Methodology: how we measure Pokémon TCG pull rates

The short version

Step 1
Find the original opening
Case breaks, store tallies, and 100-box streams. Sites that repeat someone else's numbers are traced back to the opening they came from.
Step 2
Check every count
We match each tally against the original article, video, or post. Openings we can't verify are listed but never counted.
Step 3
Combine what agrees
Verified openings are added together when they agree. When they don't, we say so instead of averaging the disagreement away.
Step 4
Show how sure we are
Every rate carries its sample size, a 95% range, and a plain-language confidence label.

What counts as evidence

An opening only feeds our rates if it passes every check on the left. Anything else is listed with the reason, so you can see what we left out.

Counts

Physical packs opened and tallied by the person who published the numbers
Counts we can check against the original article, video, or post
Openings large enough to say something about the rarity in question
Rarity definitions that match the set's official card list

Doesn't count

Repeats another source: Seven sites quoting one study is still one study.
Counts we can't verify: No article, video, or tally to check against.
Demos and models: Simulations, calculators, and “expected” rates aren't observations.
Mismatched rarities: Groups like “full arts” that mix rarities we track separately.
Self-reported: Kept as a separate series, never mixed into the main estimate.
Too small: Shown for context, never combined.

Original vs. repeated numbers

Most pull-rate pages don't open packs. They repeat a number from someone who did, often with rounding, relabeling, or the wrong rarity attached. We trace every figure back to the opening it came from and count that opening once, no matter how many sites repeat it.

Our fact-check so far

We've checked 42 pages that republish pull rates against the openings they cite.

18 match the original
13 partly match
11 get it wrong

How we calculate a rate

Counts, not percentages

We store hits and packs for each opening and rarity. When a source only publishes a rate and a margin of error, we work the pack count back out from those two numbers and mark the result as reconstructed.

One headline number

With one verified opening, the headline is the rate that opening published. With several, it's their hits added together divided by their packs added together, so bigger openings carry more weight.

A range, not just a number

Next to each rate we show a 95% range: where the true rate very likely sits given how many packs were opened. Small openings get wide ranges; huge ones get tight ranges. (For the statistically curious: a Beta posterior with a Jeffreys prior.)

When openings disagree

Before combining openings we test whether they're consistent with one shared rate. If they aren't, we flag the set as “openings disagree” and widen the range rather than hide the conflict.

Odds for one specific card

Unless an opening counted individual cards, we assume every card in a rarity is equally likely. So with 11 Special Illustration Rares, one specific SIR is 11 times rarer than “any SIR.” We label these as modeled.

How sure we are

Every rate gets one of four labels, based on how wide its 95% range is compared with the rate itself.

Label	What it means
Very sure	The 95% range is within ±25% of the rate, from at least one verified opening.
Fairly sure	The range is within ±25–50%. Good enough to plan with; expect it to move.
Rough estimate	The range is wider than ±50%. Directionally useful, not precise.
Estimate	No verified opening for this set yet. Based on similar sets, with wide ranges.

Pack value and the simulator

Pull rates tell you what you'll probably pull. Value tells you what those pulls are worth. We keep the two separate so a price swing never changes a pull rate.

Card prices: TCGplayer market prices, refreshed daily and dated on every page.
What you'd actually get: Market price minus 13% selling fees and $0.30 per card. Cards under $1.50 count as $0.05 bulk.
What's counted: Hits only. Commons, uncommons, regular rares, and reverse holos are left out, so values run slightly low.
How packs are built: Each pack has a few slots, and each slot gives at most one hit, matching how real packs are laid out.
Uncertainty: The simulator can redraw rates from their range on every run, so small studies show wider outcomes.
Exact math first: Odds like “chance of at least one SIR in a box” are calculated exactly. Simulation is only for value.

Updates and corrections

Our evidence registry is versioned. Every change to a rate is recorded with the reason, and each set page shows when its numbers were last checked. When a new verified opening lands, the rates, ranges, and labels update together.

No one pays to change a number. Prices, ads, and affiliate links never feed into pull rates, and we're not affiliated with The Pokémon Company, Nintendo, or TCGplayer.

Find the original opening

Check every count

Combine what agrees

Show how sure we are