Hello NoahA, Exactly Exactly, if I am a programmer and you pay me $20,000 to program your trading strategy/idea in NinjaTrader8, and I back test it and I see it makes $100,000 per year, I just take your algo for myself, make a million dollars with it scaling it up and then put a bug in your code give it back to you so it does not work for you, or just say I have no time to finish it, steal your trading system and take your $20K for myself and scale it up and make million dollars, not I am rich. Done. Lol, programmers do not care about no NDA. LOL HAHAHAH You can not trust people in this trading industry, that is why best to make your own money manual guessing the futures market clicking everyday. Or learn to code your own trading systems and back test it yourself.
Your reply is consistent with what I've found having some experience in trading and expert level in IT work. Traders tend to overestimate the value of their edge and manual systems, and undervalue the work of a good IT guy. A good "edge" in IT is often worth a lot more then a trading edge. I highly doubt he's moving forward with the trading edge because for one you have to understand how to use it in real time markets and two it's not going to be a marketable product or skill for him 99% of the time.
Well, that was my main concern expressed on the first page of this thread, so don't think I'm completely naive or blue eyed. As it's not a mechanical system/algorithm and requires some know-how and savvy to use, I decided to take the risk. Even after a year, it's apparent to me that the programmer does not fully understand what he have (or had) in his hands. If it was a fully automated system which would be extremely profitable without any user input I would of course never taken the chance.
One choice is to have someone set you up with the IDE, and then write the function library that you use to create the signals, order handling, P/L manager etc. Then take your scripts and port them. You will have to learn the IDE, and the new function calls and handling, including error handling. But then you will have: 1) a new development environment 2) examples of how to write additional functions 3)learn how to integrate the basic infrastructure and functions. 4) but the strategy logic is kept to you. This is the path I am following, sort of. Prototype in EasyLanguage, using all the Multichart libraries. Then port to c# IDE. I think about 50% of the work is getting the IDE setup, and writing a simple piece that enters and exits using a simple test signal. This is the part a friend could do with you. Of course, there are MANY MANY examples of this part so it is very accessible to learn and IMO worth the effort.
If you can't build it yourself then you are just the middleman with an idea where in almost all cases that idea is worthless, if something is valuable you control it yourself now the main problem is there are very few people in the world that can both create something and also make it work, it is one of the other. So the fact of someone making it work giving the code and idea to someone who creates it while being worried about where that code ends up is truly quite funny, sure there are institutions like Goldman who frontrun LTCM trades, and multi-nationals who take code from smaller players and make it their own, I know someone this happened to, but unless you have something like that, no one cares except you. Have an NDA it's smart, but in the end the probability of a programmer being able to use a profitable strategy, which there are so few of the probability of that is microscopic, is also almost zero those two ever meet.
There was a post on Twitter where, as an experiment, someone had a quote ~$2k for a software dev project. He also got a quote from a programmer whom uses chatGPT. After a week, the $2k quote became a ~$25k + 6 mths, the chatGPT assisted programmer completed the project within that week for <$2k. The chatGPT programmer was also much more communicative. Don't have the link, lost in the sands of timeline.
To add to that, my requirements are on the surface nearly identical to yours and I've been banging through gpt 4 for a month and automating analysis that I've been reluctant to farm out with two important differences Because privacy concerns are exponential using AI, the way I've done it is breaking everything down to smallest component parts so that I combine them myself, so that not only the variables but the trade logic is unknowable. This way, debugging is relatively easy too. And I found massive advantages to be able to iteratively develop like this. I'd be really surprised if you don't find similar. To give you an idea, with about an hour a day in one month, I've got around £15ks worth of dev based on previous work. First week was so torturous that I nearly gave up, but it's proved its worth beyond measure.
Actually, the guy I cooperated with allegedly used ChatGPT to help streamline his process by writing some of the code, but it didn’t seem to help him much. As a parting note he had suggestions for how things could have done differently and “improved”, but to me it only showed that he hadn’t really understood it all and wanted to solve the markets using “brute force” and machine learning. To me, programming is still a tool and I don’t think the riddle of the markets can be solved by programming alone. If that was the case – every programmer would be rich already.
I was asked to look at some ChatGPT code recently, it was useless as it didn't solve the underlying problem, the trading idea was fundamentally flawed so actually makes it worse for people, they truly believe they have something that works. Apparently Elton John was asked about the X Factor many years ago, he said these people don't have any experience as they were never on the road, never learned how it worked, here today gone tomorrow. ChatGPT is the new Simon Cowell
Yes, a tool is as good as the craftsman yielding it. Sounds like he was a beginner or junior programmer with not that much experience. Another thing to consider is feeding your code into a chatGPT assisted IDE. Many are finding it's ability to refactor, comment and streamline existing code to be an underappreciated functionality. It's all depends on how you structure the problem to be solved.