This is the list of weights of agents that has been trained for specific trading strategies.
| Name | Agent | Timeframe | Decision Period | Leverage | Description | Release At |
|---|---|---|---|---|---|---|
| v1 | Test Future Agent | 1h | 60min | 2x | Just a test weight for future agent. | 2025-02-27T05:22:45.028000+00:00 |
| v1 | Test Spot Agent | 1h | 60min | 1x | Just a test weight for spot agent. | 2025-02-27T05:22:45.052000+00:00 |
| v1 | Statera 2025.03 | 1h | 60min | 1x | Trained with BTCUSDT data from 2020 to 2024. The reward function considers the change in portfolio value, maximum drawdown of current trades, penalties for holding trades too long, and overtrading. Reward per step is the main indicator for training. Rewards per step are 0.015 and 0.005 for the verification and testing datasets, respectively. This agent is not ideal for real trading but is profitable in backtesting scenarios. | 2025-02-27T05:22:45.055000+00:00 |