayer Impact and eBay EPN; index methodology is not influenced by affiliate-source pricing per Tier 2 cross-source weighting); the control framework requires reproducible computation; internal oversight is the open-source-style audit trail of this page plus the cryptographic provenance log.

10. Data sources and license compliance

Layer 1 (publicly redistributable derived form): eBay Browse API, TCGPlayer affiliate market price, JustTCG aggregated market price (attribution: Powered by JustTCG), Cardmarket EU market price. Layer 1.5 (lab-local only; not redistributed): TCGCSV stale archive, PriceCharting derived form, 130point paid tier when applicable. Layer 2 (partner tier; future): eBay Marketplace Insights, TCGPlayer Sales API, Cardmarket commercial, Greysheet B2B, Heritage Direct, and others as partnerships mature.

Promotion from Layer 1.5 to Layer 1 requires six criteria: license inheritance permitting commercial or academic redistribution in derived form; aggregated derived form only (no raw-record redistribution); explicit attribution per emit; ToS and robots.txt and commercial-use clause review; rate-limit compliance and minimum sample size; stale-data labeling with timestamp.

11. Compliance posture

IOSCO P11 (methodology content) is the direct binding standard for this page. EU Benchmarks Regulation awareness: this index family is not a financial-market benchmark within the EU BMR scope as TCG cards are not financial instruments; tcgvibes monitors regulatory expansion. AEA replication standard is followed for reproducibility. Per IOSCO P10 plus Paper 10 strong-EMH guard, indices are ex-post structural measurements and not predictive trading signals.

12. Reproducibility

Per AEA Data and Code Availability Policy 2019 plus. Computational environment is Python with pinned dependencies; data sources are documented per section 10; cohort ledger snapshots are hashed via SHA-256; replication target is byte- equal cohort-period return ledger from a reproducer reading the public substrate plus the published Layer 1 price feed. External verifier engagement is targeted for Year 2 and beyond.

13. Version and amendment history

  • v0.1 2026-05-05 initial draft. Nine sections, IOSCO P11 baseline, MSCI five-section template, three-tier hybrid, survivorship and smoothing acknowledgment, five-axis cohort partition.
  • v0.2 2026-05-05 source-independent observable discipline (no reverse-engineerable attribution from public emit); AEA replication recovers algorithm not source-specific contributions.
  • v0.3 2026-05-05 per-tier maturity disclosure (Tier 3 active, Tier 2 partial, Tier 1 not yet producible). Year 1 first emit revised from five to nine Tier 3 indices.
  • v0.4 2026-05-05 listing-side methodology novelty acknowledgment and bootstrap-then-converge framing (Tier 3 to Tier 1 as data partnerships mature).
  • v0.5 2026-05-05 per-substrate-quadrant baseline reference disclosure: cohort-cell expected behavior is reported per substrate state (filling, over-structured, emptying, chaos) using canonical reference values from random matrix theory (Wigner-Dyson and Berry-Robnik), with explicit honest framing that these are reference priors, not empirical fits. Empirical recalibration is scheduled for v0.6 after the first quarter of forward-only accumulation.

Future amendments are scheduled per IOSCO P12 changes procedure and announced via this page plus the version log. Quarterly review thereafter.

14. Disclaimer

tcgvibes index values are forward-looking benchmarks published as ex-post structural measurements. They are not price predictions, trading signals, investment advice, recommendations to buy or sell, or financial advice of any kind. Past performance does not indicate future results. tcgvibes does not warrant the accuracy or completeness of the data and is not liable for any loss arising from reliance on this page. Affiliate revenue from TCGPlayer Impact, eBay EPN, and other partners is disclosed at the footer and on the privacy and terms pages; affiliate relationships do not influence index methodology.

For more on what tcgvibes measures, see the about page. Privacy and terms at privacy and terms.