About & methodology

About CryptoAlmanach

CryptoAlmanach is an independent research project, developed by an artificial intelligence engineer specialized in quantitative analysis, complex systems, and responsible AI.

The site relies on a proprietary analysis engine, TokenAlmanach™, designed to explore how advanced AI models can help analyze crypto and equity markets while remaining transparent about their limitations.

What types of data are used?

  • Public market data: prices, volumes, volatility, aggregated order books when available;
  • Classic technical indicators (trends, support/resistance, volatility, etc.);
  • Statistical and stochastic signals inspired by physics (Markov processes, volatility regimes, etc.).

How does the AI score work?

For each eligible asset, an AI engine evaluates the quality of the risk / opportunity profile over a short- to mid-term horizon. The AI score is an internal indicator that notably takes into account:

Methodology
The AI score is an internal risk / opportunity indicator based on quantitative signals (trend, volatility, liquidity, signal stability) and risk filters. The algorithms are tuned and tested on historical data, and may evolve as new data becomes available and models are improved. — Engine methodology: TokenAlmanach™ AI Engine
  • recent trend and volatility dynamics;
  • the minimum liquidity required for responsible trading;
  • signal stability over time (avoiding one-off “lucky hits”);
  • risk filters to exclude assets that are too illiquid or abnormal.

Algorithms are regularly tuned and tested on historical data. The methods used include stochastic modeling, neural networks, and Bayesian optimization, with strong inspiration from statistical physics.

Important limitations

  • Scores may change at any time as new data arrives;
  • Models do not take into account users’ personal situations (wealth, taxation, risk tolerance, etc.);
  • Crypto and equity markets are volatile and may behave unpredictably, including after macroeconomic or geopolitical events.

For more details about liability, personal data, and cookie management, please refer to the Charter & legal notice page.