Provably Fair and Result Verification

A commitment works like a sealed note. We pick a secret value called a server seed. Before you place bets, we show you a hash of that server seed. A hash is a one-way fingerprint. It can be shared publicly, but it does not reveal the server seed itself in practical terms.
After play, we reveal the server seed. You can hash it on your side and compare it with the earlier fingerprint. If both match, the server seed stayed the same from the start. That matters because a changed server seed could change outcomes.
Hashing also helps you because it blocks prediction. If you only see the hash, future results cannot be calculated ahead of time.

Each verifiable bet uses 4 inputs: server seed, client seed, nonce, and cursor. All 4 are important because they play different roles in the calculation.

For each bet, we use a cryptographic function called HMAC_SHA256. Inputs include server seed, client seed, nonce, and cursor. The function returns a 32-byte output shown as a hex string.
Those bytes are raw randomness. After that, the bytes get converted into numbers that games can use.

Games need numbers such as card indexes, pocket numbers, or grid positions. We take the 32-byte output and split it into chunks of 4 bytes. Each 4-byte chunk is converted into a float from 0 to 1.
Use of 4 bytes per float increases precision. Higher precision reduces rounding bias compared with 1-byte conversion.
A single 32-byte output produces 8 floats because 32 / 4 = 8. If a game needs more than 8 floats, the cursor triggers another 32-byte output.

After floats exist, the system converts them into game events. A game event is the concrete outcome you see, such as a dice roll value, a roulette pocket, a card, or a mine location.
Conversion follows a standard rule:

• Multiply the float by the number of possible outcomes for that step;
• Round down to an integer index;
• Map that index to a game event.

A 52-card example is easy to picture. A float times 52 gives a value from 0 up to just under 52. Rounding down gives 0–51. That integer maps to a specific card index.
Multi-step games repeat this process for each event needed.

Some games require unique outcomes inside a single round. A deck cannot deal the same card twice in a single-deck model. Keno cannot pick the same square twice. Mines cannot place 2 mines on the same tile.
For these cases, the pool of available outcomes shrinks after each selection. After one item is selected, it is removed from the pool. The next float is multiplied by the new pool size, which is smaller by 1. This method prevents duplicates.
We use the Fisher–Yates shuffle approach to support this no-repeat logic. The core idea is simple: each selection reduces the list of available items, and the next selection uses the reduced list.

Each bet is independent. Results do not change based on prior wins, prior losses, time spent playing, or wager size history. The algorithm always uses the same rule set, and the nonce keeps bets distinct.
A belief that a win becomes “due” after many losses does not fit with this system. The next outcome is based on the next nonce and the same inputs, not on a memory of past events.

Verification can be done on our website, and it can also be done through community tools. Community tools help because they let you confirm results outside our platform.
Community projects often publish their verification procedures and keep them open for review. Using an external tool can help you confirm that your own calculation matches the recorded result.

We also take part in industry efforts that promote verifiable fairness in crypto gambling. One example is the Crypto Gambling Foundation network, which focuses on education and standards around provably fair systems.

Network participation does not replace your own checks. Your verification remains the most direct way to confirm results.

Provably fair confirms that results match the published algorithm and inputs. It proves that the server seed was committed before play and later revealed without change. It confirms that your client’s seed was part of the calculation.
Provably fair does not guarantee wins. It does not remove the house edge in games where a house edge exists. It does not turn gambling into a low-risk activity. It also does not remove the need for responsible play.

We use provably fair so you can verify results independently. The method relies on committed server seeds, player-controlled client seeds, and repeatable calculations through HMAC_SHA256. When you choose to verify, you can confirm that a bet outcome follows the same rule every time, with no hidden adjustment after the bet is placed.