PROphet: Dual Linear Perceptrons with Separate Buy and Sell Optimization

Summary

PROphet is an Expert Advisor built around two independent linear perceptrons that classify candlestick-shift inputs. One model distinguishes buy signals from flat or sell conditions; the other distinguishes sell signals from flat or buy conditions. The stated design keeps long and short classification separate instead of asking one perceptron to learn both directional decisions. A movable stop loss is also included in the optimization parameters.

The prescribed workflow optimizes the buy-side weights and stop separately from the sell-side weights and stop, using recent data on an M5 timeframe. Both directions are then enabled, and the resulting settings are intended for the following week. The document gives procedural parameter ranges and a weekly re-optimization schedule, but supplies no performance results, validation method, transaction-cost assumptions, or comparison against simpler rules. Its short optimization window may also make settings sensitive to recent market conditions, so the proposed procedure alone does not establish robustness.

Key ideas

• The Expert Advisor uses two separate linear perceptrons to classify candlestick-shift features.
• One perceptron handles buy classification, while the other handles sell classification.
• Buy weights and stop settings are optimized separately from sell weights and stop settings.
• Optimization is specified for recent M5 data, with parameters intended for use in the next week.
• No performance evidence or robustness analysis is provided.

Tags

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