walk-forward

Execute walk-forward validation against the Udgaard API to test strategy edge persistence.

Updated Jun 10, 2025
One-click install
npx skills add https://github.com/skrymer/trading --skill walk-forward
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: walk-forward
Source: https://github.com/skrymer/trading/tree/main/.claude/skills/walk-forward
Command: npx skills add https://github.com/skrymer/trading --skill walk-forward

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Run walk-forward validation to determine if a strategy's edge persists out-of-sample by delegating interpretation to the dedicated walk-forward-analyst sub-agent. This helps ensure robustness before live trading.

Core Features & Use Cases

  • Out-of-sample edge verification: Test edge persistence across multiple windows and cadences.
  • Automated orchestration: Coordinate backtest walk-forward runs against the Udgaard API and collate results for decision making.
  • Analyst delegation: Leverage the walk-forward-analyst to generate per-window insights and a verdict with recommended next steps.

Quick Start

Command the system to run a walk-forward validation for a given strategy and cadence, then review the saved results and analyst report.

Frequently Asked Questions about walk-forward

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run walk-forward validation to check if my strategy edge persists out-of-sample?

Walk-forward validation tests strategy edge persistence out-of-sample by executing backtest runs against the Udgaard API. It applies in-sample and out-of-sample splits across various cadences and window configurations to verify robustness before live trading.

What is walk-forward edge validation and when do I need it for strategy analysis?

Walk-forward edge validation is a backtesting process that determines whether a trading strategy maintains its edge on unseen out-of-sample data. You need it to ensure strategy robustness and prevent overfitting before deploying to live trading environments.

Can I configure custom in-sample and out-of-sample window splits for backtesting?

Yes, walk-forward validation supports various cadences and window configurations for backtesting. You can apply custom in-sample and out-of-sample splits along with optional ranking and size parameters to tailor the edge validation to your specific strategy design.

How are walk-forward backtest results analyzed and saved for review?

Walk-forward backtest results are saved locally for review after orchestration. A dedicated walk-forward-analyst sub-agent delegates interpretation to generate per-window insights, a final verdict, and recommended next steps for decision making.

Does walk-forward backtesting work without external dependencies?

Walk-forward validation orchestrates API calls specifically to the Udgaard backtest endpoint and delegates analysis to a walk-forward-analyst. You need access to the Udgaard platform to execute the out-of-sample edge verification runs.

What is the best way to validate out-of-sample strategy robustness before live trading?

The best way to validate out-of-sample robustness is walk-forward validation, which coordinates multiple backtest runs across different windows. This automated orchestration collates results and leverages analyst delegation to generate a clear verdict with recommended next steps.