-
Notifications
You must be signed in to change notification settings - Fork 0
FAQ
The wheel result is a pure cryptographic function of a server seed (which we commit to before you spin), your client seed, and a nonce. Because the commitment is published first, we cannot change the outcome afterward. You can recompute every result yourself — see Verifying Your Spins.
If we revealed it up front, the outcome of every future spin would be predictable. Instead we publish
SHA-256(serverSeed) (the commitment). When you Rotate, the raw seed is revealed so you can verify
the whole history — and the hash proves we never swapped it.
No. The server seed is fixed and its hash is published before you spin. Changing the seed would break
SHA-256(serverSeed) == commitment. The nonce is a server-owned counter that only ever increases, so
no result can be replayed.
No. The browser verifier (index.html) runs entirely offline — view source and confirm it makes no
network calls. Or run the Node CLI. Or re-implement the algorithm in any language.
They are the relative frequency of each wheel segment (the paytable). They are published in full in
weights.maczo.json and embedded in the site's verify links, so the exact segment can be reproduced.
It doesn't hide them — the prize values are printed on the wheel and the weights are published. You need the weights only to map the random number to a segment; the cryptography above guarantees that mapping is honest.
It only advances in the extremely rare case that all four 8-byte words of one HMAC digest fall in the biased tail and are rejected. It guarantees the sampling never runs out of randomness while staying perfectly unbiased.
Double-check you used the revealed server seed (after Rotate), the exact client seed and nonce shown for that spin, and the published weights. If it still fails, contact support with the values — but the math is deterministic, so a correct input always reproduces the published result.
Yes — the same HMAC-SHA256 commit-reveal family, with rejection sampling to remove modulo bias. The implementation is open source so anyone can audit it.