Data intelligence platform for investors
Our predictive models are backtested over independent historical periods before generating a single recommendation. The result is a traceable signal, not a market intuition.
Technical pillars
The platform does not replace the investor's judgment; reduces the noise of large volumes of data so that this criterion is applied to already refined information.
The models process prices, volume and connected macroeconomic variables to identify recurring patterns before issuing a signal. Each signal is linked to the variables that originated it.
Each recommendation includes an exposure range and a loss threshold derived from the historical volatility of the asset, so that the risk assumed is defined before acting.
The infrastructure incorporates new data and recalculates the suggested portfolios incrementally, without the need to reprocess the entire history with each update.
How we work
CONTPAQi combines data engineering and statistical modeling applied to financial markets. The goal is not to predict with absolute certainty, but to quantify the probability and risk of each scenario before the investor decides.
Each module of the platform documents its assumptions: the data period used, the variables considered and the limits of validity of the model. This documentation is delivered together with each recommendation, not separately.
Know the complete approachValidation framework
The rigor of backtesting determines the reliability of any predictive model. That is why each strategy goes through three different phases before reaching the end user.
Connection with market sources, corporate reports and structured and unstructured macroeconomic data. The data is normalized and cleaned of duplicates before entering the modeling engine.
Each strategy is backtested on historical periods independent of the training set, in order to reduce the risk of overfitting to past data.
Signals that pass the validation phase are translated into allocation recommendations with defined risk parameters, ready for review before application.
What the panel shows
The specific values of each metric vary depending on the asset, time horizon and investment mandate, and are detailed in the technical report delivered during the demo.
Calculated over the selected backtesting period, documented by strategy and time horizon.
Average interval between the arrival of new market data and the update of the current recommendation.
Visualized using standard industry ratios, such as Sharpe and Sortino, on the simulated history.
Number of markets and asset classes included, configurable according to the defined investment mandate.
Usage scenarios
The platform applies to both capital allocation decisions and internal operational management issues. Both scenarios share the same requirement: historical evidence before committing resources.
A manager evaluating the incorporation of crypto assets into a diversified portfolio needs more than a qualitative thesis. The system allows a proposed strategy to be backtested over different market cycles before allocating real capital, and documents the strategy's behavior in periods of steep declines, not just in uptrends.
A corporate finance team uses predictive models to anticipate liquidity needs and adjust exposure to macroeconomic variables that affect its operating costs. The objective is not to speculate, but to reduce uncertainty in quarterly planning decisions.
Resolution of technical doubts
Before integrating any predictive analytics tool, it is reasonable to ask how the data is treated and how the model can be audited.
The data is processed in accordance with the General Data Protection Regulation (GDPR). Access to each client's information is segmented and encrypted both in transit and at rest, and only the records necessary for generating and auditing recommendations are kept.
Yes. Each signal includes the variables that originated it, the backtesting period used and the risk parameters applied. This algorithmic traceability is part of the technical report delivered along with each recommendation.
The connection is made through programming interfaces compatible with market data sources and internal financial management systems. The exact scope of the integration is defined during the technical consulting phase, based on the client's existing infrastructure.
Models are continuously monitored against their expected historical performance. When a relevant deviation is detected, the strategy is flagged for review and notified before new recommendations are generated based on it.
A technical demo allows you to review the behavior of the models on real historical data before making any integration decisions.
