SHADOW Q-Score: how FreedomCore ranks Hyperliquid wallets by realised account outcome

SHADOW Q-Score: how FreedomCore ranks Hyperliquid wallets by realised account outcome

Most wallet leaderboards on perp DEXes show one number. Profit and loss for some recent window. Sort descending. Done. This is the easy answer and it produces a brittle ranking that flips constantly with luck.

FreedomCore SHADOW takes a different approach. The Q-Score is a six-component composite that ranks Hyperliquid wallets by quality of realised outcomes, not by raw PnL. The score is the foundation of the tiering system that decides which wallets SHADOW actually copy-trades.

The six components

The Q-Score combines six measurements that each capture a different dimension of trading quality.

Wilson lower bound on win rate. Raw win rate is misleading on small sample sizes. A wallet that wins 3 out of 4 trades looks like a 75 percent win rate, but with only 4 trades the confidence interval is enormous. The Wilson interval gives a statistically honest lower bound that scales with sample size. A wallet with 30 wins on 40 trades will outrank one with 3 wins on 4 trades even though the raw rate is the same.

Sharpe ratio over a rolling 30-day window. Risk-adjusted return matters more than absolute return. A wallet that compounds at 5 percent monthly with low volatility is worth more than one that doubles and halves randomly. Sharpe captures this directly.

Realised PnL on the SHADOW account. Not the wallet's own claimed PnL on its own books. The actual outcome on SHADOW's account when it copied that wallet's trades. This is the only PnL number that survives audit, because it is grounded in trades SHADOW actually placed.

Average holding time. Wallets that scalp in seconds and wallets that swing over weeks live in different worlds. SHADOW's copy infrastructure handles holds between roughly ten minutes and three days well. Anything outside that range gets penalised in the score because it is harder to copy reliably.

Trade frequency. A wallet that trades once a month cannot meaningfully drive a copy-trading account. The score requires a minimum thirty trades in the last thirty days to be eligible at all, and the per-trade frequency above that floor adds a modest signal-quality kicker.

Volume profile consistency. A wallet that switches between micro-size and whale-size positions randomly is harder to follow than one that maintains consistent sizing. The score rewards consistency.

Tier comes from account quality, not leaderboard rank

This is the key insight. Many copy-trading services tier wallets by their leaderboard rank on the exchange. Top 100 gets premium. Top 1000 gets standard. And so on.

SHADOW does the opposite. Tier is determined by the SHADOW account's realised result when copying that wallet over a seven-day evidence window. A wallet that is rank 50 on the Hyperliquid leaderboard but whose trades produce negative outcomes on the SHADOW account gets demoted regardless of its public reputation. A wallet that is rank 800 but whose trades produce consistent positive outcomes on the SHADOW account gets promoted.

The seven-day evidence window is short enough to catch behavioural shifts quickly and long enough to filter out single-trade luck. Every wallet tier change is logged in the wallet_tier_actions table with the score components that triggered it. The audit trail is complete and visible to operators.

Why this matters for public proof

The ARENA leaderboard at arena.freedomcore.io shows the result of this scoring publicly. Each wallet page lists its current tier, the components of its Q-Score, the seven-day account evidence, and a frozen snapshot of stats from the last profile run. The page exists whether the wallet is currently being copied or has been demoted to archive.

This matters because public proof requires inspectable methodology. Anyone reading an ARENA wallet page can see exactly how the wallet was ranked, what the score components are, and whether the wallet is currently active or archived. There is no hidden ranking that varies by who is looking.

Where the score evolves

The Q-Score is not static. Three known limitations are being worked on.

The first is venue diversity. Hyperliquid is currently the only data source. Adding other perp DEXes would broaden the wallet universe.

The second is the seven-day evidence window. Shorter is more responsive but more noisy. Longer is more stable but slower to react to behavioural shifts. Whether seven days is the right number remains an open question.

The third is the weighting between the six components. The current weights are reasonable but not formally optimised. A future version may re-tune these based on accumulated SHADOW account evidence.

These limitations are documented in the SHADOW operations manual rather than hidden behind marketing copy. The point of this note is to expose the methodology so that the result on ARENA can be evaluated against the method that produced it.