Adapting Trading Strategy Parameters with Daily Feedback
Summary
This brief discussion proposes a feedback loop for a short-horizon stock strategy: observe the result of a trade, such as a buy today and sell tomorrow, then use the realized daily return to adjust strategy parameters. The intended outcome is a model that updates as market conditions change. The response names dynamic model updating and deep reinforcement learning as possible approaches and points readers toward two papers on reinforcement learning for quantitative trading.
The post is conceptual rather than an implementation guide. It does not specify the state representation, reward function, update schedule, constraints, or safeguards needed to prevent unstable changes. Nor does it provide trading results or compare reinforcement learning with simpler adaptive methods. Daily returns can be noisy, and repeatedly tuning a strategy on recent outcomes risks overfitting. Any feedback system would need careful out-of-sample evaluation and controls for transaction costs and risk before deployment.
Key ideas
- A strategy can use realized trade returns as feedback for updating its parameters.
- The example considers a stock strategy that buys one day and sells the next.
- Dynamic model updating and deep reinforcement learning are suggested as approaches.
- The discussion gives no implementation details or evidence of trading performance.
- Frequent updates require safeguards against noisy feedback and overfitting.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.