Methodology & transparency

The formula

A booster's Expected Value (EV) is a weighted sum: for every card in the set, we multiply its probability of showing up by its market value, then add it all up. Pack EV = Σ (pull rate × card value).

A booster box's EV is the pack EV times the real number of packs in the box (usually 36). An ETB follows the same logic with its own pack count. We only value the packs — accessories and goodies count as zero.

A card with no quoted price counts as zero, so our EV is a cautious floor, never inflated.

The slot-by-slot model

A booster isn't a uniform draw: it's made of slots (commons, uncommons, the reverse slot, the rare or hit slot). Each slot carries its own rarity distribution, documented per set. Within a rarity, every card is assumed equally likely.

Reverse slots are modeled with the value of the matching commons and uncommons; premium replacements (illustration rares, ACE SPEC…) get their documented share of the odds.

Pull rates

The price API doesn't provide pull rates. They come from a versioned static dataset, one file per set, built from public quantitative sources: large community opening campaigns and official Japanese rates where they exist.

Every file cites its sources and carries a confidence level. When no reliable source exists, we invent nothing: the set shows "EV unavailable".

Prices

Each side of the site reflects its own language's market. In English, the TCGPlayer market price of English cards, in dollars. In French, Cardmarket's European market in euros — and, where a reliable listing exists, the price of the French-language print itself.

That French-card price comes from a daily reading of Cardmarket listings. Cardmarket publishes no per-language quote: a card has a single page pooling eleven languages. The figure is therefore the cheapest French near-mint listing — an asking price, not a sales average. So that a nonsense listing never shows, it is kept only when it sits within a plausibility band around real market signals; otherwise the European trend stays. Coverage grows day by day, within the API's strict quota.

When a card is quoted on only one marketplace, we convert the other at the day's exchange rate so the value stays real. If a European quote freezes for more than three weeks while the US market keeps moving, it is rebuilt from the latter rather than serving a stale price. The snapshot timestamp is shown everywhere.

Card condition

Every price shown is for raw cards in Near Mint condition. Never graded, never played.

It is the only condition consistent with EV: a card pulled from a booster is new by definition. Using PSA 10 graded prices would multiply expected value by five to twenty and have you opening products on a false promise — grading costs money, takes months, and most freshly pulled cards do not come back a 10. Conversely, starting from played-condition prices would understate what you actually pull.

One point of honesty: the Cardmarket trend aggregates sales across all conditions, whereas the TCGPlayer market price is Near Mint. The French-side EV is therefore structurally a little more conservative than the English one. The gap errs toward caution, never toward overstatement.

Price history

Since July 2026 the site has archived every card's price on both markets daily, along with each set's EV and sealed-product prices. The graphs on card and set pages draw that series and nothing else: no purchased backdata, no curve reconstructed after the fact.

Every curve stays in its language's market. On the English side, the TCGPlayer series in dollars. On the French side, the European series in euros, which will switch automatically to French-card data card by card as soon as it runs deep enough — the source line under each graph always says which of the two you are looking at.

A graph only appears from five readings spread over at least a week. Below that, two connected dots would look like a trend without being one.

The confidence index

Every EV ships with a confidence index blending three weighted components: pull-rate quality (40%), the share of the set's cards with a quoted price (35%) and price-snapshot freshness (25%).

Above 80 the data is solid; 55–80 is decent; below that, take it with a grain of salt.

The OPEN / KEEP verdict

A single comparison decides the verdict: your price paid against the opening EV. If expected value beats what you paid, it's OPEN; otherwise KEEP, with the margin shown. The sealed market price is shown for reference and never changes the verdict.

Probability of profit & limits

EV is an average. A single booster's variance is huge: it's all about the big hit. We approximate the distribution of the sum of packs with a normal law (central limit theorem): reliable for a box, indicative for a single pack.

Poké EV is a statistical estimation tool for information only, not financial advice. Prices are volatile and pull rates are community estimates. Open your boosters for the fun of it first.