Mystery box house edge

Pool EV = Σ (item value × probability)
House edge = (Box price − Pool EV) ÷ Box price

A box priced at 50:

House edge = (50 − 42.89) ÷ 50 = 14.2%

50 in, 42.89 of expected value out.

The grand prize dominates the site's marketing and contributes 4.00 of the 42.89 — under a tenth. Over 71% of outcomes are the base tier, worth roughly a third of the box price.

That distribution is the product. It isn't hidden, and almost nobody calculates it.

The published odds are usually honest. Item values are where mystery boxes diverge from cleaner verticals like CS2 skins, because physical goods don't have a live market price.

RRP instead of market price. Items are commonly valued at manufacturer's recommended retail price. Real market price on the same item is frequently 20–40% lower. A box with a nominal 14% edge, valued at RRP against goods trading 30% below it, carries a real edge closer to 40%.

Buy-back rates. Most platforms let you sell an item back for credit rather than shipping it. The buy-back is typically 70–90% of the listed value. Take it and the pool's effective EV drops proportionally.

Credit, not cash. Buy-back usually pays site credit, not withdrawable money. An item sold back at 80% of an inflated valuation, into credit that can only be spent on more boxes, is worth far less than its line in the EV table suggests.

Shipping and duty. On physical goods, the recipient often bears both. A 45 item costing 15 to receive is a 30 item.

Run all four adjustments before treating a calculated EV as real. The stated pool EV is the ceiling, not the estimate.

Mystery boxes look comparable to CS2 cases on paper. They're usually worse in practice, because skins have a liquid market that constrains how far a site can inflate valuations and physical goods don't.

→How case site odds actually work covers the same calculation with a liquid market underneath.

Standard independence rules apply and get misread constantly.

A 0.05% grand prize is 1 in 2,000 per box, not a guarantee at 2,000 boxes.

100,000 spent for a 63% shot at an 8,000 item. No pity timer, no accumulated progress, no due-a-win state.

Mystery box rates run lower than CS2 or casino rates, and the reason is commission structure rather than the house edge.

The vertical's affiliate contracts pay less. Margins are wide, but a large share goes to inventory cost, fulfilment, shipping and returns rather than to acquisition. There's less commission available to share, so the rebate is smaller even though the product is more expensive.

Over 1,000 of box opening:

Note also that some box platforms quote cashback as a percentage of turnover rather than of theoretical loss — the "3% down to 0.25%" style figures. Convert before comparing to a casino's "40%".

→Why mystery box cashback rates are lower.

1. Published odds per tier. If they aren't published, stop there.
2. Calculate the pool EV. Ten minutes with the odds table.
3. Check item values against real market prices. Search two or three listed items.
4. Buy-back rate, and whether it pays cash or credit.
5. Shipping and customs terms. Who pays, and how much on a mid-tier item.
6. Withdrawal restrictions. Whether credit can leave the platform.

Points three and four move the answer more than anything else on the list.

At a nominal 14% edge, 1,000 spent on boxes returns roughly 860 in listed value. After realistic valuation and liquidation adjustments, the figure is frequently closer to 600.

Buying the item you want directly is better value by a wide margin, every time, before and after cashback. Boxes are a lottery mechanic wrapped around retail goods, and the entertainment is the product.

That's a legitimate purchase if it's what you're buying. It isn't a way to acquire things cheaply, and cashback narrows the gap by a few percentage points at most.