How Conviction Scoring Works
A transparent look at how QSA calculates S/A/B/C grades using domain-diversity analysis, category rarity multipliers, and sqrt scaling.
The Core Idea
Most scanners grade signals based on a single indicator's strength. QSA takes a fundamentally different approach: domain diversity is conviction.
When a coin triggers a momentum scanner, that's a signal. When it also triggers a smart money scanner, a structure scanner, and a regime scanner — that's conviction. Four different analytical domains independently reaching the same conclusion is far more meaningful than one domain screaming loudly.
The scoring engine rewards breadth over depth. One S-tier signal from a single category will never outscore three B-tier signals from three different categories. Diversity is the signal.
The 9-Step Scoring Pipeline
Scanner Base Score
Each scanner that fires on a coin contributes a base score based on its historical reliability. High-performing scanners like Momentum Ignition or Whale Tracker contribute 6-8 points. Simpler volatility or session scanners contribute 3-5 points.
Domain Diversity
The core of conviction scoring. Each unique scanner category that fires earns diminishing points: the first category adds 15 points, the second 12, third 9, and so on. Having signals across 4 diverse categories (e.g., momentum + smart money + structure + regime) is far more meaningful than 4 signals all in the momentum category.
Category Rarity Multipliers
Not all categories are equal. Rare, harder-to-trigger categories earn multipliers on their domain diversity points. A smart money signal contributing to domain diversity is worth 2.5x what a volatility signal contributes. This rewards signal quality over quantity.
Grade Bonus
The best internal grade from each category contributes bonus points. If a coin has 3 momentum scanners firing and the best one is grade A, that category adds 8 bonus points. Only the best grade per category counts — duplicates don't stack.
Regime Alignment
The current market regime (QUIET, NORMAL, CHAOS) affects scoring. Breakout scanners in QUIET regimes earn full points. Mean reversion in NORMAL earns full points. Misaligned regime-strategy combinations earn 0. CHAOS applies a penalty across all types.
Volume & Timeframe
Volume above the 20-period average adds up to 5 conviction points (2x+ volume = max). Higher timeframes contribute more: 4h and 1D signals add up to 5 extra points vs. 5m/15m signals which add 1-2 points.
Bias & Freshness
Alignment with BTC's daily trend adds up to 5 points (BTC bullish + LONG signal = full points). Signal freshness adds up to 3 points for signals detected in the current scan cycle, with diminishing points for older signals.
Bearish Subtraction
SHORT signals subtract from the total raw score. If a coin has multiple LONG scanners firing but also has SHORT scanners, the bearish signals reduce conviction. This prevents inflated scores when directional evidence is mixed.
Sqrt Scaling
The raw score is normalized using square root scaling. This compresses very high scores (preventing easy 100s) while spreading out lower scores for better tier differentiation. The result is clamped to 0-100.
Category Rarity Multipliers
These multipliers determine how much each scanner category contributes to the domain diversity score. Rare, harder-to-trigger categories are worth more.
Diminishing Domain Points
Each unique scanner category that fires on a coin earns diminishing points. The first unique category is worth 15 points, the second 12, and so on.
Total possible domain diversity: 15 + 12 + 9 + 6 + 3 + 1 = 46 raw points (before rarity multipliers)
Final Grade Assignment
Example: How BTC Gets an S Grade
Disclaimer: Conviction grades reflect pattern quality and cross-domain confirmation, not price prediction certainty. Even S-grade setups can and do fail. QSA is an analytical tool, not financial advice. Always do your own research.