Methodology & Risk Factor Breakdown
A transparent, data-driven approach to Bitcoin risk assessment using five analytical pillars.
What the G-Score is
The GhostGauge G-Score is a daily composite from 0–100 where higher = higher market risk (more crowding, leverage, froth). It's informational context—not advice. GhostGauge publishes a daily UTC intraday production snapshot. Individual inputs can have different source vintages and update cadences; their freshness and scoring availability are shown separately.
How it's made (in one breath)
Each enabled factor uses its factor-specific production logic to produce a 0–100 score. Only factors classified fresh under the production source-cadence rules contribute to the composite; their versioned weights are normalized over the included set for that snapshot. A missing required factor is omitted and fails strict publication health. It is not replaced with a neutral score. Cycle and Spike adjustment mechanisms remain implemented but are disabled and contribute zero points. The current BTC price observation is a UTC intraday snapshot. Completed daily price history is maintained separately for production calculations that require historical closes.
The Five Pillars (SSOT)
Five Pillars of Risk
Our methodology evaluates Bitcoin risk across five analytical pillars, each contributing to the composite G-Score.
- Liquidity/Flows — 30%: Stablecoins (18%), ETF Flows (7.7%), Net Liquidity (4.3%). Captures "is money coming in or out?" breadth and participation.
- Momentum/Valuation — 30%: Trend & Valuation (BMSB 60%, Mayer 30%, RSI 10%). Anchors where price sits vs regime and long-trend stretch.
- Term Structure/Leverage — 20%: Futures/derivatives term structure and leverage conditions; stress here often precedes spot.
- Macro Overlay — 10%: Macro context relevant to Bitcoin (e.g., liquidity proxies, rates) summarized into a single overlay.
- Social/Attention — 10%: A light measure of narrative intensity and attention—supporting signal, not the star.
Adjustments (disabled)
Cycle and Spike adjustment mechanisms remain implemented but are disabled. They contribute zero points to the current score. Reactivation would require a separate versioned methodology decision.
- Cycle (Power-Law): Disabled — contributes 0 points.
- Spike (Volatility): Disabled — contributes 0 points.
Band Legend (how to read the number)
Band labels and recommendation text describe general market-risk context. The official Risk-Based DCA framework below is narrower: it changes only new monthly contribution size. It does not create an automatic sell or trim rule for Bitcoin already held.
BTC G-Score Calculation
How It Works
Weighted Average
Each enabled factor contributes according to the weights defined in the current production configuration.
Score Interpretation
The G-Score provides a normalized risk assessment that accounts for multiple market factors simultaneously.
Risk Bands
Risk bands provide context for interpreting G-Scores and help categorize market conditions.
Band labels and recommendation text describe general market-risk context. The official Risk-Based DCA framework below is narrower: it changes only new monthly contribution size. It does not create an automatic sell or trim rule for Bitcoin already held.
To see how today's score maps to the official monthly Risk-Based DCA contribution scaling used in Strategy Analysis (not personalized advice), open Today's official DCA stance on the dashboard.
How the official monthly DCA framework uses these bands
The score bands above also drive GhostGauge's published monthly Risk-Based DCA framework — the same rules used in Strategy Analysis's official Baseline vs Risk-Based comparison. The framework scales a base planned monthly contribution by a band multiplier. It is context for planning, not a buy/sell bot and not personalized financial advice.
These multipliers apply to new monthly contributions only. They do not represent automatic sell or trim rules for Bitcoin you already hold. In this framework, Reduce Risk means sharply reduce new adds; High Risk means pause new adds.
Official six-band multipliers (monthly SSOT)
| Band | Multiplier (vs base monthly contribution) |
|---|---|
| Aggressive Buying | 1.5× |
| Regular DCA Buying | 1.0× |
| Moderate Buying | 0.75× |
| Hold & Wait | 0.5× |
| Reduce Risk | 0.25× |
| High Risk | 0× |
Example (hypothetical, not a recommendation): if your base monthly contribution were $1,000, the Hold & Wait band (0.5×) would size that month's new contribution at $500 under the published rules.
For the current production snapshot band and stance on the dashboard, or Strategy Analysis for the full monthly comparison detail.
Key Risk Factors
The current production G-Score combines seven enabled scoring factors across five analytical pillars. Their weights are defined by the versioned production configuration. On-chain Activity remains defined in configuration but is disabled at weight 0 and does not contribute to the current score.
Factor Weight Distribution
Trend & Valuation
What we look at
- • Price vs 200-day SMA (Mayer Multiple)
- • Distance to Bull Market Support Band (20W SMA / 21W EMA)
- • Weekly momentum (RSI proxy)
Why it matters
Measures price extension relative to long-term trend and momentum within the GhostGauge model.
How it affects risk
↑ stretch above trend ↑ risk; below trend ↓ risk
Update cadence & staleness
Freshness is source-aware and governed by the current production configuration; see the dashboard Input Status for the current factor state.
Primary sources
Coinbase daily close (UTC); rolling SMAs/EMA computed from that series
Caveats
Trend can stay elevated in strong markets.
Stablecoins
What we look at
- • Elapsed 7-day and 30-day supply change from the latest positive market-cap print
- • Market-cap weighted supply growth across the frozen seven-coin basket
- • Percentile versus dated calibration rows strictly before the observation date
Why it matters
Provides crypto-native liquidity context from stablecoin supply growth and composition.
How it affects risk
↑ supply growth ↓ risk; contractions ↑ risk
Update cadence & staleness
Freshness is source-aware and governed by the current production configuration; see the dashboard Input Status for the current factor state.
Primary sources
Stablecoin market data used by the current production factor.
Caveats
Exchange behavior and chain migrations can add noise.
Net Liquidity
What we look at
- • Fed balance sheet (WALCL), native Wednesday observations, ×1e6
- • Reverse Repo (RRPONTSYD), official FRED Wednesday-ending weekly average, ×1e9
- • Treasury General Account (WTREGEN), native Wednesday observations, ×1e6
- • Exact common-Wednesday intersection only; a missing RRP print makes the factor unavailable
Why it matters
Provides a broad macro-liquidity backdrop used as one input to the GhostGauge model.
How it affects risk
↓ net liquidity / negative Δ ↑ risk; ↑ liquidity ↓ risk
Update cadence & staleness
Slower public-data/FRED cadence. Freshness is source-aware and governed by the current production configuration; see the dashboard Input Status for the current factor state.
Primary sources
Caveats
Macro proxy; indirect for BTC.
ETF Flows
What we look at
- • US spot BTC ETF net creations/redemptions
- • 21-day business-day flow momentum
- • Weekend/holiday exclusion logic
Why it matters
Provides a regulated-vehicle flow proxy for institutional participation.
How it affects risk
↑ sustained inflows ↓ risk; persistent outflows ↑ risk
Update cadence & staleness
Business-day source. Freshness is source-aware and governed by the current production configuration; see the dashboard Input Status for the current factor state.
Primary sources
SoSoValue supplies the live ETF observations. Farside supplies the preserved frozen historical calibration and is not the live source.
Caveats
Holidays/reporting lags.
Term Structure & Leverage
What we look at
- • Completed 30-day funding mean from one eligible provider
- • Realized spot volatility from 31 prices and the 30 returns between them
- • The frozen stress component on that same completed window
- • Calibration against exactly 60 prior valid reference windows
Why it matters
Captures derivatives funding, realized-volatility, and leverage-stress conditions used by the model. Each component is scored from one provider's own history.
How it affects risk
↑ funding + ↑ volatility + ↑ stress ↑ risk
Update cadence & staleness
Raw-provider freshness is the latest funding observation from the selected provider. The completed scored cutoff is the daily window that is scored. Binding lastUpdated is the earliest of the scored funding, stress, and spot observations. Those three times stay distinct. Freshness is source-aware and governed by the current production configuration; see the dashboard Input Status for the current factor state.
Primary sources
Provider preference is BitMEX, then Binance, then OKX. Fallback is independent. Funding history is not spliced across providers.
Caveats
Fewer than 60 prior valid windows leaves the factor unavailable. A failed provider does not contribute rows to another provider.
Social Interest
What we look at
- • Bitcoin trending rank and social attention
- • 7-day vs 7-day price momentum
- • Search attention patterns
Why it matters
Social Interest scores only when CoinGecko search rank and 7-day price momentum are both observed. A missing component does not become a neutral score. The factor time is the earlier of the search acquisition time and the score-eligible price observation. A cached search response keeps its original acquisition time.
How it affects risk
↑ search attention + ↑ momentum ↑ risk
Update cadence & staleness
Freshness is source-aware and governed by the current production configuration; see the dashboard Input Status for the current factor state.
Primary sources
CoinGecko trending attention/rank plus BTC price-momentum context.
Caveats
Attention signals can be noisy and are a supporting input, not a standalone read.
Macro Overlay
What we look at
- • Dollar strength (DXY momentum)
- • Interest rate environment (2Y/10Y Treasury)
- • Risk appetite gauge (VIX level + momentum)
Why it matters
External macroeconomic factors. Dollar strength, rising rates, and market fear affect risk asset performance.
How it affects risk
↑ dollar strength + ↑ rates + ↑ fear ↑ risk
Update cadence & staleness
Slower public-data/FRED cadence. Freshness is source-aware and governed by the current production configuration; see the dashboard Input Status for the current factor state.
Primary sources
Caveats
Macro factors are indirect and can have delayed effects on BTC.
On-chain Activity
Disabled / 0% — not included in current G-ScoreOn-chain Activity remains defined in configuration but is disabled at weight 0 and does not contribute to the current score. It is not a current scoring source and is not part of the Momentum pillar contribution.
Data Status & Source Handling
Quality Controls
- • Schema tripwires for format changes
- • Z-score tripwires for outliers
- • Cache fallbacks for API failures
- • Retry logic with exponential backoff
Source-Handling Features
- • Multi-source fallback chains
- • Business-day aware calculations
- • Historical baseline comparisons
- • Source freshness and input-status checks
Factor Weights
Current Weights (v1.2.0)
Liquidity/Flows 30%, Momentum/Valuation 30%, Term Structure/Leverage 20%, Macro Overlay 10%, Social/Attention 10%. Seven enabled scoring factors contribute according to the versioned production configuration. Only factors included under the production freshness rules contribute to a snapshot, with weights normalized over the included set.
Current Weights
How weights are applied
Each enabled factor contributes according to the weights defined in the current production configuration. If a factor is excluded from a snapshot, remaining enabled weights renormalize for that snapshot. Weights are not recalculated from live performance, correlation studies, or claimed historical predictive power.
Data Sources & Methodology
Our methodology combines multiple data sources and analytical approaches to provide a comprehensive risk assessment.
Data Sources
- UTC intraday BTC price snapshot, with completed daily closes maintained separately for historical calculations
- Stablecoin market data used by the current production factor
- SoSoValue live ETF observations, with Farside kept as the frozen historical calibration
- FRED public-data series for Net Liquidity and Macro Overlay
- One eligible funding provider for Term Structure & Leverage: BitMEX, then Binance, then OKX, without cross-provider splicing
- CoinGecko trending attention/rank plus BTC price-momentum context
Update Frequency
- Daily UTC intraday production snapshot
- Individual inputs can have different vintages and cadences
- Current factor state is shown on dashboard Input Status
- Not a continuously streaming real-time feed
Model version: Current production: v1.2.0 / semantic-correctness-2026-09 · five pillars weighted 30/30/20/10/10. Live ETF source: SoSoValue. Farside is the frozen historical calibration, not the live source. Macro uses FRED for dollar and rates, and direct Cboe official daily history for VIX, with a guarded FRED VIXCLS fallback. Stablecoins, Net Liquidity, Social, and Term use the v1.2.0 successor semantics described above. On-chain remains disabled at weight 0.