How the odds, prices, draws and buyback work, with every number behind them.
Our entry tier, and its own published table — odds v8. Every other tier publishes its own; they are listed under Live published odds below.
The bottom figure is each band's share of the average pull, rounded to the cent — they are the parts the average is made of. The average pull on this tier is $27.50 against a $25 pack, and 66% of pulls land below what the pack cost. Selling every pull straight back would return 92.4% of the price.
Everything below is the long version, one topic per section. Open any of them.
Every card is a real graded slab held by us. Its value is our recorded market value for that card in that grade, taken from recent sales, and a market watch rechecks it regularly. Every change of value is recorded, and the value that counts for a pull is the one recorded at the moment it is drawn.
Every pack is built from a pool of real graded cards, each carrying an indexed dollar value. Cards are sorted into up to five value bands purely by value relative to the pack price — the same structure at every price point. A formula (not a person) solves the bucket probabilities so that the average pull returns a fixed, published percentage of the pack price. That percentage, and what selling every pull back returns, are listed below for every pack. They move with the pool. Each pack page shows them live.
Inside a pack, every card's own chance comes from one formula: within its value band, it is proportional to 1 ÷ (value − c), where c is a single dollar figure per pack, below its cheapest card. A dearer card is therefore always rarer than a cheaper one in the same band, and two cards of the same value have the same chance. The band totals are the published ones. When the table is solved they are exactly the sum of the curve over the whole pack; as cards are drawn and replaced between solves, each band keeps its published total and its cards share it by the same formula. c is set each time the table is solved, so the pack returns its published percentage — Standard c = $9.31, Premium c = $26.19, Elite c = $55.61. A pack's featured card, where it has one, is only a label: its chance comes from the same formula as every other card's. Every draw's proof page recomputes the card's chance from c. The odds you see and the odds the draw uses come from the same published table — it is architecturally impossible for them to differ. When the pool changes enough to move the real odds, a new version is published with a timestamp; history is kept forever.
| Pack | Average pull | Sell-back return | Pulls below price |
|---|---|---|---|
| Standard $25 | 110.0% | 92.4% | 66% |
Every pack page carries a "Full disclosure" panel showing the complete picture of what a pull is likely to be worth, before you spend anything. Every figure in it is computed live from the published odds table — the same one the draw runs on. In plain language:
We publish the full outcome distribution so you can judge a pack before you buy.
The card list for Standard changed against the one published before it. Every card below is a change to the list itself, not a change of price.
Added
One structure: every price point uses the same value-bucket template and the same published-odds machinery — only the pool gets deeper and the ceilings higher. Where a tier's pool cannot honestly fill its top bucket from real prices, that bucket is merged downward and the published odds say so — values are never invented to fill it.
Cryptographically provable (you can check the math yourself, on every single draw, from the live feed):
Not provable — this part rests on our word: that the physical slabs exist, match their listings, and sit in our vault. Cryptography cannot prove a fact about the physical world. The plan for that part is intake photos and video, inventory audits published where anyone can read them, and independent third-party checks once volume justifies the cost — none of which is in place yet. Until it is, this is the part you take on our word.
Every pull carries an instant buyback offer at the tier's rate — 84% for 7 days on Standard, 86% for 7 days on Premium and 89% for 7 days on Elite — counted from the moment of the draw, under the terms in force when the pack was opened. Elsewhere, as of July 2026, the common buyback deadline is 72 hours or less, and one platform gives full value for only two minutes. The price is locked to the index value recorded at the instant you pulled — it never gets re-quoted against you, even if the market moves. Accept it and the credit lands in your balance immediately; skip it and the slab stays in your vault.
You sign in through Privy, our sign-in provider, with your email, your phone, Apple or Google. Privy holds that login; we keep the Privy account id and the display name you choose. Before you can add money or buy, you complete an identity check with Stripe Identity — a photo of your ID and a live selfie. Stripe keeps the images; we keep the result, the name on the ID and the date of birth it shows. Card and crypto payments are taken by Coinflow, our payment provider: we never see or store your card number, and we keep a record of each payment — its amount, method and status. If you ship a card to yourself, we keep the address you give us for it.
Two other things are stored, and only if you hand them over. If you complete the age check on your account page we keep the date of birth you enter, when you entered it, and whether it was typed or documented — that last part matters because a typed date and a verified one are not the same evidence. If a jurisdiction check runs we keep the region code it resolved, how it was resolved, and when — and the earlier codes too, instead of writing over them, so that where you said you were when you bought something is still answerable afterwards. With real money the age check and the region check are required before you can add money or buy.
We only store the address after you prove you control it. Asking to add one sends a single-use link that expires in an hour; until you click it, the address is held against that unclicked link and never attached to your account. The link itself is not stored — we keep only a one-way hash of it, so nobody reading our database can reconstruct a working link.
Confirming your address is not agreeing to hear from us. They are two separate actions, deliberately. Consent to be contacted is a distinct, unticked choice on your account page, it is off unless you turn it on, and clicking a confirmation link never turns it on as a side effect. You can withdraw it at any time on the same page, and we stop. We keep the record that you once gave it — that record is the only proof that the consent we relied on was real — but withdrawal is what governs whether anything is sent.
Signing out ends the session on our server, not just in your browser: the credential is revoked, so a copy of it is worthless afterwards. Confirming an address also signs out every other session on the account. Sign-in links and expired sessions are deleted on a schedule rather than kept forever, including addresses from links that were never clicked — a mistyped address belonging to someone who never signed up here does not live in our database indefinitely.
Four controls, all enforced on our servers at both the deposit and the pack-purchase boundary — not in the page's buttons, so bypassing the interface changes nothing. Every parameter below is read from the same configuration the enforcement runs on.
Your cards and your open buyback windows are unaffected by any of these — a block stops new spending, it never touches what you already hold. Set yours or see your spending.
| Premium $50 | 111.0% | 95.5% | 66% |
| Elite $100 | 106.0% | 94.4% | 71% |
Each pack page prints the same three figures, computed live from the odds.
Display names are public here (they appear on the live feed and on collection pages), and a name is never a way in: you sign in through Privy or with an emailed link.