🤓 Latest Submissions
trdrbot - a sELF improving options trading agent
trdrbot senses markets and news, forms a falsifiable thesis, simulates several option structures expressing it, sizes the winner using Kelly sizing and its own earned track record, executes and manages it through Alpaca, then scores afterwards whether its view was right, its structure was right, or it just got lucky. It has a number of innovative concepts like having a wiki knowledge base of long term information and patterns that it wishes to remember. It also uses a short term immediate memory for agents that that evolves like human memory called elfmem. This helps the agent make decisions and learn in a similar way to humans. This agent goes through a number of steps. First it senses the environment around it and gets the latest news pricing information and knowledge stored in its own wiki. It updates any regime documents or ads and improves company related information. It then uses a Muse engine, to come up with potential thesis. These theses are evaluated and learning is applied to our selection criteria to reason and evaluate to find the best opportunity for an options trade. trdrbot will then use deterministic approaches to size Are trade and spread, and it will apply the appropriate risk appetite. It will then use the alpaca MCP to place the trade or spread. It has knowledge on different types of trades and techniques and will use the appropriate ones accordingly. The system then continually schedules a wake up and will check the market every five minutes during market hours and update the trades if If they pass certain thresholds. Outside of trading hours, the schedule downgrades to thirty minute gaps, and on weekends this is hour gaps. Finally, there is a feedback loop after the trade that evaluates the effectiveness of the thesis and the prompts against the overarching goals, and it adjusts and improves these accordingly. This way, the system continually sELF-improves.
4 Sep 2026
