01 / Trend Following
Trend
The current public trend stack consists of a five-day medium-term model and a weekly long-term model. Both use shared regime bands on a normalized `-1.00` to `+1.00` scale.
Methodology
How the models are built, tested, and interpreted.
AstravaQuant uses systematic frameworks designed to convert market data into repeatable measurements of trend, liquidity, valuation, and regime.
The objective is not to predict every market movement. The objective is to define market conditions consistently and evaluate how those definitions have behaved historically.
01
Trend
02 / Mortgage Rate Pressure
Mortgage Rate Pressure
MRPI is a separate weekly system anchored to the 10-year Treasury backdrop. Its five published regimes run from Strong Tightening through Neutral to Strong Easing.
03 / Valuation
Valuation
The current valuation system averages 17 standardized input readings across fundamental, technical, and sentiment categories. The app publishes the workbook score without converting it into a dollar target.
Backtesting
Backtesting.
AstravaQuant evaluates historical behavior using dated score observations that already exist in the published workbooks. The public app does not fabricate return series, fill missing dates, or imply live track records from hypothetical simulations.
Versioning
Methodology changelog
Added read-only MRPI workbook ingestion, six-criterion verification, dated MRPI history, and public tightening versus easing publication.
Added read-only Google Sheets ingestion, automated Medium-Term Trend and Long-Term Trend publication, relative-strength normalization, and dated score diagnostics.
Initial public methodology release covering Medium-Term Trend, Long-Term Trend, MRPI, interpretation scales, and the manual fallback snapshot.
Auditability
Preserve prior methodology.
Material methodology changes should create a new version entry rather than silently rewriting the past. Prior versions remain part of the research record.
Release discipline
Every material input, threshold, interpretation, or model-version change is recorded here before the revised framework is published.