Three proprietary metrics that fuse 15+ data sources into actionable intelligence.
Overview
Thrive Metrics are three proprietary market scores computed exclusively on the Thrive platform. They synthesize data from 15+ premium sources — derivatives positioning, smart money flows, on-chain analytics, social sentiment, technical structure, and macroeconomic indicators — into actionable intelligence that cannot be replicated anywhere else.
While traditional indicators show you one dimension of the market at a time, Thrive Metrics fuse multiple dimensions into composite scores designed to answer three fundamental trading questions:
⚡
Thrive Power Score
How strong is this setup? A 0–100 conviction score that tells you whether conditions are favorable across all data sources. Graded S through D.
🎯
Thrive Alpha Signal
Which direction is the edge? A -100 to +100 directional bias score that identifies asymmetric risk/reward and tells you whether to lean long, short, or stay flat.
🌊
Thrive Regime Pulse
What playbook should I use? Identifies one of six market regimes (Breakout, Momentum, Accumulation, Distribution, Capitulation, Chop) and tells you how intensely that regime is expressing.
Where to Find Thrive Metrics
Asset detail pages — full metrics breakdown with sub-component analysis and historical charts
Asset screener — TPS badge on every asset row, sortable by Power Score
Workbench — query thrive_metrics and thrive_metrics_history tables in SQL notebooks and dashboards
Live data feeds — stream real-time Thrive Metrics updates in workbench live widgets
Alerts — set threshold alerts on TPS and TAS values
Update Frequency
Thrive Metrics are recomputed every hour for all 30 tracked assets. Each computation pulls the latest data from all integrated sources. Historical values are retained for 90 days, enabling time-series analysis and backtesting against metric changes.
Data Confidence
Each metric includes a confidence score (0–100%) that reflects data completeness. If a data source is temporarily unavailable, the affected dimension receives reduced weight rather than producing an inaccurate reading. A confidence below 70% is flagged in the UI.
