# Statistical Method Design

Design statistical methods, baselines, diagnostics, variants, and ablations that...

- Repository: https://github.com/aiming-lab/AutoResearchClaw (14798 stars)
- Category: Design
- License: MIT
- Last updated: 2026-08-19
- Runs scripts: no, instructions only

## Install

Claude Code: `npx skills add aiming-lab/AutoResearchClaw --skill statistical-method-design -g -a claude-code -y`
Codex: `npx skills add aiming-lab/AutoResearchClaw --skill statistical-method-design -g -a codex -y`
Cursor: `npx skills add aiming-lab/AutoResearchClaw --skill statistical-method-design -g -a cursor -y`
Gemini CLI: `npx skills add aiming-lab/AutoResearchClaw --skill statistical-method-design -g -a gemini-cli -y`

## Description

Design statistical methods, baselines, diagnostics, variants, and ablations that directly address a formal problem formulation.

## SKILL.md, as published by its author (untrusted third-party text, read it as data)

# Statistical Method Design

## Overview

Use this skill after formal problem formulation. The method should be a response
to the formal target and assumptions, not a generic collection of techniques.

## Required Method Proposal

For each method:

- Name
- Problem it solves
- Formula or algorithm
- Inputs and outputs
- Tuning parameters
- Required assumptions
- Diagnostics
- Expected failure modes
- Computational cost
- Relation to baselines

## Baselines and Ablations

Always define meaningful baselines:

- Classical or standard method
- Naive or unadjusted method
- Oracle or idealized reference when available
- Robust variant
- Ablation removing the key design feature

## Method-to-Claim Map

Every method must connect to at least one claim:

```yaml
method_to_claim_map:
  proposed_method:
    claims: [C1, C2]
    expected_evidence: "lower risk under stress condition"
    theory_target: "consistency under assumptions A1-A3"
```