Skip to content

Model Overview

This page is the top-level map of the StackSats model.

The model computes daily DCA weights using MVRV and trend context while the framework enforces invariants (sum == 1.0, feasibility clipping, and lock semantics).

Reading Path

  1. Signals and Features
  2. Allocation Kernel
  3. Backtest Runtime

Model Architecture

flowchart LR
    subgraph Inputs
      P["Price history"]
      M["MVRV data"]
    end

    subgraph Features
      F1["Signals + lagging"]
      F2["Confidence and volatility"]
    end

    subgraph Intents
      I1["Raw daily intent"]
      I2["Dynamic multiplier"]
    end

    subgraph Framework
      K1["Feasibility clipping"]
      K2["Locked-prefix preservation"]
      K3["Remaining-budget enforcement"]
    end

    O["Final weights"]

    P --> Features
    M --> Features
    Features --> Intents
    Intents --> Framework
    Framework --> O

Core properties

  • Deterministic output for identical inputs.
  • Historical allocations are immutable once locked.
  • Future allocations are feasibility-aware and budget-constrained.