Mystery pack odds — what's posted vs what actually happens

Posted odds describe a pool. Observed odds describe reality. A mystery pack's published odds are a snapshot of the pool as it was stocked. The moment people start ripping, the two numbers begin to diverge — and every chase card that leaves makes the posted figure a little more fictional for the next buyer.

What are the real odds on a Courtyard pack?

Stated plainly, using the two best-sampled packs in our index. On Pokémon Starter Pack, a $25 pack with 6,087 indexed pulls, a card worth at least the price came out 25.7% of the time and a card worth double came out 2.6% of the time. On Pokémon Ultra (Spicy), a $249 pack, 37.9% beat the price and 8.5% doubled it.

Two things stand out. First, the losing bar is the biggest one on both packs — that is normal for mystery packs and not evidence of anything wrong. Second, the more expensive pack has better odds of beating its price, which is the opposite of what most buyers assume. Price tier and odds quality are not the same axis, and neither one is visible from a product page.

Why static odds go stale: pool depletion

Imagine a 1,000-pack pool holding exactly one $5,000 grail. The posted odds say 0.1%, and on day one that is exactly right. After 900 packs have been ripped and the grail is gone, the real odds of hitting it are zero — while the posted number still reads 0.1%.

No deception is required for that to happen. It is arithmetic operating on a fixed pool, and pool depletion is the name for it. The same mechanism quietly drags EV down over a pack's life, which is why a lifetime average keeps crediting a pack for cards that stopped appearing weeks ago — and why our calibrated EV weights recent pulls more heavily than old ones.

Restocks complicate this in the buyer's favour. Pools do get replenished, and when they do, odds improve sharply and without announcement. Depletion is not a one-way ratchet — it is a cycle, and the useful skill is knowing which phase a pack is in. That is what the 48-hour drift column on each pack page is for: a tier falling over two days is the signature of good cards no longer coming out, which is consistent with a draining pool, though nobody outside Courtyard can confirm the inventory directly.

Are the pulls actually random? What the data can and cannot show

“Is it rigged” is one of the most common things people search about mystery packs, so it deserves a direct answer rather than a reassuring one. Nothing in 35,075 indexed pulls looks manipulated. The value distribution has the ordinary right-skewed shape a prize pool produces; the cards carry verifiable grading identifiers; the large hits are real and do land.

What we can honestly claim is narrow, and stating the limits matters more than the reassurance. We observe outputs — the cards that came out and what they were worth. We cannot see Courtyard's inventory, the number of packs remaining, the seeding of any pool, or the mechanism that selects a card. Nobody outside the company can. So we can say the outputs look like a normal prize pool, and we cannot certify a process we have no window into. Any site claiming more than that is overstating what public data supports.

In practice, what people experience as rigging is almost always the mean-median gap. A pack advertising a healthy average while delivering eight straight below-median rips feels engineered. It is simply a long tail doing what long tails do, and understanding that turns an infuriating experience into a predictable one.

Posted odds vs observed odds: which should you trust?

Neither blindly. Posted odds are a statement about the pool at stocking time — genuinely useful when a pack has just launched, and progressively less informative every day after. Observed odds are a measurement of what has actually come out — genuinely useful once enough pulls exist, and meaningless when they do not.

The practical rule: on a brand-new pack, the posted odds are the better guide because there is no observed sample yet. After a few hundred pulls, observed odds win, because they reflect the pool as it stands rather than as it was printed. PullValue refuses to publish odds below five indexed pulls and flags anything under thirty as an early read, for exactly this reason.

How to use odds like a buyer rather than a gambler

Three checks, in order, before committing money to a pack:

1. The hit rate. The share of pulls that met the price — the odds number that pays the bills. Anything under 25% means three of every four rips come back light, and you should want compensation for that in the form of a strong cash-out ratio.

2. The median pull against the price. Your realistic evening, uncorrupted by grails. If the median sits far below the price, the pack is a lottery ticket whatever its average says.

3. Freshness and sample size. Odds computed on last week's pool beat odds computed on launch day's, and odds computed on four thousand pulls beat odds computed on forty. Both are printed on every pack page; neither is optional context.

The scanner puts all three beside every tracked pack, and the methodology page documents the exact tier bucketing, including what we refuse to compute and why. The underlying discipline is simple: treat odds as a measurement with a date and a sample size attached, never as a fixed property of the product.