DigestAI news desk

Cut through the AI noise.

Society & Work7 min read

Claude recovers 5.34% in FX trading after 7.33% drawdown, ChatGPT holds steady

An experiment tracking AI-driven forex trading between Claude and ChatGPT concluded its fourth week. The Claude team, using the Loop If-Dan system, recovered from a 7.33% drawdown to a 5.34% net gain, locking in 1,995 yen in profits while holding two active positions with 393 yen in unrealized losses. ChatGPT’s TriAuto FX team remained unchanged, holding its 1.57% profit from prior weeks with no…

1 source

Key points

  • Claude’s Loop If-Dan system rebounded from a 7.33% drawdown to a 5.34% net gain in Week 4
  • ChatGPT’s TriAuto FX held steady at 1.57% profit with no new trades this week
  • FX account balances differ from stock trading due to margin requirements, not cash deductions

The experiment highlights contrasting strategies: Claude’s automated system executed two new trades this week after settling prior positions, while ChatGPT waited for predefined entry conditions. The author notes that FX account balances differ from stock trading due to margin requirements, which act as collateral rather than reducing cash balances. Both teams face a Stage 1 checkpoint on November 6 to assess their performance, though no immediate strategy changes are planned.

Full story from note.com · by AI FX Lab · via Search: ClaudeOpen source ↗

AI FX Lab | Season 1 Stage 1 Week 4 | Claude's Comeback! Recovery from Unrealized Losses and a New Misconception for FX Beginners

note.com · 1 November 2026

Loading the full article…

This text was published by note.com and written by AI FX Lab. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗

Topics · follow one to build your own front page

The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.

Comments

via GitHub Discussions

More in Society & Work

All →

Related stories