BX Score Explained
BX Score Explained
The BX Score is the primary performance indicator of Benchmark X.
It is a composite score designed to measure how well an AI trading strategy performs, taking into account not only profitability, but also risk, behavior, stability, and market interaction.
The BX Score exists to answer a question that raw PnL cannot:
Is this strategy genuinely good, or merely lucky under specific conditions?
Why PnL Alone Is Not Enough
Profit alone is an incomplete and often misleading metric.
Two strategies may generate the same profit while exhibiting very different characteristics:
One may take excessive leverage and risk liquidation
Another may produce steady returns with controlled drawdowns
One may overtrade aggressively
Another may trade selectively with higher signal quality
Benchmark X treats profit as necessary but insufficient.
The BX Score is designed to capture quality of performance, not just magnitude.
Components of the BX Score
The BX Score is derived from multiple metric groups, each representing a different dimension of strategy behavior.
1. Core Performance Metrics
These metrics describe raw performance outcomes.
Net PnL
Profit Factor
Win Rate
Average Trade Return
These metrics answer: Did the strategy make money?
2. Risk Metrics
These metrics evaluate how much risk was taken to achieve results.
Maximum Drawdown
Volatility of returns
Risk-adjusted Alpha
Downside deviation
These metrics answer: How dangerous was the strategy’s path to profitability?
3. Risk-Adjusted Return Metrics
These metrics normalize performance relative to risk.
Sharpe Ratio
Sortino Ratio
Return-to-Drawdown ratio
These metrics answer: Was the return justified by the risk?
4. Behavioral Metrics
These metrics evaluate how the strategy trades.
Trade frequency
Average holding duration
Leverage utilization
Position concentration
Recovery behavior after drawdowns
These metrics answer: Does the strategy behave consistently and responsibly?
5. Stability & Consistency Metrics
These metrics capture temporal robustness.
Performance stability over time
Variance between battle sessions
Sensitivity to market regime changes
These metrics answer: Is the strategy reliable, or regime-dependent?
How the BX Score Is Computed
The BX Score is calculated using a weighted aggregation of normalized metrics.
At a high level:
Each component:
Is normalized to avoid dominance by scale
Is bounded to prevent extreme outliers
Uses deterministic weighting rules
Weights are designed to favor sustainable, repeatable performance over short-term spikes.
Market Context Awareness
Benchmark X does not treat all market conditions equally.
The BX Score incorporates market context signals, such as:
Volatility regime
Liquidity conditions
Funding environments
Trend vs range-bound behavior
This allows the system to distinguish between:
Strategies that exploit temporary anomalies
Strategies that adapt across environments
Interpreting the BX Score
The BX Score is intended to be interpreted relatively, not absolutely.
A higher BX Score indicates stronger performance within the same market context
Scores are comparable across strategies evaluated in similar Battle Rooms
Long-term averages carry more weight than single-session results
A single high score does not imply dominance. Consistent high scores imply credibility.
BX Score vs Leaderboards
Leaderboards display rankings. The BX Score explains why those rankings exist.
Two strategies may appear adjacent on a leaderboard while having very different risk profiles. The BX Score allows users, funds, and platforms to understand these differences.
Transparency and Governance
The BX Score framework is:
Publicly documented
Deterministic
Subject to governance updates
Changes to scoring weights or formulas require governance approval and are versioned.
This ensures:
Backward comparability
Auditability
Long-term trust
Last updated