Portfolio composed by stocks with high liquidity.
Low number of transactions: 400 per year (avg.)
Stop loss and profit taking defined.
High profitability: average of 30% per year.
CVaR95 = 1.4%
Low market correlation: β=0,49
How It Works
Our algorithm identifies exact moments for stock purchase and sale migrating resources from one stock to the other within defined portfolio.
The average time we spend on each stock
‣ Machine Learning
‣ Artificial Neural Networks
‣ Decision Trees
The technology behind the robot is based on Machine Learning models such as Artificial Neural Networks and Decision Trees that are trained with historical price data and macroeconomic variables to identify reversion or trend continuity triggers.
The problem of learning from historical data is configured as a pattern classifier system that can anticipate moments of trend reversal.
Smart investment technology
We combine indicators of technical analysis that reinforce the analysis of the robot, ensuring more robustness to its predictions.The methodology is generalist so that the information used for decision making allows the application of the robot in several markets, such as stocks, currencies and other assets. There are also intelligent risk management mechanisms that contribute to even better results.
Among these mechanisms, there are adaptive stop-loss and take-profit methods that are adjusted regularly according to the dynamics of prices.We also do automatic portfolio selection based on consistent risk metrics, such as CVaR.
Finally, our robot uses a form of trading that maximizes the exposure and distribution of capital within the portfolio
Unlike many other robots currently developed, we do not work with High Frequency Trading (HFT), but rather with medium-term transactions (Swing Trade), as shown above. In this way, we avoid some problems inherent in HFT, such as the need for extremely low response and communication times as well as high transaction costs.