Criesnyse Basics
Overview
Criesnyse is a platform where you create or subscribe to machine learning models that predict market conditions. We provide the technology to train, validate, and operate these models at scale.
We don't predict future prices or generate signals. We provide tools to create and run sophisticated classification models that evaluate current market conditions based on historical patterns.
The following table describes what Criesnyse is and what Criesnyse is not:
CRIESNYSE IS | CRIESNYSE IS NOT |
|---|---|
Cloud infrastructure for machine learning models - Criesnyse handles training, validation, and continuous execution of models on historical and live market data. Models generate daily predictions from purpose-built datasets, and we provide analytical features built on this prediction data. | A signal provider or trading bot service - Criesnyse does not send trading signals, provide rule-based bots, execute trades, or connect to brokerage accounts. We provide prediction data and analytics, not signals or automation. |
A closed data generation platform - Models operate independently and generate prediction data internally. You interpret and use this data for your own trading decisions. There are no social features, influencers, or community. | A copy trading or social trading service - You cannot follow other users' decisions or copy positions from external traders. Models work in isolation without social elements. |
"Build and forget" approach - We handle backtesting, deployment, and continuous operation. Any model created will run continuously - we take responsibility for keeping it operational. | A backtesting platform for professional quants - Unlike platforms focused on research and backtesting promising strategies, Criesnyse intentionally emphasizes deployment and continuous operation. Any model runs after deployment, regardless of performance. |
Data
Models are trained and operate on Criesnyse's proprietary dataset, specifically prepared for the platform. Training data spans multiple years of historical market data.
Live data is released on weekdays, several hours after US stock market close. Models generate predictions on this data early the following trading day.
This dataset is obfuscated - models trained on it are designed specifically for Criesnyse and cannot be used for trading outside the platform. Even if you own a model as an NFT and run it locally through our desktop application (planned feature), the model requires Criesnyse's data structure to function.
Models
A model is a machine learning algorithm trained to classify market conditions. Models use supervised learning - during training, you label historical data by assigning market states (Buy, Sell, etc.) to different time periods. The model then learns to recognize patterns in the data that correspond to these labeled states.
Every model goes through a structured lifecycle: training on historical data, validation to ensure it generalizes, freezing into an immutable binary file, backtesting to establish performance baseline, and deployment for continuous live operation.
The key characteristic of Criesnyse models is immutability after freezing. Unlike human traders who can change strategies, or adaptive algorithms that modify behavior, frozen models consistently apply the same learned patterns regardless of market conditions or recent performance.
FAQ
Do I need coding skills or algorithmic trading experience?
No. You don't need programming skills to use existing models. Creating models requires basic understanding of machine learning concepts (like "label", "feature", etc.), but the platform handles all computational complexity.
What is a "prediction"?
A prediction is the model's classification of current market conditions for a specific asset. It indicates whether the model classifies conditions as Buy, Sell, etc. based on patterns it learned during training.
Who creates the models?
Anyone can create models - regular users and Criesnyse team. All models go through the same training, validation, and freezing process regardless of who created them.
How many models can I create?
Unlimited. There are no restrictions on the number of models you can create.
What's the catch with unlimited model creation?
There is no catch. We intentionally allow any model to run continuously after deployment, regardless of performance. This "build and forget" approach is core to our platform design. Models that perform poorly simply become part of the data - they don't cost us differently than successful models.
What markets are available?
Currently US stocks, cryptocurrencies, and forex. Technically, the platform can support any financial market - we may add more markets in the future.