Seeking Feedback on Dynamic Stop-Loss Adjustment Algorithm (with Test Data)

Hi Kasran! Thank you for your insightful questions.

I ran the simulation on the test dataset, and here are the key results (I’ve also attached the results in graph form in my original post):

• Duration Tested: The dataset includes a total of 41 trades.
• Profit and Loss Distribution:
  • Win rate: 63.41% of the trades were profitable.
  • Loss rate: 36.59% of the trades resulted in losses.
• Payoff Ratio: The average payoff ratio is 0.54, meaning that for every unit of loss, the corresponding profit is 0.54 units. This suggests that losses are typically larger than profits in this simulation.

The algorithm is designed with a long-term focus on increasing and preserving capital. The idea is to dynamically adjust risk based on performance: during a winning streak, the risk-taking ability increases, allowing for greater leverage, while during a losing streak, the risk appetite decreases to protect the core capital.

There’s a “secret sauce” within the algorithm to fine-tune the exact rewards for a trader’s previous profitable trade and the penalties imposed for a previous loss. The core of the algorithm lies in finding the right balance between how much extra reward should be given for a win and how much penalty should be imposed for a loss. This precise adjustment ensures the right incentives for winning trades and risk mitigation during losing periods.