I Developed a Trading Bot: Here Is What It Earned Me
Mastering different additional income streams, I am facing one specific point all the time:
Whatever you do, whether that is a passive income or not, you need to master that area perfectly — or not to start it at all.
But, obviously, that is not always the solution — especially when you are combining several main occupations and possess your passive income stream as a small on-site project. Happily, we live in a period of automatization — and my strong opinion is that having at least a minor programming skill is obligatory to live a wealthy life in a fastly-moving digital world.
Last time, I posted an article about trading as a way of earning money easily. The thing is unless you know how to trade, this is a casino with absolutely no guarantees. And even mastering the art of analysis, short-stock trading, etc., you cannot guarantee the result and still waste tons of nerves.
The idea this time was super simple: all the trading on short stocks is based on a simple regression and reflexive analysis. Simply saying, you are taking the plot for the previous day / hour / 20 minutes, applying basic analysis rules, and forecasting the way how your plot will act within the next period of time. To do this manually, you need some good knowledge of analysis, forecasting, time-series, predictive analytics, and — what is most important — act extremely fast. Usually, when it is about binary currency plots, this is totally impossible.
But machines are available to do that much faster. All the algorithms were developed long before this article was created. Nothing new — just a Lego to build an analytics tool of. And most importantly — the computer has no emotions and will not throw away your deposit on nerves.
So, my plan was simple as always:
Connect exchange provider API — implement legacy data gathering — add several beautiful forecasting algorithms — make script place the bids and run trades for me
I will not be telling exactly, what code was I writing and what libraries connecting. There is a huge amount of cool stuff to be used and play around, and there is completely nothing complex. I have used Python, Matplotlib, several additional libraries for visualizing and data processing (e.g. PyPlot) — and one self-written time-series library.
All this staff has been connected to Quotex — but you can use any other exchange provider or trading service with API you want. Several analysis and trading algorithms have been implemented. The funny thing is that I have just taken the libraries supporting the algorithms, without even writing new algorithmic code. Lego, just as I’ve said.
All in all, the algorithms used were:
- Scalping
- Alligator Analysis
- Rate of Change Behavioral Analysis
- AROON Data Displayer
- RSI Indication
- Time-Series Forecasting
- Fractal Forecasting
- Auto- and Moving Regression Methods
All of these approaches are easily searchable — and you do not even need to know how they work in complex (though that is worth reading).
Finally, I threw 50 bucks into the system — and launched the trading algorithm a week ago. Here is what I got today.
On the first day, it threw away more than half a sum. I was shocked cause there was something like 10$ left. First, I wanted to stop the algorithm at all, but then decided to play hard and proceed to the show.
On the next day, the algorithm returned the whole sum. It is important to mention that there is no bidding limitation, meaning one day the algorithm can bid all the deposit money at once — and lose all — but that is exactly how regular traders sometimes work, huh?
For the next several days, algorithm work stabilized. I have added a simple improvement, which was defining a daily income goal, so the algorithm kept trading until the goal was reached. Additionally, my goal was to disable trading on weekends. The stock market is highly unstable these days, meaning you should not trade — because there is nothing to predict, the market turns into a real casino.
All-in-all, as of the 8th day of trading, I have ended up with 83$ on my account. Not too many? Probably. But I did literally nothing except running the algorithms and enhancing some parameters of the algorithm. How will it behave in the next week, two, a month? Will it lose the money or make me extremely rich? I am not sure. Such experiments should be possessed just as experiments — and that is the key to success.
By the way, there are some cool open-source projects you can take a look at if you do not create the trading bot on your own. Here are the ones I know:
Take a try — and good luck!