We are traders , sometimes we are using high leverage or high trade volume, and when we doing so we are risking our capital, gainium.io has to ensure us there will be very minimal bugs and errors, a little error can liquidate our positions while we were unaware of the situation. I have dealt with two recent incidents
1 , once strategy didn’t open a deal , and it was 15% profit and it would have eased my overall drawdown. But it didn’t happened I reported that bug and @aressanch made sure it will not happen again
2, second time when bot opened a deal and didn’t closed at 0.34% loss , and it kept open for hours and caused me 10% loss i had to close it manually.
This overall scenario has increased my drawdown upto 26% it could have lower if such incidents don’t happen so often.
If you are not willing to have a loss because of a bug then I’m sorry to tell you this platform is not for you. The reason why we can add features faster than anyone is because the dev knows all code base and he can release features fast. But to release features fast we do minimal testing. If we did through testing, that would take weeks for each feature and we wouldn’t have even half of what we have today. So bugs are part of the process.
I love this platform, i don’t want to leave it, I can also deal with bugs, I want to be part this beautiful project. If bugs are part of it , i can deal with them by reporting and by constantly monitoring unless we are stable with no turbulence . I hope you are not offended by my concern in this post.
Yes no worries, I understand the frustration. We do have a policy to give credits 2x the loss caused by a bug. If you let me know what was the loss I can credit your account.
What about using AI to create unit tests with a high line and mutation coverage? If those were integrated into your built process, you could easily find out whether a change affects other parts of the application as well and verify whether that is an expected result or not.
We already have unit tests, which would trigger for the basic problems. But it’s impossible to cover edge cases using those. We are thinking about another solution, a separate server running thousands of random configurations, we may be able to detect problems before users run into them.
So like executing automated monkey, performance or load tests as addition to a fix set of unit tests?
Even empty unit tests would be enough to get mutation testing started, where it is meant to measure the quality of the unit tests and can help to find a minimal set of those to detect possibly unwanted changes while testing. The main use case of fix unit and e2e test cases though remains to reliably verify whether the code is still doing what it was meant to do.
I can tell you that there aren’t platforms which give you any kind of compensation even when bugs are disruptive as happened several times in 3commas and other platforms.
Same for exchanges - people got hacked several times and the platform just ignore the customers