Methodology

How a card gets into the vault

Vault 10 Capital adds one PSA 10 Pokémon card to the vault each day. The selection is mechanical: a card either clears every gate below or it is not eligible, regardless of how good the story sounds. The gates are published here in full. The specific thresholds behind them are what subscribers pay for.

What the screen looks for

A real graded population

Enough PSA 10s in circulation to form an honest market

Below a certain population a handful of sales can invent a price that does not exist. We require a population large enough that the quoted value reflects a market rather than an anecdote.

A structurally low gem rate

Cards that are genuinely hard to grade clean

A low PSA 10 rate means the top-grade float stays small even as more copies are submitted. That is a supply argument, not a hype argument, and it is the single most durable edge we know of.

Genuine trading depth

Consistent recorded PSA 10 sales, not sporadic ones

Recorded sales counts come from PriceCharting's third-party feed. A card that trades a few times a year has a price on paper only. We want a card that changes hands regularly enough for its mark to mean something.

A settled grading wave

Past the release rush, before the repricing

Immediately after a set releases, submissions are still pouring in and no honest gem rate exists yet — the early numbers are noise. We wait for the excitement to settle and a stable gem rate to emerge before a modern set becomes eligible at all. Vintage and older sets are always past that point, but must still show meaningful recent activity.

A closing PSA 10 float

Population growth, not just population level

A low gem rate describes a ratio. It says nothing about whether the PSA 10 count is doubling right now. Vault 10 Capital stores a dated population reading every time PSA is called, so the float can be diffed across months rather than read as a snapshot — and a card whose top-grade supply is actively expanding is rejected however scarce it looks today. Nobody else publishes this number because it only exists if you have been recording it.

A durable subject

Characters that hold bids across cycles

Charizard, Pikachu, Umbreon, Eevee, Mew, Lugia and the other blue chips have found a bid in every downturn we can measure. Mid-tier Pokémon have not. Named trainers behave like characters too. Ordinary subjects are screened out, because a long-hold thesis on a card nobody asks for by name is not a thesis.

A scarce printing

Bespoke art over the base run

Alt arts, illustration rares, full arts and secret rares behave nothing like the regular holo out of the same set — they carry the durable collector floor. Print scarcity is scored on every card, and the base run ranks last.

Limited reprint exposure

Cards that cannot simply be re-issued

A card that can reappear in a future set or a promo tin has a ceiling on it. 1st Edition and shadowless printings are structurally unrepeatable and are treated as such; bespoke artwork is rarely re-used; everything else carries a stated reprint risk in its write-up.

A patient entry

Bought into weakness, near the card's own floor

We do not chase. The engine favours chase cards whose price history shows a steady drift down that has landed close to their own low, so the entry is made when attention is elsewhere rather than at the top of a run.

The anchor intelligence layer

Each set has anchor cards — Pikachu and the other highly visible chase cards — whose prices are monitored daily. When an anchor makes a sharp move on real sales, the engine scans the rest of that set for popular cards that have not moved proportionally and ranks the largest remaining gap first. If 1st Edition Jungle Snorlax runs, Jungle Pikachu and Eevee get examined immediately — but they are only added if they independently clear supply, gem rate, depth and maturity. The lag is a tiebreaker, never a bypass.

How a pick reaches release day

Picks are not decided on the morning they are published. Every day the full PriceCharting Pokémon guide is re-synced and one price mark is recorded for each card in the universe, which builds the price history the screen reads. Once a week a deep pass takes the eligible sets, examines the top chase cards in each of them, and keeps only the ones whose recent history shows a steady decline that has landed near the card’s own low.

Those survivors are checked against cached PSA population and gem rate, and the best of them are written into a multi-week queue with the rationale, filters and risk notes already computed. PSA population is fetched as a slow daily drip against the shortlist rather than on demand, because population barely moves week to week — so a card is only ever looked up once, and no pick waits on a grading-data call.

On release morning the queued card gets one fresh valuation so the vault books a live entry price rather than a stale one.

Concentration limits

A book that never sells needs limits on what it accumulates, or a year of “modern alt arts look cheap” quietly turns the whole vault into a single bet. Vault 10 Capital caps how many positions can share a set and how many can share a character, counted across everything already held plus everything already queued. When a card clears every gate but its set or its character is full, it is passed over and the next candidate takes the slot. A better card is sometimes skipped for a merely good one, and that is the intended trade.

This is a long-term portfolio

Vault 10 Capital does not trade. There is no exit signal, no target price and no sell discipline, because there is nothing to sell — every position is held indefinitely and marked to market. The thesis on every card is a multi-year one: a small PSA 10 float meeting persistent collector demand. Judge the vault on its cumulative curve over years, not on any single position or any single month.

One consequence worth stating plainly: positions are allowed to go against us and stay there. A card that is down is not a failed trade to be cut — it is a holding whose thesis has not played out yet, and it stays in the vault and in the reported numbers either way. Nothing is ever quietly removed.

How cards are valued

The entry price booked into the vault is the median of the card’s verified PSA 10 sold prices over the three weeks ending on its entry date, taken from PriceCharting’s completed-sales record for the exact card and version. A median rather than the single last sale, because one mis-scoped listing or one bundle can sit 40% away from the market and we are not willing to book that as a cost basis. When a card is thin enough that no sale cleared in that window, the last confirmed sale is used and the position is dated accordingly.

The daily revaluation uses PriceCharting’s PSA 10 figure for the same card. It is a modelled value that smooths across recent sales — not a last-sale price — which means it deliberately resists single-sale noise and can lag a fast move in either direction. We label it as a valuation rather than a sale everywhere it appears, along with the date it was read.

Alongside it we publish a live eBay ask floor: the median of the cheapest few qualifying PSA 10 listings, filtered to fixed-price US listings with clean single-slab titles and established sellers. Because a PSA 10 of a given card is close to fungible, that floor is a fair read on what you would pay right now. It is still an asking price, not a sale, so it never sets the vault's value — it exists as an independent cross-check. When the two disagree materially we flag it, because that gap usually means the market has moved and the smoothed valuation has not caught up yet.

Population and gem-rate figures come from PSA's public population API. When any feed is unavailable, the vault carries the last recorded figure forward unchanged and says so. No number here is ever simulated or estimated.

What we publish and what we don't

Everything above is the complete list of things the screen cares about — there is no hidden extra factor. What stays behind the subscription is the calibration: the exact population floor, the gem-rate ceiling, the volume minimums, the maturity window, the float-growth ceiling, the character and print scoring weights, the trend and floor maths, the anchor and lag thresholds, the concentration caps, and the forward queue of cards already selected. Every entry, mark and result is published in full, including the losers.

Timing

The daily job runs before the public tracker refreshes. Subscribers receive the pick and the full analysis immediately; the public portfolio and the public Today's Pick page release the same card the following morning with a shorter summary. This delay is the entire value of the subscription, which is why the sharing policy below matters.

Subscriber sharing policy

This daily pick and full analysis are for paid subscribers only. Please do not share the card recommendation or detailed rationale publicly on social media, Discord, or forums until the public portfolio updates. Limited sharing hurts the project for everyone. Thank you for helping keep Vault 10 Capital sustainable.

Required disclaimers

  • §This is not financial advice. Nothing on Vault 10 Capital is a recommendation to buy, sell, or hold any asset.
  • §Past performance of the hypothetical portfolio does not guarantee future results.
  • §Pokémon card values are highly speculative and can decline significantly, including to a fraction of the entry price.
  • §The portfolio is hypothetical. Vault 10 Capital does not necessarily purchase the cards it recommends, and reported entry prices are the median of recent confirmed public sales, not a fill we received.
  • §Affiliate disclosure: Vault 10 Capital may earn commissions from eBay Partner Network and Amazon Associates links at no extra cost to you.
  • §Users should do their own research and consider consulting a licensed professional before making any purchase.