Trust and safety

Fair draws

A fair draw needs unpredictable numbers and every line counted.

Two things must be true#

The numbers are unpredictable. They come from a trusted random source, requested only after sales close.
Every line is counted. Each accepted line is in the final count, and nothing is added later.

Checking the count doesn’t prove the numbers were random. Both checks matter.

Check a draw yourself#

The blockchain keeps a public record of every draw. On the test network, you can read it on Stablescan (opens in a new tab).

The round, its closing time and its number ranges (normalMax, bonusMax).
NumbersRequested came after the close, and NumbersDrawn matches that request.
The queue matches QueueProved.
SettlementVerified, the prizes owed and ClaimTreeRoot agree.
If the owner used the forceNumbers fallback, the request had really failed.

Where it stands today#

Heads up

The test network gets its numbers from ManualEntropyV2, a trusted operator behind the Pyth adapter. It is not production Pyth randomness, and it must not be described as such.

A real proof checker is deployed, but the full end-to-end test hasn’t been run. An audit and production testing are still needed.