Hello everyone, my name is Theo, developer by profession, passionate about Machine Learning (a form of Artificial Intelligence) and present on the crypto scene since late 2012. I got interested in Bitcoin early on, without imagining the size and variety of the ecosystem we know today.
Mid-2019 with, among others, my first autonomous trading bot on Nano (XNO). Then more recently, with the creation of several bots on more varied assets.
Like many, trying my hand at trading has been a bumpy ride for me. With big gains and equally big losses. Working on one’s own portfolio has the advantage of making the trader evolve quickly.
In my case, the realisation was clear: the biggest unknown was the trader himself. So I tried to remove myself from the equation as much as possible by turning to automation.
Binance Futures strategy
So my strategy differs from other traders here in its systematic approach, using an algorithm. Positions are re-evaluated every day, based on simple signals, but read by a computer model.
So I don’t make the decisions directly.
This model has been trained on all available data from 2017 to 2020, and tested on 2021. It does not know the prices and yet it is able to position itself correctly, to remain neutral, and even to short regularly.
The goal is not to win every time, but to avoid losses at all costs. No reading of graphs every morning either.
The goal here is to build a solid model that makes (statistically) the right choice in the long run. I don’t interfere with decisions, I don’t cut positions if they fall (except in extreme cases at the request of the risk manager of course). My job will be to generate quality data, then (re-)train the model, and adapt its parameters to optimise profits/drawdown/etc.
Performance over 2021 ( in progress, as at 5 December)
The performance so far has been very satisfactory, with a growth that could be described as consistent given the volatility of the crypto. There have been a few exceptional months, particularly May 2021 (+63%), as it has managed to go Short on several occasions during the fall and Neutral the rest of the time. A few empty months too (-10% in September), as there was a lot of price movement, but no obvious trend.
The pattern remains the best performer during clear trend periods. I’m looking forward to a bullish trend again. In the meantime, it remains on a maximum drawdown of -19% for the year, and an average monthly performance of +17%.