How Do Case Opening Sites Work? Odds Explained
How case opening sites set their odds, how provably fair rolls work, how to verify one yourself, and what the house edge on a case really is.
A case opening site sells virtual cases, each containing a list of CS2 skins with a drop chance for each. You pay the case price, a roll picks an item, and you can keep it, sell it back for balance or withdraw it to Steam.
Case odds and expected value
Each item has a drop chance and a market value. Multiply each value by its chance and add them up: that's the case's expected value. The difference between expected value and case price is the house edge.
| Item | Value | Chance | Value × chance |
|---|---|---|---|
| Knife | $200.00 | 0.5% | $1.00 |
| Rifle skin | $20.00 | 5% | $1.00 |
| Pistol skin | $2.00 | 30% | $0.60 |
| Common skin | $0.20 | 64.5% | $0.13 |
| Expected value | $2.73 |
If that case costs $3.00, the edge is 9%. Most cases are priced so the large majority of opens return less than the case price.
How provably fair works
1. Before you open, the site shows a hash of its secret server seed. 2. You have a client seed, which you can usually change. 3. A nonce counts your rolls. 4. The roll is calculated from server seed, client seed and nonce. 5. After you rotate seeds, the site reveals the server seed. You hash it yourself: if it matches the hash shown earlier, the site couldn't have changed the outcome.
How to verify a roll
Copy the revealed server seed, your client seed and the nonce into the site's verifier or an independent one, and check the result matches your roll. If a site doesn't let you do this, its "provably fair" label is decoration.
What provably fair doesn't tell you
It proves the roll wasn't changed. It doesn't tell you whether the odds are generous: that's the expected-value calculation above. And it says nothing about whether withdrawals work, which is why we test those separately.
Why "provably fair" matters for skins
Skin sites aren't licensed or audited like regulated casinos, so provably fair systems are the main way a player can check results. They prove that the site committed to an outcome before you played and couldn't change it afterwards. They don't prove that the odds are generous, so the expected-value calculation still matters.
Server seed, client seed and nonce explained
| Part | Who controls it | Purpose |
|---|---|---|
| Server seed | The site, hidden until rotated | Secret input fixed before play |
| Server seed hash | Shown to you before play | Commitment you can check later |
| Client seed | You (changeable) | Stops the site choosing results to suit itself |
| Nonce | Counts up each roll | Makes each roll unique |
How sites turn seeds into a result
The three inputs are combined and hashed; the hash is converted to a number, and the number is mapped to an outcome, such as an item in the case's drop table or a slot on a roulette wheel. Because hashing is deterministic, anyone with the same inputs gets the same result.
Red flags in case opening
- Odds not listed, or listed but not adding up to 100%.
- No way to change your client seed.
- No server-seed hash shown before you open.
- A verifier that only works on the site and doesn't explain its method.
- Case prices far above the expected value of their contents.
Worked example: verifying one roll
1. Before opening, note the server seed hash shown on the fairness page. 2. Set your own client seed, for example a random word. 3. Open a case; note the nonce for that roll. 4. Rotate seeds; the site reveals the old server seed. 5. Hash the revealed server seed and confirm it matches the hash from step 1. 6. Put server seed, client seed and nonce into the verifier; the result should map to the item you received.
If any step fails, stop playing on that site and report it.
Expected value vs. what you feel
Most opens return less than the case price, while occasional big hits make the average look better than most sessions feel. A case with a 90% return might give you less than half your spend back in a typical session of 20 opens, with a small chance of a large hit. That gap between average and typical outcome is why budgets matter.