One number, 0–100, for a single question: is this trading code correct? Not whether it's profitable — correctness only. The same deterministic method scores all 7 platforms, so a score means the same thing everywhere. This page is the whole method, in the open, so anyone can reproduce and cite it.
Every finding is one of four severities. Each carries a fixed penalty.
| Severity | Penalty | Means |
|---|---|---|
| Critical | −35 | The code acts on information it could not have had, or won't run as intended — look-ahead bias, future indexing, same-bar fills. A backtest built on it is misleading. |
| High | −18 | A real correctness risk that can make live behaviour diverge from the backtest — intrabar repaint, unsynced higher-timeframe reads, missing stops on an order strategy. |
| Medium | −8 | A meaningful hygiene or robustness issue worth fixing — un-warmed indicators, swallowed exceptions, fragile version usage. |
| Info | −2 | A minor note or style signal that rarely changes results but is worth knowing. |
score = clamp( 100 − 35·critical − 18·high − 8·medium − 2·info , 0 , 100 )Start at 100, subtract each finding's penalty, floor at 0. It's deterministic — the same code always produces the same score, with no model, randomness or opinion in the number. Rules are conservative: a finding fires only when the pattern is unambiguous, so the score is honest, not alarmist.
The green Verified seal is reserved for a passing result: score ≥ 90, with zero critical and zero high issues. Anything below that is honestly labelled Checked · score X — the badge always shows the real number, so the seal stays meaningful. Every badge attests to a specific code fingerprint (SHA-256), so it proves this exact code was checked, and can be re-verified through the public API.
The identical model scores all seven, so a Strategy Health Score is comparable across platforms.
The Strategy Health Score™ measures code correctness, never profitability, expected return, or edge. A high score means the code does what it says without the common bugs that make a backtest lie — nothing about whether the strategy makes money. It is software and education, not financial advice.
Validate through the public API and mint verifiable badges from your own tools.
Strategy Health Score™ Standard v1.0 · code-correctness only · not financial advice · not a measure of profitability.