Methodology / Trend

Trend following.

Trend models attempt to systematically identify the direction and persistence of market movement rather than forecast an exact future price.

AstravaQuant stacks multiple trend-following indicators and oscillators across complementary horizons, because market structure can change with the window.

Medium-Term Trend

Purpose A five-day read of intermediate crypto trend structure.
Inputs A stacked set of trend-following indicators and oscillators. The public site publishes the normalized score and dated observations.
Signal Construction Inputs are combined into one standardized five-day score. Detailed transformations and parameterization remain proprietary.
Normalization The public app reads the workbook score on a normalized `-1.00` to `+1.00` scale and does not rescale it in the browser.
Regime Thresholds `-1.00 to -0.75` Extreme Risk-Off, `-0.75 to -0.25` Defensive Contraction, `-0.25 to +0.25` Neutral Transition, `+0.25 to +0.75` Constructive Expansion, `+0.75 to +1.00` Full Risk-On.
Update Frequency The public site refreshes the workbook feed every five minutes by default and carries through the worksheet’s published update label when present.
Historical Test Period The public site uses verified dated observations only; it does not fill or create historical values.

Long-Term Trend

Purpose A weekly read of slower structural crypto trend.
Inputs A complementary stack of trend-following indicators and oscillators, read on a weekly horizon.
Signal Construction Inputs are combined into one standardized weekly score. Detailed transformations and parameterization remain proprietary.
Normalization The public app reads the workbook score on a normalized `-1.00` to `+1.00` scale and applies the same regime bands used by MTPI.
Regime Thresholds `-1.00 to -0.75` Extreme Risk-Off, `-0.75 to -0.25` Defensive Contraction, `-0.25 to +0.25` Neutral Transition, `+0.25 to +0.75` Constructive Expansion, `+0.75 to +1.00` Full Risk-On.
Update Frequency The public site refreshes the workbook feed every five minutes by default and carries through the worksheet’s published update label when present.
Historical Test Period The public site uses dated rows from `LT Total Forward Testing` only.
Current Interpretation LTPI is the slow structural read in the public stack and is intended to be read beside MTPI and the derived NSPI.
Interpretation Different trend horizons can disagree. That disagreement is information rather than automatically a model failure.