ProofOfEdge

Each strategy receives a single composite score from 0 to 100 and a letter grade, but that headline number is built from six component scores that each measure a different dimension of quality. A strategy that prints PnL by taking reckless risk will score very differently from one that earns it consistently and reproduces well for followers.

Edge quality

Realized PnL adjusted for how it was actually earned, after fees and funding. Returns that come from genuine edge are rewarded more than returns that come from one lucky outlier.

Risk quality

Drawdown behavior, leverage discipline, and liquidation history. Tight, controlled risk scores higher than large swings, even at the same headline return.

Copyability

How faithfully the strategy reproduces for the people copying it: fill slippage, divergence between the manager and follower results, and copy success rate. A strategy that is hard to follow accurately is penalized.

Consistency

Stability of returns across the sample window. Steady performance scores higher than the same total return delivered in a single spike.

Data confidence

The quality of the inputs behind the numbers: index freshness and oracle quality during the window. Low-confidence data lowers the score.

Volume quality

Whether there is enough genuine, non-suspicious activity to trust the sample at all. Thin or manipulated-looking volume caps how much the rest can be trusted.

Alongside the score, every strategy carries a status that tells you how much to trust it right now. Status and grade always trace back to indexed fills, so they move as the data does.

Provisional

The default for a new strategy. It has a score, but not yet enough verified history to be relied on. Treat provisional strategies as early signal, not proof.

Verified

The strategy has accumulated enough clean, indexed history to back its score. This is the status the verified badge reflects.

Degraded

Something in the data or risk picture has weakened, such as stale inputs or a risk flag. The score is shown with reduced confidence until it recovers.

Blocked

A hard penalty applies, such as a liquidation, a drawdown breach, or suspicious volume. A blocked strategy is not presented as copyable proof.

A read-only indexer reconstructs every fill and position from on-chain activity. Because Signalhouse only reads the chain, there is nothing a manager can self-report or edit. Before any reconstructed activity counts toward a score, a shadow-verify pass independently cross-checks it against the chain, so a flaw in the indexer cannot quietly inflate a strategy. Only activity that survives that cross-check feeds the score.