Beezie

Pick a machine, pay, and it opens immediately to one item — a graded slab, or sometimes a sealed box. You either keep it, in which case it stays vaulted at Brink's until you redeem the physical item, or you swap it back at its published value minus a 6% fee.

The four machines differ only in price and the pool they draw from: Wildcard at $30 mixes Pokémon, basketball and sealed product, while Gold and Platinum at $250 and $500 are almost entirely graded Pokémon and One Piece singles. Stock is published per machine and visibly moves as items are pulled.

Each machine publishes five tiers with a probability and a dollar band: on Wildcard that's 72.85% at $18–30, 21.17% at $31–60, 4.12% at $61–99, 1.54% at $100–199 and 0.32% at $200–2,000. Both ends of every band are bounded, and the probabilities sum to exactly 100.00%.

We sampled 1,000 real pulls from each machine and sorted them into those same bands. The observed mix tracked the published odds closely — 70% of Wildcard pulls landed in the bottom band against 73% published, and the other three machines matched within a couple of points. That is the check that matters, because a feed showing only the good pulls would produce a flattering distribution and a meaningless return.

On value, the aggregate came out at 106%, 108%, 102% and 101% across the four machines against 110–111% advertised. Both figures are gross, before the swap fee.

This is the number to hold onto. Across those same 4,000 pulls the MEDIAN return was 77%, 77%, 78% and 72% — meaning half of all pulls came back with less than about three-quarters of what they cost.

The gap between a 106% average and a 77% median is the entire product. Roughly seven pulls in ten land in the bottom value band, and the average is dragged up by the rare items at the top: a 1999 Shadowless Charizard at $4,850 from a $250 machine, a Mario Pikachu at $9,000 from a $500 one. Those are real and they do happen. They are also why the average is a poor description of what happens to you.

Then the 6% swap fee comes off whatever you pull, if you cash out rather than keep the item.

Swapping returns 94% of the item's published value — a flat 6% fee on every machine, stated up front. Claiming keeps the item in your name, vaulted, with the physical card redeemable whenever you want.

The feed shows which route people take, and the split is worth noticing: the overwhelming majority of pulls are swapped straight back, while the genuinely valuable ones tend to be claimed. That is exactly what you'd expect, and it means the value we measure is not distorted by only counting cash-outs — we count both, because the value is published either way.

It delivers real graded items, vaults them with Brink's, publishes odds that sum correctly with no open-ended top band, states its swap fee plainly, and publishes a live feed complete enough that we can check the odds against reality. When we did, they agreed. That's a stronger position than most of this category.

The honest caveats: values are Beezie's own swap figures rather than an independent price source, settlement runs in USDC through a wallet so the cash-out path is yours to manage, and the advertised return describes the machine over thousands of pulls rather than your session. The typical pull loses money — that's not a criticism of Beezie specifically, it's what a 110% average with a 77% median means arithmetically.