A 0-100 combined score that measures whether a live strategy’s edge is still holding or quietly decaying over time.
Health Score is a 0-100 combined score Quantprove calculates from your live trade sequence to measure whether its edge is still intact or is getting worse. Quantprove tracks two components, Rolling EV and Consistency, then compares early performance against your recent performance to catch decay the moment it starts.
A backtest tells you the past, a validation check tells you the present, but neither tells you what will happen next.
Edges go down… markets shift regime… a setup gets crowded… volatility changes… and a strategy that printed money for two years stops working.
The trader almost always finds out too late, after the account is already deep in drawdown, because they were watching their P&L instead of watching their edge.
Health Score watches the edge. You feed in your live trades as they happen, and Quantprove tracks whether the strategy’s expected value is still strong and still steady, or whether it has started to slip below the baseline it set early on.
In simpler words: it turns “I think something feels off lately” into a number you can act on.
Health Score is built from two weighted components, where one tells you how big the edge is and the other tells you whether you can trust it, or whether it relies on isolated, aggressive moves to print. The compound of the two elements, run through our mathematical calculation, gives the Health Score.
Health Score weights size higher because an edge that is large but a little choppy is still tradeable, while an edge that is steady but barely positive is not worth the risk.
Health Score uses an adaptive rolling window sized to your trade count:
It needs a minimum of 100 trades to run, because a handful of recent trades cannot separate real decay from normal behavior.
You can use this kind of rolling window for other metrics in your own analysis if you wish, a very useful concept.
Now… this is the magic behind it, and where it gets a bit technical.
To flag decay, Health Score compares the average of the first three windows against the average of the last three. If recent performance has dropped meaningfully below the first baseline, the edge is fading and the score reflects it. This is the same principle behind why a true edge should persist out of sample: performance that erodes over time points to overfitting or a regime the strategy was never built for, a failure mode Marcos López de Prado documents at length in Advances in Financial Machine Learning (2018).
Health Score is the core output of Mode 3 (Monitor).
It is where a strategy goes once it has earned a strong Edge Score in Mode 1 (Backtest) and a strong Stability Score in Mode 2 (Validation). From there, Mode 3 tracks every new trade against the baseline and tells you the moment the edge starts to fade, while you can still cut size or pull the strategy.
Edge Score grades the backtest. Stability Score checks whether live results match it. Health Score watches the live edge over time and tells you when it starts to break. The scores guide shows how to read all three together.