Palworld TCG Pull Rates: What 3,600 Opened Packs Actually Show

Dawn of Palpagos pull rates from four public datasets: RR 1 in 3.1 packs, SR 1 in 5.4, OSR 1 in 11. Plus why the per-box odds you're reading are wrong.

Bushiroad has published no official pull rates for Palworld's first set. Across four public datasets covering roughly 3,600 packs as of August 7, 2026, a Double Rare shows up in about 1 in 3.1 packs, a Super Rare in 1 in 5.4, and an Over Super Rare in 1 in 11 — roughly one OSR per 12-pack box. The two rarest tiers are still moving, and not in the direction you might expect.

Palworld OFFICIAL CARD GAME launched worldwide on July 30, 2026, and one week in its first set — BP01 "Dawn of Palpagos" — is being opened at scale. The community tracker anchoring this article has grown from 951 logged packs at launch to 3,378. This is what the numbers say now, which datasets they come from, and the one widely repeated per-box figure that does not mean what people think it means.

What is in a pack, a box and a case?

Confirmed product structure, from the official product page and retail listings:

The rarity ladder runs C (Common), U (Uncommon), R (Rare) and RR (Double Rare) in the base set, then the parallels: SR (Super Rare), OSR (Over Super Rare), SP (Special) and SSP (Super Special Parallel). The chase cards are parallels — alternate-art versions of cards that already exist at base rarity.

The pack itself follows a fixed slot model: 3 Commons, 3 Uncommons and 1 Rare, with one Uncommon slot upgraded to Double Rare or better in roughly 6 of every 10 packs. That model is worth stating because it is checkable, and it checks out — see the card-level data below.

What are the Palworld TCG pull rates?

Four independent public datasets exist. Here is what each reports, converted to the same unit — expected cards per 12-pack box.

The top three rows agree within about 10% across four samples gathered by different people using different methods. That is the strongest signal in this data: RR, SR and OSR rates are settled. Expressed per pack, the community tracker's figures as of August 7, 2026 are RR 32.4% (1 in 3.1 packs), SR 18.4% (1 in 5.4) and OSR 9.47% (1 in 11) — each within a point of where it stood at launch on a sample 3.5 times larger.

The bottom two rows are a different story, in two ways. Across sources they still disagree — SP lands anywhere from 1 in 24 packs to 1 in 97, SSP from 1 in 48 to 1 in 150. And they are the only rows that moved as the tracker's sample grew: between July 30 and August 7 its SP rate went from about 1 in 58 packs to 1 in 31, and SSP from roughly 1 in 200 to 1 in 54, while RR, SR and OSR held still. Two readings fit that pattern. Either the launch-week sample genuinely undershot the rare tiers, or — more likely, given that only the chase tiers inflated — people log the pack with the SSP in it more readily than the eleven packs that whiffed. Treat the tracker's SP and SSP rates as the optimistic end of their ranges, not the middle.

Why the "per box" odds you are reading are wrong

This is the part worth your attention, because the most-cited source for Palworld pull rates prints a per-box column that is not measured.

The community tracker reports, alongside its per-pack rates, a "chance per box": as of August 7, 2026 that column reads 99% for RR, 91% for SR, 70% for OSR, 32% for SP, 20% for SSP. Those five numbers are reproducible to the decimal with one formula — the probability of at least one hit in 12 independent packs, or 1 − (1 − p)¹².

Run it on OSR: 1 − (1 − 0.0947)¹² = 69.7%. On SR: 1 − (1 − 0.184)¹² = 91.3%. On SSP: 20.2%. Every figure matches, at launch and again on the sample 3.5 times larger. The per-box column is therefore a model output derived from the per-pack rate, not an observation of boxes — and the model assumes each pack is an independent draw.

That assumption is contradicted by the same site's own finding: reported case openings yield 4 copies of every Double Rare — a complete playset — plus around 6 of each Rare, which is the signature of a collated case rather than a randomly filled one. If packs within a box are collated, hits are distributed more evenly than chance, and the true per-box odds are higher than the binomial says.

The observed boxes support exactly that. All 16 fully documented boxes — one 12-box case plus a separate four-box opening — contained at least one OSR. If the real per-box probability were 70%, the chance of 16 out of 16 is about 0.3%. Either boxes are collated, or the public sample is biased toward lucky boxes. Both are probably true to some degree; the playset evidence makes collation the better-supported explanation.

A week after launch, the tracker's own page now says both things at once. Its live calculator still prints the binomial 70% for OSR, while its static summary table lists the OSR box odds as "~100% — Every box" and its FAQ states that "community openings so far report at least one Over Super Rare in every booster box." The observed-box reading is the one the evidence supports; it is just worth knowing that the two claims share a page.

The practical takeaway: treat "about one OSR per box" as a reasonable planning assumption, treat "70% of boxes have one" as an artifact of a formula, and treat neither as a guarantee, because Bushiroad has confirmed nothing.

One new tier appeared in the tracker this week: the "God Pack", its label for an all-hit pack. Its logged data reports one in roughly 29 boxes (3.4%) as of August 7, 2026 — while the same site's static table still lists the tier as "no confirmed sighting yet." Logger-entered, publisher-unconfirmed, and contradicted on its own page: file it next to the SSP rate as a number to watch, not one to buy sealed product on.

How reliable is community pull data?

Two opposite biases are visible in this dataset, and they partly cancel.

Reddit photo posts skew lucky — people post the box that hit, not the box that didn't. One commenter on the four-box thread made the opposing observation: every YouTube box opening he had watched contained exactly one OSR and never two, while Reddit photo posts routinely show more. Video openings are harder to cherry-pick after the fact, so they skew toward the honest floor.

The community tracker is the best-constructed of the four sources: it states plainly that its percentages "come solely from the numbers people have entered on this page. No estimates, no manufacturer figures, no scraped videos." That statement is still on the page a week after launch, and it is why the tracker's 3,378 packs anchor the table above. The self-reporting it describes is also, of course, exactly the mechanism behind the chase-tier drift noted earlier — a methodology can be honest and biased at the same time.

One caveat on the total: these datasets are not cleanly additive. At least one of the Reddit openers was asked to log the same pulls into the community tracker and agreed, so some packs appear twice across the four sources. Roughly 3,600 packs is the honest scale of the public record as of August 7, 2026, not a sum — and the three Reddit-sourced samples are frozen at launch size, so the tracker is now more than 90% of it.

What does a card-level sample add?

The 97-pack sample is the only one that recorded every card rather than just the hits, so it is the only way to check the slot model. Across 679 cards:

Multiply by 7 cards per pack and the slot model falls out precisely: 2.98 Commons, 1.02 Rares, and 0.58 cards of RR-or-better per pack — that last figure being the "roughly 6 in 10 packs" upgrade, arrived at independently. The Uncommon count lands at 2.41 rather than 3, which is exactly the 0.58 upgraded slot being taken out of the Uncommon pool.

That is a useful thing to know before you buy: the upgrade replaces an Uncommon, so a pack with an SSP in it is not a pack with extra cards, it is a pack with one fewer Uncommon.

The same sample gives the non-rarity splits: 56.11% of cards are Pals, 21.21% Structures, 11.49% Events, 10.90% Gear, and 9.87% are Lucky Pals — about one in ten cards. Colors are near-even at 22–26% each except Colorless, which sits at 3.24% and is effectively a trial-deck rarity.

What about box price and card values?

Cardlogs does not track Palworld yet — it is planned, not live, so nothing here comes from Cardlogs platform data and we are not going to dress up someone else's numbers as ours.

What we can offer is our own hand-gathered eBay read, published in full in the Palworld launch guide: as of August 6, 2026 — one week in — the English booster box asks a median $172.50, up 44% from $119.89 on day two, with the cheapest listing at $155 sitting above the old median; the Japanese print asks $105 and Simplified Chinese $39.13. The top chase asks have tripled, led by the Gold Cattiva Soul SSS at $2,000+, all JP prints. Every one of those figures is an asking price, not a completed sale — eBay's sold view is not readable without a signed-in session, so no sold median exists for this game yet, from us or anyone.

One collector's note: the first print is misprinted

Worth knowing before you sort your sealed product. The official BP01 product page carries an errata notice: "A printing error (Misspelling of the word 'Palworld') was found on the packs and box of the product." Bushiroad's stated remedy is that it "will adjust this in future reprints".

That makes every launch-window pack and box a first print with an officially acknowledged, visually identifiable error — the exact conditions under which packaging variants become collectible. Whether a premium develops depends on how quickly the reprint arrives and how much product exists, but the identifying feature is documented by the publisher, which is unusual and useful. Riftbound collectors are already paying up for a first print distinguished by nothing more than bolder text and a different importer block.

A week in, sellers have already turned it into a tier: English box listings now lead with "Misprint / 1st Edition" at $160–$300. Until a corrected reprint actually exists, that premium buys the default product — per Bushiroad's own errata, the misspelling covers the entire first run.

Where this data comes from

This article aggregates four public sources: a community pull tracker with 3,378 logged packs (re-pulled August 7, 2026; 951 at launch), a card-level Google Sheets sample of 97 packs (679 cards) compiled from filmed openings, one complete 12-box case log, and a four-box opening drawn from a single master case. Product structure is from the official BP01 product page and retail listings. All per-box figures in the comparison table are expected counts, converted to a common unit by the author; the binomial analysis of the per-box column and the launch-to-week-one drift comparison are ours. No Bushiroad pull rates exist to check any of it against.

Cardlogs aggregates prices nightly from TCGplayer, Cardmarket and eBay sold listings across the games we cover — One Piece, Dragon Ball, Riftbound and vintage Pokémon — with graded population data and set-level EV. Palworld is on the roadmap; when it lands, these pull rates become the input to a real Box EV calculation, which is the number that actually tells you whether to open or hold. Until then, the honest position is the one above: three tiers measured, two tiers guessed, and nothing official. For how the Box EV calculation and the multi-source pricing behind it actually work, see how Cardlogs tracks TCG prices. Follow along at cardlogs.com, free on web, iOS and Android.