Adaptive Market State Graphs for Signals, Randomness Tests, and Trade Exits
Summary
This article describes an Expert Advisor that represents market conditions as vertices in a state-transition graph. It discretizes normalized trend, volatility, and momentum measures, counts transitions between states, smooths and decays those counts, then uses a k-step random walk to estimate directional bias. Entropy provides a confidence measure that can suppress weak forecasts. A second graph tracks how trades progress through R-multiple milestones and estimates the probability of reaching the next one, which can inform exits.
The method also compares the clustering of a thresholded market graph with Erdos–Renyi random graphs of similar size and edge density. This z-score is offered as an optional filter, not proof of predictive power. The article reports a backtest in which the randomness gate did not improve performance and the outcome-based exit rule underperformed a simple fixed target. It emphasizes that thresholds remain fixed inputs, results require out-of-sample validation, and transition or survival probabilities alone do not establish positive expected value.
Key ideas
- Market states are formed from discretized, volatility-normalized trend, volatility, and momentum measures.
- A smoothed transition matrix supports k-step forecasts, while row entropy measures forecast uncertainty.
- A random graph comparison tests whether network clustering differs from a density-matched chance baseline.
- An outcome graph estimates trade survival across R-multiple milestones to guide position exits.
- The reported tests do not establish an edge, and the article calls for out-of-sample validation and expected-value analysis.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.