is it sensible to build a Trading System completely from scratch?

I’m learning QT along with C++ right now as a part of my Computer Science Major.

I was wondering if it is sensible to build a trading system completely from scratch or there are open-source trading systems available for free which I can develop to my needs. if yes, could you name some open-source trading systems? C++ or Python.

1 Like

I cant help with the coding part but i can give you my experience starting from scratch.

I have been developing my own system for the last few months (Manually).

Id say the biggest advantage of this is understanding why something works. Each month I get a deeper understanding of what I want in a system. This means constantly going back to the start to improve though.

Id say if you are not in a rush try developing your own system. Then once you understand what makes a good system you can go out and try other peoples.

If you develop a open source system then you could also learn these skills but it might take longer.


In short, yes

“Trading System” could mean a few different things. Can you tell us a little more about what you’re envisioning?

Either way is good as long as you profit. There are hundreds of free source codes in Codebase in MQL5 site. Check it out.

If you build strategy in C++ or python, how you will implement this code to MT4 or 5 terminal? btw, can tell us more about you? Regards Greg

1 Like

Of course, it is. You may not get all that you want from a readymade trading system. In such a situation, a trading system that is built from scratch on the basis of your trading requirements can do wonders. You just have to be sure about your trading requirements so that you don’t end up messing it all.

1 Like

It is definitely sensible given the fact that you can inculcate as many features as you want in it. But here’s a catch. You really gotta be a pro at programming to do this. It is not a layman’s job.

Of course, it is. But if you are just starting out, you can use the systems that are already there in the market. Gain some experience before building one for yourself.
From my personal experience, I can say that python is better for building a trading system.

coding doesn’t help much in constructing a trading system,i mean ,a profitable one.

Coding is a tool, knowledge about building and testing is more important.

Absolutely. What’s better than making your own system entirely dedicated to your needs. I believe that you already have some good programming skills or at least you are confident that you can learn them easily, that’s why you’ve made your mind for it. I would still suggest you to try out an open-source system. You will get a better idea how to achieve your goal.

you dont need a gui to write a trading bot… they run from the shell just fine… unless you are using metatrader, in which case your bot will run inside of the metatrader platform…

btw, the programming language that is built into metatrader is almost identical to straight C language and it has trading functions built into it that allow you to manage your trades… it is a very nice programming language.

when you use the phrase “from scratch”, that can mean a couple of things to traders…

a trading bot simply automates the process of manually looking at an indicator(such as macd or rsi) and then deciding when to open and close trades.

the really difficult part of creating a trading bot from scratch is when you set out to design your own trading indicator as an alternative to using macd or rsi, etc…

the great majority of people simply do not have the moxy to design their own indicator, and therefore are stuck using the basic indicators that are supplied with their trading terminals.

hint: the basic indicators DO NOT WORK.

all of the trading bots that are available for sale on the internet are built using the basic indicators and therefore they do not work and are a waste of your money.

good luck.

Do you have proof of that? Or standard indicators do not work for you, and you are trying to convince everyone that the standard indicator did not work. I am building strategies from standard indicators, and they are profitable. Good luck with looking for the own path to profitable trading but do not assume that something did not work because, did not work for you. Regards Greg.

please delete my forum account, so that you can go back to doing whatever it is that you do here.

thank you.

"if any combination of these basic indicators could be used reliably, you would know what that recipe is…

so, do tell me, which indicators with which parameters works reliably?

also, if you notice in my original posts here in the introductions section… two users came at me asking how my code works…

well, if there was a well known combination of the basic indicators that actually works reliably, they dont need to know what i am doing… but they do want to know… because trading is almost impossible…

also, with all of the mentors running around in here… why would you need them if the basic indicators worked reliably?

and out of all of the indicators that are for sale thru my metatrader application, some for as much as $30k, i dont see not one single positive feedback review."

One of my strategies uses the Keltner channel in standard settings, another MACD, another Bollinger Bands, even comparing prices, you can build a strategy that trades long-term profitably. Need C ++ or python to trade, your choice. If you want to delete the account, write to the administration, I am only a user. Regards Greg

1 Like

Usually by automated “trade system” we mean expert advisor written in MQL4/5 (C#ish language for Metatrader). It has the advantage of whole backtesting and live trading functionality out of the box.

There is nothing wrong to write everything from scratch, but you will need to write at least strategy AND trade execution with your broker (some of brokers expose their API even for free).

My idea to leverage the above is to create strategies in MQL4/5, backtest within Metatrader, identify profitable ones and then rewrite in python to run 24h from eg. Raspberry Pi lying on the shelf :slight_smile:

keltner channel is built using moving averages.

moving averages do not predict the future.

building a trading system from scratch means that you must design a trading indicator that DOES NOT use moving averages.

moving averages work fine in a ranging market, but will fail during a black swan event.

We finally got to the essence of your problem, you want to predict the future while I have a statistical advantage in the long run. That’s why the standard indicators are effective for me and not for you. There is a long way to go if you want to become a profitable trader.

I think indicators, standard or otherwise, only make sense when traded with context.
Blindly taking any signal is likely to yield, at best, break-even, at worst a slow account blow-out.

Taking indicator signals contextually, such as Stochastics OB/OS (reversion) only when the market appears to be in a range, and MA type signals, only when the market appears to be in trend, make sense.

I also think that when the market appears to be in a tight range, a break-out type style makes more sense, and any MA related signals should be ignored.

Market context needs to be defined before deciding if a signal is valid or not.
This is my intuition, as I learn this business.