How does it handle Google Colab integration?

It provides standardized patterns for GPU checks, Google Drive mounting, and secure API key parsing using specific _read_keys_from_file methods.

Why is the CLOSE action handled differently?

The skill ensures CLOSE actions are processed before drawdown guardrails, allowing the system to exit positions even if trading is otherwise halted for safety.

Does this skill support legacy trading models?

Yes, it includes a fallback mechanism for legacy 3-action models that do not possess a model_spec attribute, ensuring backward compatibility.

What specific bug does this skill address?

It fixes a critical mapping error where 8-action models had signals misinterpreted (e.g., BUY_50% treated as SELL) because the engine only expected 3 actions.

Trading Backtest & Colab Optimizer v4

Name: Trading Backtest & Colab Optimizer v4
Author: smith6jt-cop

bysmith6jt-cop

•

Data Science & ML

Corrects multi-action trading logic bugs and optimizes backtesting notebooks for Google Colab environments.

This skill provides Claude with specialized domain knowledge to resolve 8-action space bugs within trading backtest engines and migrate research notebooks to the Colab v4.0.0 architecture. It ensures precise mapping of complex trading signals—including tiered buy/sell percentages and explicit close actions—by implementing the interpret_action pattern. It is particularly useful for developers working on Alpaca-integrated trading systems who need to ensure backtest accuracy across different model versions while maintaining robust position sizing and guardrail logic.

Key Features

01Fixes 8-action mapping bugs for complex trading signal interpretation

02Optimizes Google Colab notebooks with Drive and Alpaca API patterns

03Implements uniform model handling via interpret_action() integration

04Ensures critical CLOSE actions bypass drawdown guardrail halts

051 GitHub stars

06Enables dynamic position sizing using size_mult parameters

Use Cases

01Debugging silent failures in backtesting engine signal interpretation

02Upgrading algorithmic trading bots from 3-action to 8-action logic

03Standardizing trading research notebooks for Google Colab v4.0.0

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add smith6jt-cop/skills_registry backtest-notebook-v4

For use in Claude.ai and ChatGPT

Download Skill