How to verify a slot result
Updated 5 Nov 2026. Examples use live figures from the Engine fairness API.
Stake Engine slots use a provably fair process: every round is drawn from a published outcome table using inputs you can check afterwards. This guide walks through how you verify a round yourself, what you need to collect, and how to read the result. If you want the background first, start with what provably fair means.
What you are checking
Verifying a round answers one question: did this result come from the published table, using a casino seed that was fixed before I bet? To answer it you need four things:
- The server seed that was active when you played, in its revealed (unhashed) form.
- The hashed server seed you were shown before you played.
- Your client seed for that period of play.
- The nonce of the specific round, which is its position in the sequence of bets under that seed pair.
You also need to know which game and which mode the round was played in, since each mode has its own table.
Step by step
- Write down the hashed server seed before you play. Your casino's fairness settings show the hash of the current server seed. Copy it somewhere you control. This is your record of what the casino committed to.
- Note your client seed. It appears in the same fairness settings. You can keep the default or set your own. Setting your own is a simple way to make sure the casino could not have predicted it.
- Play, and note the nonce of any round you want to check. Bet history usually lists the nonce for each round. The first bet under a new seed pair typically starts at 0 or 1 and each bet adds one.
- Rotate the seed to reveal it. A server seed is only revealed once it is retired. Rotating it in your fairness settings ends the current pair, shows you the old server seed in plain form and commits a new hashed seed for future play.
- Check the hash. Run the revealed server seed through SHA-256 and compare the result with the hash you wrote down in step 1. They must match exactly. If they do, the seed was fixed before you played.
- Replay the round. Feed the server seed, client seed and nonce into a replay tool for the same game and mode. It recomputes which entry of the outcome table was selected.
- Compare the outcome. The replayed entry should show the same payout, and usually the same symbols and features, as the round you saw.
Where the replay comes from
Engine, formerly branded Stake Engine, runs a public fairness API at fair.engine.io. Conceptually it offers two things that matter for verification:
- An outcome replay. Given a game, a mode and the three inputs, it returns the entry those inputs select. Because the calculation is deterministic, anyone sending the same inputs gets the same answer.
- The full event tables. For each mode, the complete list of possible outcomes and their weights. This is what lets you check not only that a round came from the table, but what the table contains.
You can use the replay directly, use a tool that calls it for you, or, if you are comfortable with code, recompute the HMAC-SHA256 step yourself against the published table. All three routes should agree.
How the inputs become a result
It helps to know roughly what the replay is doing, so you can tell whether a mismatch is a real problem or a mistake in your inputs:
- The server seed, client seed and nonce are combined with HMAC-SHA256, with the server seed acting as the secret key.
- The output bits are converted into a number within the total weight of the mode's table.
- That number falls inside the range belonging to one table entry, and that entry is the round.
Change any one character of any input and the result jumps to an unrelated entry. That sensitivity is why a match is meaningful and also why typos are the most common cause of a failed check.
If the replay does not match
Before drawing any conclusion, work through the usual causes:
- Wrong nonce. Check whether your bet history counts from 0 or 1, and that you picked the right round.
- Wrong seed pair. If you changed your client seed or rotated the server seed mid-session, the round may belong to a different pair.
- Wrong mode. A bonus buy and a base spin use different tables. The same inputs give different results in each.
- Copy errors. Stray spaces, a missing character or a line break will change the hash completely.
- Game version. If a game has been updated, make sure the replay uses the version that was live when you played.
If everything checks out and the replay still differs, keep a record of the inputs, the hash you noted and the round details, and raise it with your casino's support with those specifics.
Checking the table, not just the round
Round verification tells you that a draw was honest. The table tells you what you are drawing from. Because the event tables are published, the figures that describe a mode can be recomputed by anyone:
- RTP of the base mode of Waylanders Forge: 98.00%.
- Pays anything: 1 in 16 rounds.
- Reaches its top payout of 80,085×: 1 in 2,002,135 rounds.
Each of those comes straight from the published table, computed by us rather than copied from a declared figure.
The RTP checker shows these figures for any game and mode.
Good habits
- Set your own client seed when you start, and change it from time to time.
- Note the hashed server seed before playing so you have an independent record.
- Rotate the seed when you want to check a batch of rounds; you cannot see the active server seed until you do.
- Spot-check a few rounds rather than none. A check you never run proves nothing.
Frequently asked questions
Why can't I see the current server seed?
If you could, you could compute upcoming results before betting. It stays hidden until you rotate it, and only the hash is shown in the meantime.
Does verifying a round change future results?
No. Rotating the seed starts a new pair, but the odds of every round are still set by the same table.
Do I need to verify every round?
No. Any round you choose can be checked, so occasional spot checks are a reasonable approach.