A booster pack is not a random handful of cards. It is a fixed number of slots, each filled from its own pool, and the shape of those slots has changed at least six times since 1999. Every odds figure on this site is computed against the specific shape of the pack you are opening, so this page is the list of those shapes and where each one comes from.
If you want to know what a pack is worth, the tempting method is to take every card in the set, multiply each price by that card's pull rate, and add it up. That overcounts badly, because it treats a pack as though every card in the set had an independent chance of showing up. A real pack has a fixed number of positions, and most of them can only ever be filled by a common.
So the count matters, and so does the split. A nine-card EX-era pack with five commons is a different proposition from an eleven-card Base Set pack with seven, even before you look at a single price. Getting the shape wrong does not shift a figure slightly — it shifts every figure for that set at once.
This is also why we store pack shape per set rather than hardcoding one modern layout. The sets where the default is wrong are not footnotes; they include some of the most-opened products in the game's history.
One real consequence: Legendary Collection was computing live daily expected value against the wrong pack shape until we checked it against Bulbapedia and found it had a reverse holo slot the era default did not. The fix changed every number on that set's page.
Scarlet & Violet and Mega Evolution packs hold ten cards: four commons, three uncommons, two reverse-holo slots and one rare-or-better. The second reverse slot is the recent change — it replaced a common, which is why a modern pack has four commons where a Sword & Shield pack had five.
Sword & Shield packs are otherwise the same ten cards: five commons, three uncommons, one reverse holo, one rare or better. If you opened packs through 2020 and 2021 and the modern ones feel slightly better, that is not nostalgia in reverse — there genuinely is one more foil in there now.
Sources: PokeBeach and TCGplayer's reporting on the Scarlet & Violet pack-composition change, corroborated by ScreenRant and poke.rip.
Each row is the shape our model actually runs on for that era. "Rare+" is the slot that resolves into a rare or anything above it; the chase cards all come out of that one position.
| Era | Years | Cards | Common | Uncommon | Reverse | Rare+ |
|---|---|---|---|---|---|---|
| Base Set – Neo Destiny | 1999–2001 | 11 | 7 | 3 | — | 1 |
| Legendary Collection | 2002 | 11 | 6 | 3 | 1 | 1 |
| e-Card & EX series | 2002–2007 | 9 | 5 | 2 | 1 | 1 |
| Diamond & Pearl – Sun & Moon | 2007–2019 | 10 | 5 | 3 | 1 | 1 |
| Sword & Shield | 2020–2022 | 10 | 5 | 3 | 1 | 1 |
| Scarlet & Violet onward | 2023– | 10 | 4 | 3 | 2 | 1 |
| Japanese SV / Mega | current | 5 | 3 | 1 | — | 1 |
Three things in that table are worth saying out loud. The reverse holo has a birthday: Legendary Collection in 2002, a year before the EX era, and it is the single exception to the vintage shape. Bulbapedia states it directly — "Each booster pack offers 11 cards, including one reverse foil card, one rare single, three uncommon cards, and six commons."
The EX-era pack is the shortest the game has had at nine cards, and it is the only era that drops to two uncommons. And the modern second reverse slot was paid for with a common — packs did not get bigger in 2023, the mix got better.
Sources: Bulbapedia's era pages for Diamond & Pearl and Legendary Collection; PokeBeach's January 2017 article on the Sun & Moon composition change; Neo Genesis product listings for the vintage shape, cross-checked against the Common-tagged energy rows in our own synced catalogue.
Celebrations (2021) — four cards, every one of them a holo. No commons, no uncommons, no reverse slot, because there is nothing ordinary in the pack to reverse. Four slots all drawing from a twenty-five card set is a completely different probability problem from a normal booster, and any model that assumes a standard shape will be wrong about it by a wide margin.
30th Celebration (2026) — five cards, all foil, with one slot guaranteed to be one of thirty different Pikachu cards, plus a foil basic energy that sits outside the five. Again: no uncommons, no reverse, because the whole pack is foil.
Mega Dream ex (Japan) — a "High Class Pack": ten cards per pack, but only ten packs per box instead of thirty. There is no separate uncommon or plain-rare tier at all; the floor is commons only.
These are not rounding errors. A set whose pack is four all-holo cards has an expected value several times a normal pack's, and presenting it next to standard sets without accounting for the shape would make the comparison meaningless.
A Japanese Scarlet & Violet or Mega-block pack holds five cards, not ten: three commons, one uncommon and one hit slot. There is no reverse holo — it is an English-only mechanic.
This matters more than it sounds. It means a Japanese pack and an English pack of nominally the same set are not comparable units, and neither are their prices. Half the cards, a different foil structure, and a different box size. Any per-pack figure that treats them as the same thing is comparing two different products.
Source: samuraiswordtokyo.com's pull-rate breakdown, corroborated by TCGplayer's own seller blog on the differences between Japanese and English foil treatments.
The same caution applies across games, not just regions. One Piece packs hold six cards in Japan and twelve in English — we checked Bandai's own product pages for both and they state different contents for what looks like the same set.
Pack shape and pull rates are different questions, and we answer them to different standards. A shape can usually be confirmed from a product listing or a publisher's own page. A pull rate needs a stated sample size from a named source, which a great many sets simply do not have.
So there are sets on this site where we know exactly what the pack contains and still show TBD for the odds. That is not an oversight. Inferring a set's rates from a similar set would produce a number, and the number would be invented.
If you want the full reasoning for the odds side, how we calculate every number covers the model, and pull rate sources lists every source we use and several we rejected.