Claude/crypto analysis requirements 9 he7 h #4
Conversation
…mplement-features-from-readme-in-branch-claude/crypto-analy-hriywu
…adme-in-branch-claude/crypto-analy-hriywu Add PROGRESS_REPORT.md summarizing Phase 1/2 progress
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: f6c4e8e744
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| rs = avg_gain / avg_loss.replace(0, pd.NA) | ||
| rsi_values = 100 - (100 / (1 + rs)) | ||
| return rsi_values.fillna(0.0) |
There was a problem hiding this comment.
When avg_loss is zero (e.g., a monotonically increasing price series), avg_loss.replace(0, pd.NA) forces rs to NA and rsi_values.fillna(0.0) returns an RSI of 0. This inverts the signal: RSI should be 100 in the “no losses” case, so any downstream strategy using RSI will misclassify strong uptrends as oversold. Consider explicitly setting RSI to 100 when avg_loss is zero (and only using fillna for the initial window).
Useful? React with 👍 / 👎.
1 participant
No description provided.