Process
Six stages, no surprises in the last one.
The whole design of the engagement is to move discovery forward — earlier findings, cheaper fixes — so that the expensive stage, evaluation, has nothing left to discover. Start it at design and most of it happens inside your own sprints.
Where to start
You can join this process at three points.
The six stages below describe a full engagement from a finished build. They compress considerably if you bring us in earlier — and expand if you bring us in later.
At design
Stages one and two happen alongside your sprints, and stage three becomes continuous rather than a block at the end. Most findings never get written because they never get built.
At code freeze
The standard engagement described below, start to finish. Everything is fixable; some of it is tedious because the build is already stable.
After a knock-back
We start at the laboratory's findings rather than our own, work backwards to root cause, and prepare you for a clean next submission. It works — it just costs more than it needed to.
Scope & jurisdiction lock
We fix exactly which standard, which version and which jurisdiction appendix your submission will be judged against, and confirm it with the receiving laboratory or regulator. Everything downstream traces back to this document. Ambiguity here is the single most expensive thing in the whole process.
Evidence intake
You hand over the build, the math model, the paytable and reel data, the RNG description and whatever documentation exists. We inventory it, hash the build, and produce a gap list of what is missing before any testing starts. Work is done under NDA on isolated infrastructure.
Readiness testing & RTP verification
The main pass. Requirement-by-requirement testing against the locked rule set, alongside Monte-Carlo simulation of your engine for RTP, volatility and hit-frequency verification. Findings are graded and issued as they are confirmed, not held to the end.
Remediation to standard
The stage that decides whether you pay for one evaluation or three. Every finding comes with the clause, the reason and a concrete recommended change; your engineers implement it and we verify the change against the criterion before it is called closed. Where a fix touches the math, verification is re-run rather than assumed. Where a finding is a judgement call, we document the rationale you will put to the laboratory.
Dossier assembly
The submission pack is built in the receiving body's expected structure: game description, math report, RNG statement, build manifest with hashes, change log, declarations, screen captures and test evidence — with a final reconciliation pass so the binary, the paytable and the numbers all agree.
Submission support
We stay on through evaluation. Requests for information are drafted technically and returned quickly, clarifications are tracked against the register, and any conditional finding gets a documented resolution. When approval lands, the evidence set is archived for your next market.
Working with us
Practical terms
- Under NDA from first contact. Math models and reel data are the crown jewels; we treat them that way.
- Isolated environments. Client code and builds are kept segregated, and destroyed or returned on completion at your instruction.
- Fixed scope, named deliverables. You know what documents you are getting before work starts.
- Findings as they land. Blockers reach you the day they are confirmed, not in a report six weeks later.
- Reproducible by anyone. Every number we publish ships with the seed and build hash needed to regenerate it.
What to send for a first conversation
- Target market(s) and intended submission date
- Game list, with bet modes and buy-feature variants
- Declared theoretical RTP per game and per mode
- Engine language and how a headless simulation run is invoked
- Whether the game has been submitted anywhere before, and any prior findings
- Existing documentation, however incomplete
None of this is needed to start a conversation — it is what lets us give you a scoped answer instead of a range.
Next step
Send us one game and one target market.
Finished, half-built or still on a whiteboard — we will come back with a scoped plan, the document list your submission needs, and an honest view of what has to change before it is worth paying a laboratory to look at it.