Would you collaborate with a programmer if it meant sharing your system/research?

[fan27]

The likely scenario is the programmer works on the project and loses interest, both in working on the project and in using the software and probably trading in general. The odds are low that he is going to take your software and commercialize it against your wishes. Most people would likely not be interested in your software. Would you be interested in a custom software solution built for someone else? I doubt it.

That being said, I bet he would stay interested long enough for you to get your new features free of charge.

 

[ET180]

Like we didn't have entire terabytes of personal information at our disposal from our jobs and we don't do anything with it.

I think in most cases, programmers can't actually use what they are working on at work for direct personal benefit. If I work at Microsoft and have full access to the Microsoft Word source code, it would be hard to take that and sell it as a competing product. Or if I develop automated market making trading software at Goldman and had full access to their proprietary trading software, it might be hard to use it without their infrastructure. Although there was a case, I think about a decade ago where someone ran off with some of their software and it was a concern.

My advice would be to learn to code. My guess would be a few months if you can dedicate a couple hours per day to it. C# is an easy language to pick up and the long term benefit is that you'll be able to maintain and experiment with your own code. When you have an idea, you'll be able to try it probably a lot faster than it would take you to explain it to someone else who is not familiar with your system.

Most people would likely not be interested in your software. Would you be interested in a custom software solution built for someone else? I doubt it.

If I knew that someone else used it profitably, I'd be interested. If I had the source code and/or knew how it worked, I would at least try it out.

 

[nooby_mcnoob]

Learn to code. You'll thank me later.

I second this.

Coding is a necessary skill for traders to level up.

 

[studentofthemarkets]

I wasn't really planning on doing anything more at the moment, but have now been approached by a programmer who's very eager to collaborate with me.

Being approached by someone would be a red flag for me. I wouldn't assume bad intentions but try to discern his intentions, as you are doing.

 

[newwurldmn]

If there’s no risk of your alpha decaying because he’s trading it, the risk is small. But what developer will take a trading strategy as compensation instead of cash.

No disrespect to you, but 90percent of strategies are crap and 90percent of Those that aren’t are likely to perform poorly under a different manager.

 

[themickey]

Actually, I'd be surprised.

I have done some basic coding in VBA, but I suspect that it would take me a long time to be able to produce a model similar to what I have today on my own. It needs to import data, update data, make calculations, store it in a DB and display it in a front end. I'd be very surprised if I'd be able to create this on my own without first making a huge investment in learning to program.

That said, I do want to learn to program. But time is at a premium these days and I'm not sure I'd be able to invest enough time to get anywhere fast.

Maybe, instead of paying a developer to code up your work, request you pay him to teach you how to code up your work.
In other words you pay him to tutor you to code just what you want.

Another thing, ask this person what their style of trading is, if they trade a totally different method or market, chances are they will not be interested in your method.

Plenty of people share their methods on ET, I've never hardly ever been interested as it's poles apart on how I prefer to trade.

 

[rb7]

Actually, I'd be surprised.

I have done some basic coding in VBA, but I suspect that it would take me a long time to be able to produce a model similar to what I have today on my own. It needs to import data, update data, make calculations, store it in a DB and display it in a front end. I'd be very surprised if I'd be able to create this on my own without first making a huge investment in learning to program.

That said, I do want to learn to program. But time is at a premium these days and I'm not sure I'd be able to invest enough time to get anywhere fast.

Probably that 95% of your stuff is general data manipulation.
Any developer could do that without knowing the 5% of you system that you don't want to disclose and you could code yourself that last 5%.

 

[SimpleMeLike]

Hello all.

Long story short I have a custom trading system/program written in C# which was built to me by a professional firm over 10 years ago by now. Since the first iteration I've had subsequent upgrades up until the point where I'm mostly satisfied as it covers most bases. This is not an automated trading system, but what I call a statistical model.

The main problem at the moment with the software is that the program outgrew its original specifications (which was rather simple), such that particularly updating of data isn't as smooth as it should be considering that I use it intraday as well. Also, I've had to add some of my own calculations later on in Excel. However, it does work, although not optimally.

I wasn't really planning on doing anything more at the moment, but have now been approached by a programmer who's very eager to collaborate with me. Roughly, the deal will be that he re-writes my current model/system taking care of all present issues and adds a few new features. In return, I pretty much give him everything which I spent so many years developing and researching.

I have two concerns:

1) How can I be certain that this guy doesn't leave me in the dust as soon as I've divulged the current source code/calculations?

I know it's possible to create contracts, but I'm still a bit worried here.

2) Simply the act of sharing everything I accumulated is something that makes me conflicted. I spent a great deal of time, trial and error to deveop what I have. And then I'm simply going to hand it out to some stranger?

The obvious benefit to me is that I may be abe to finally complete my system and have at hand a much better and more efficient program without investing any money.

Additionally, there's the possible benefit of developing something even better together. If we continue working together he can quickly implement new ideas for testing, etc.

Any thoughts?

Hello Laissez Faire,

Good questions and concerns.

I have no opinion here. I have thought about hiring a programmer many times. But I simply do not trust them with my algo and trading ideas.

Therefore, I program everything myself. Yes, it takes awhile and it is alot of work. But just rather do myself. For things that is hard, I rather hire a programmer to do, but I do not send them all my code.

A programmer may take your algo and sell to the market or sale to a friend of a friend to sell to the market.

 

[Laissez Faire]

If there’s no risk of your alpha decaying because he’s trading it, the risk is small. But what developer will take a trading strategy as compensation instead of cash.

No disrespect to you, but 90percent of strategies are crap and 90percent of Those that aren’t are likely to perform poorly under a different manager.

It's not a trading strategy per se, but a system or way of analyzing/understanding the market as it is through all market conditions. I call it a predictive/statistical model.

But since it's only a tool it and not an automated strategy it does require a vast amount of experience and knowledge in order to know how to exploit it. I spent a great deal of time developing it and all the pieces involved, so I know it's not 'crap' since it's already working.

I'm on holiday at the moment, so will respond to the rest of the good comments later on, but let's just say I've become more skeptical towards a collaboration, although it could be a fast-track for me to complete and further refine what I already have. Hence why I was tempted to move forward with it.

 

[NoahA]

It's not a trading strategy per se, but a system or way of analyzing/understanding the market as it is through all market conditions. I call it a predictive/statistical model.

But since it's only a tool it and not an automated strategy it does require a vast amount of experience and knowledge in order to know how to exploit it. I spent a great deal of time developing it and all the pieces involved, so I know it's not 'crap' since it's already working.

I'm on holiday at the moment, so will respond to the rest of the good comments later on, but let's just say I've become more skeptical towards a collaboration, although it could be a fast-track for me to complete and further refine what I already have. Hence why I was tempted to move forward with it.

I'm surprised that nobody has yet asked the simple question about win rate or profit expectancy of the system. But this post now answers that question.

So what exactly would the programmer get if he does all the coding and understands the significance of it? Is it simply that the model states that there is an 80% chance of the market finishing higher today than yesterday? Or that there is only a 20% chance that the market won't test the lows of yesterday? If this is the type of information the model produces, I'm not sure how valuable this would be to the programmer. You still after all have to enter the trade and know how to manage it. Just because something has an 80% chance of happening, if you bet 20% of your account on it, and it ends up being a loser, you're back in a bad position and fear will set in for the programmer.

And lets not forget that just because the model says an 80% chance of the market finishing higher means something totally different it if opens 50 points below the close yesterday, or 20 points higher on a gap up. Entering a trade at those two different locations is a totally different thing. It sounds like all the entry criteria is in your head, in which case, the model is only really useful in your hands.

My hunch is that your statistical model helps you trade because you have faith in your reads of the markets and you know what a good entry is. Without thousands of hours of screen time, just because a model says there is an 80% chance of the market finishing higher doesn't mean many will be able to make money with it. How often do we have a mid-day plunge that ends up getting bought up? That plunge could every well destroy a trader who just assumed the model stated it would be a straight up day.

Since the model isn't spitting out entries and exists but just probabilities of the upcoming day and probabilities of levels being hit, its not that valuable to a programmer and I would hire someone to do the job. If its gonna help you make a ton load of money, and you say it will only take the programmer a couple of weeks, in a month you should be raking it in. If you aren't making a killing a month after the project finished, then its clearly going to be of limited use to the programmer anyway.

Perhaps if you share more precisely what type of into the model generates, we can all better understand how useful it would be to a non-trader. I can tell you, much of what I read here at ET is totally useless because of the way I read the market. Unless someone says "I just went short with a 10 point stop", what value is there in knowing that we are approaching some zone, or some retracement level, or that some indicator is flashing red, or that Wave 5 just completed, or .....