How are training results stored?

Archives are organized in a timestamped directory structure containing a master index, raw JSON summaries, model assessments, and human-readable Markdown reports.

Is the gating configuration customizable?

Yes, users can define custom ModelGatingConfig objects to set stricter or more relaxed thresholds based on their specific risk tolerance.

How does the skill handle overfitting?

It monitors fitness declines from peak performance; if a decline exceeds 5% or oscillation exceeds 10%, it triggers an unstable training flag and recommends an earlier checkpoint.

What metrics are used for model gating?

The system evaluates Fitness (>=0.70), Profit Factor (>=1.8), Consistency (>=85%), and Max Drawdown (<=8%) against calibrated thresholds for production approval.

Trading Model Training Gating

Name: Trading Model Training Gating
Author: smith6jt-cop

bysmith6jt-cop

0•

Data Science & ML

Automates the archival and quality classification of algorithmic trading models based on performance metrics and risk thresholds.

The Training Archive & Model Gating System provides a robust framework for managing AI training lifecycles within algorithmic trading environments. It implements a rigorous 'APPROVED/REVIEW/DROP' gating mechanism to ensure only high-performing, consistent models reach production. By tracking metrics like Profit Factor, Consistency, and Max Drawdown, and offering built-in overfitting detection with checkpoint recommendations, this skill transforms ephemeral training outputs into a structured historical reference, preventing the deployment of unstable or overfitted strategies.

Key Features

010 GitHub stars

02Structured historical archival for long-term performance tracking

03Overfitting detection with checkpoint rollback recommendations

04Automated generation of human-readable assessment reports

05Automated multi-metric gating (Fitness, PF, Consistency, MaxDD)

06Customizable quality thresholds for specific trading strategies

Use Cases

01Detecting training instability or overfitting during model development

02Determining production readiness for new algorithmic trading strategies

03Auditing historical training runs to compare model iterations over time

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add smith6jt-cop/skills_registry training-archive-gating

For use in Claude.ai and ChatGPT

Download Skill