Using Functors and Monoids to Set Position Size from Projected Stop Loss

Summary

The article applies category theory concepts to an MQL5 money-management example. It treats monoids as rule-based selectors for choices such as timeframe, lookback period, applied price, and indicator. Their selected values, together with a normalized indicator reading, become inputs to a multilayer perceptron. The network is trained to estimate a stop-loss gap, which is then converted into trade volume using tick size, tick value, free margin, and a user-defined risk allocation. A margin check and maximum-lot fallback are also described.

The example uses BTCUSD and an RSI-based entry and exit system, with Bollinger-band readings also discussed as an input option. The author frames the approach as experimental and notes that preliminary EA performance is highly sensitive to position sizing; the text also points to remaining hurdles before practical use. It does not establish that the neural estimate improves returns or risk-adjusted performance, and the excerpt gives no complete comparative results. The method should be treated as a design exploration rather than a validated sizing model.

Key ideas

• Monoids select candidate trading inputs such as timeframe, lookback, and applied price.
• A multilayer perceptron maps selected inputs and an indicator reading to a projected stop-loss gap.
• The projected gap is converted into lot size using tick economics and allocated free margin.
• The example includes a margin check with a maximum-lot fallback.
• The proposed sizing approach is experimental and lacks reported evidence of improved trading performance.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.