Skip to content

docs: add CLAUDE.md memory on NPF normalization (ADF vs second-moment)#29

Draft
mspinola wants to merge 1 commit into
mainfrom
claude/crucible-project-website-mbnaxw
Draft

docs: add CLAUDE.md memory on NPF normalization (ADF vs second-moment)#29
mspinola wants to merge 1 commit into
mainfrom
claude/crucible-project-website-mbnaxw

Conversation

@mspinola

Copy link
Copy Markdown
Owner

Adds a CLAUDE.md project-memory file recording the NPF (net non-commercial / open interest) discussion, so future Claude sessions in this repo have the context:

  • The accepted Discord feedback: normalizing by OI adds little in an ADF test (raw net is mostly level-stationary already); the real benefit is second-moment — era variance stability and lower kurtosis — which is what keeps full-history percentile thresholds and the gray zone meaningful.
  • The remaining caveat: NPF fixes scale, not trader composition (2004–06 commodity-index influx is a structural break no denominator addresses).
  • The pending npf_check.py verification (ADF, era variance ratio, kurtosis across 8 major markets) handed off to a trusted-network session.

Documentation only — no code changes.

🤖 Generated with Claude Code

https://claude.ai/code/session_01M3dvdfiAyaEuGfUaY9Z9g9


Generated by Claude Code

Records the accepted feedback on the NPF/OI report: dividing net by OI
adds little in an ADF test (raw net is mostly level-stationary already);
the real benefit is variance stability across eras and lower kurtosis,
which is what keeps full-history percentile thresholds meaningful. Also
notes the 2004-06 index-fund structural break caveat and the pending
npf_check.py verification.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01M3dvdfiAyaEuGfUaY9Z9g9
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants