The market intelligence layer for Pokémon investing.
Disciplined, quantitative, restrained. Often compared to a Bloomberg terminal for collectibles.
From short-cycle markets to long-duration intelligence.
I've been into Pokémon for years, more on the game and Pokémon GO side than the cards, to be honest. My background is finance, mostly building trading bots for markets like Polymarket and forex. When I started looking into Pokémon investing, I noticed the same thing I'd seen elsewhere: a real market with barely any proper data behind it. TCGinvest is my attempt to fill that gap.
Martin has spent years inside markets — first as a daytrader, then as the operator of Polymarket bots. The focus today is longer-horizon portfolio development inside Pokémon TCG: building real positions, holding cards that matter, the quiet satisfaction of owning something physical that has compounded value — while bringing the same quantitative discipline to a market that has been waiting for it.
The earlier work was concentrated where inefficiencies are densest: short-cycle crypto markets, NBA-game oracle arbitrage, and other venues that reward pattern recognition, fast iteration, and unforgiving attention to execution friction. Markets that do not tolerate models which have not been pressure-tested against real fills, real slippage, and real settlement timing.
That discipline applies to fast markets. It compounds even more in slower ones — where most participants do not bring rigorous analysis at all, and where a small structural edge held patiently can matter for years.
Pokémon TCG fits that description precisely. A multi-billion-dollar alternative-asset category, thirty years of historical price data, real cultural demand from a fanbase now reaching peak earning age — and almost no platform applying the analytical tooling investors expect from public markets. The discourse is dominated by hype channels, screenshot culture, and PSA-10 anchoring. None of which a serious investor would tolerate in any other asset class.
TCGinvest is what we wanted to use ourselves. It applies the same quantitative discipline developed in shorter-horizon work — but pointed at a longer, structurally inefficient market. The output is not trade signals or hype prediction. It is the disciplined market intelligence that lets a serious investor think about Pokémon cards the way a portfolio manager thinks about any alternative asset.
A research platform, not a price tracker.
TCGinvest provides market intelligence for the Pokémon Trading Card Game secondary market — a multi-billion-dollar alternative-asset category that has historically lacked the analytical tooling investors expect from public markets.
We track ~25,000 cards across the entire English-language Pokémon TCG catalogue, refresh pricing daily from licensed marketplace and graded-population sources, and apply a disciplined three-layer model: raw data acquisition, quantitative analytics, and interpretive market intelligence.
Data is the floor. Intelligence is the moat.
The platform is organised into three concentric layers of increasing analytical depth. Each is necessary; each alone is insufficient.
Daily-refreshed Cardmarket EUR prices, PriceCharting graded comparables, PSA populations, sales counts, listing depth. Most competitors stop here. Commoditising.
Composite scores, sub-score decomposition, five-year projection bands with explicit confidence labels, era-conditional fundamental ratings. Honest about uncertainty.
Continuous interpretation: market pulse, interpretive alerts, rotational analysis, portfolio benchmarking, opportunity decay. The reason serious investors check daily.
Five principles that shape the methodology.
Stated explicitly so readers can evaluate whether the methodology is consistent with their own framework.
Every downstream metric — score, projection, leaderboard — is computed from the cheapest available raw Near-Mint English-language Cardmarket listing. Graded slabs and non-English copies inform context but never the headline number.
Where the data is thin, we say so. Cards with sparse sales evidence carry a low-confidence flag. Projections are presented as ranges with explicit confidence bands rather than point forecasts.
We do not tell users what cards will be worth. We tell them which are structurally strongest by current fundamentals, which categories are rotating, and where confidence is high versus low. The user decides what to do with that.
The product avoids language that promises returns or substitutes hype for analysis. Same register a research note would use — institutional, not promotional.
When scoring weights change, when filters are added, when the approach to a market regime is revised — users see the history. Standard practice in quantitative research; rare in collectible tooling.
Raw markets vs graded slabs.
TCGinvest is built primarily on the raw (ungraded) Near-Mint English-language market. Raw listings reflect what a transaction would actually clear at, today, without the multi-week round-trip of grading. PSA-10 prices are not the price of a card — they are the price of one specific graded outcome, contingent on submission, grading throughput, and gem-rate yield.
Certain high-end collectibles — particularly trophy-tier Gold Stars, shadowless Charizards, Holon-era chase variants — trade primarily as graded slabs because the raw market has effectively disappeared at that price point. These cards remain visible on TCGinvest with a SLAB MARKET FROM €N label, but are deliberately excluded from raw-market investment rankings to preserve scoring integrity. The absence of a raw price is a market-structure signal, not a data gap.
The discipline behind the numbers.
A research platform is only as honest as its worst data point. TCGinvest publishes the gates and procedures that keep the numbers displayed on every card defensible, so readers can evaluate the quality of the underlying pipeline.
Freshness gates. Every regime label, mover signal, and watchlist alert is evaluated against price data verified within the last 21 days at minimum — alerts specifically require sub-24-hour freshness before firing. Stale data cannot trigger a user-facing notification. Cards whose price hasn't been verified recently are surfaced with an explicit freshness note rather than silently treated as current.
Liquidity floors. Cohort regimes and movers exclude variants with fewer than three verified raw sales in the last 180 days. A single rogue listing on an otherwise illiquid card cannot move our regime label or fire an alert. Trophy-tier cards in particular are treated as observation-only until cross-marketplace sales accumulate.
Contamination quarantine. A write-side database trigger rejects any price update that deviates more than 5× from the smoothed weekly trend. Rejected writes are logged with the source URL for inspection, not silently dropped. The smoothed trend itself is computed from Cardmarket's own 7-day weighted average, which is more conservative than any single scrape.
Trend computation transparency. The “Live NM ask” figure on each card is the Cardmarket lowest Near-Mint English-language ask. The 30-day change percentage compares that figure against Cardmarket's own 30-day average, not against a derived model. We do not smooth, infill, or interpolate price history to make charts look cleaner.
The things we deliberately refuse.
A research platform's voice is defined by what it refuses to say as much as what it says.
- ✕Predict prices or claim to know future values.
- ✕Produce buy / sell signals or trade recommendations.
- ✕Promise returns, guarantee outcomes, or use hype language.
- ✕Anchor portfolio value to PSA-10 multiples.
- ✕Provide regulated financial advice or fiduciary services.
- ✕Display sub-€4 bulk cards in investment ranking surfaces.
Honest about the edges.
The scoring model targets multi-year fundamental quality under normal market conditions. It does not protect against bubble unwinds or severe regime collapses. Individual realised returns vary widely; the tier badge and relative ranking are the load-bearing signals at the card level, not the projection point value. Trading cards are illiquid alternative assets — invest only what you can afford to lose.
Full terms in the Terms of Service and methodology detail in our research notes on the Ledger.
Questions, feedback, or partnerships?
Direct line to the operator. Response typically within a working day.