Five metrics, one score
Every live trade updates five independently-scored dimensions: return rate, maximum drawdown, win rate, return volatility, and trade frequency. Each is compared to your uploaded backtest baseline and normalized to a 0-100 component score. The overall health score is a weighted composite.
Weights are not democratic. Drawdown and volatility get the most weight because they are the first things to move when an EA starts breaking. Return rate matters but is noisy over short windows. Win rate and trade frequency round out the picture — a strategy trading half as often as its backtest is usually a sign something upstream changed.
Why composite beats single-metric monitoring
Any single metric can mislead. Profit factor looks fine during a losing streak if the few wins are large. Win rate looks fine while average loss grows. Drawdown is backward-looking. Individually, each can be gamed or masked. Together, they move in correlated ways when an EA is genuinely breaking — and diverge in ways that reveal specific failure modes.
A composite also reduces false positives. Any single metric may breach threshold for normal variance reasons. Requiring several components to degrade in concert is a much stronger signal than any one alarm.
Reading the score
A healthy strategy sits at 75-100. Scores between 50-75 indicate the strategy is trading within an acceptable but wider-than-expected envelope — worth a look, not an alarm. Below 50, something specific has gone wrong: compare the component breakdown to see whether it is drawdown, trade frequency, or volatility that is driving the degradation. The component view almost always points at the root cause faster than inspecting raw trades.