strateva.INDEPENDENT REVIEW · BY COMMISSION
We examine your results, try to reproduce them and look for faults in the data, methodology and code. So you know what supports your backtest. And what does not.
For researchers, developers and small teams with algorithmic strategies or research/shadow systems.
01 / SERVICES
Three kinds of engagement. The same commitment to evidence. No subscriptions or permanent platform access.
What do my existing results actually support?
No code execution, backtest reconstruction or repository audit.
Target: 3 business days
Request a reviewCan the result be reproduced and withstand the agreed tests?
Portfolios, new adapters and intraday data need a separate scope. Not a full code audit.
Target: 5–7 business days
Request a reviewWhat faults could invalidate your system’s evidence?
Not an unlimited review, certification or comprehensive cybersecurity audit.
Initial report target: 7 business days
Request a reviewFees exclude applicable taxes. Compatibility, scope, final price and start date are confirmed in writing before payment. Turnaround starts on the agreed date with complete materials; client waiting time is excluded.
| Coverage | Review | Validation | Audit |
|---|---|---|---|
| Critical evidence review | Included | Included | Included |
| Historical reproduction | No | Accepted route | Agreed scope |
| Code | No | Focused review | Agreed modules |
| States and controls | No | No | Agreed selection |
| Retest of client fixes | No | Separate quote | One bounded retest |
| Completed future holdout | No | No | No |
Compatibility enquiries are open. Code-execution engagements are accepted only after checking the environment, isolation and reproduction route. There is no hosted self-service product.
02 / THE PROCESS
Send a summary without private code or secrets. This initial check is free; it is not an in-depth technical review.
Freeze the version, data, tests, deliverables, fees and start date. Exclusions are written down too.
We examine the evidence and run only the accepted tests. An impossible test is reported, not hidden.
Receive findings, reproductions or references, limitations and priorities. Coverage distinguishes tested, examined and not tested.
It can support an analysis of returns, not proof of how they were generated. We label it as imported-results analysis, not independent reproduction. A screenshot of a curve may be insufficient even for a documentary review.
Written rules, reports and code for the initial review. Python, Pine Script and other languages are not assumed executable: dependencies, data and compatibility are checked before quoting.
We document findings and explain next steps. Implementing fixes, optimising parameters or building an adapter requires a separate scope. Errors attributable to Strateva are corrected without a retest charge. Keep your frozen version for the retest: it does not extend the 7-day retention period in the privacy policy.
If we know the question cannot be answered before accepting, we decline or redefine the work. Later, we document what is not evaluable. Price does not depend on passing; cancellation and unperformed work are covered by the written agreement.
There is no checkout on this site. The proposal specifies taxes, milestones and terms: review before starting; validation and audit normally 50% at the start and 50% on delivery. No unagreed extras.
Access, confidentiality, retention and permission for AI tools are agreed before private code is shared. Do not send trading keys, credentials or secrets. NDA terms are reviewed before access is granted.
03 / LET’S TALK ABOUT YOUR STRATEGY
Tell us what you want checked, the market, instruments, frequency and the code or data you have. We will explain whether it fits and what the scope would be.