Methodology
Every derived figure on Oxford Ledge — a score, a threshold, a chip, an answer — is computed by code we are willing to show you. These pages document each method in plain language and name the exact source files behind it, so the figure and the explanation cannot quietly drift apart. Where a page makes a numerical claim, a contract test pins that claim to the code constant, and CI fails if the two diverge.
Raw data we report unchanged (SEC filings, FRED series, Treasury yields, USAspending obligations) is not listed here — it is sourced, not computed. These pages cover the signals Oxford Ledge derives on top of that data.
Value Creation Score
The 0–100 composite that measures your research practice — breadth, depth, education, consistency, and community — not your portfolio’s returns. Covers the five weighted components, the soft-log normalization curve, the reference ceilings, the named levels, and the full changelog of every change to the equation.
Mastery Thresholds
How the LEARN system decides a concept is mastered. Covers the Bayesian belief update across the seven practice layers, the probability threshold a concept must clear, and why the loop is built to reward steady practice rather than a single lucky answer.
Insider Transactions (Form 4) Chip
How the insider-activity chip on a company page is computed from SEC Form 4 filings — the net-flow window, what counts as a buy versus a sell, the thresholds behind each label, and the cases where the chip stays neutral rather than over-claiming a signal.
Management Quality Grade
How the A–F management-quality grade on a company page is computed — four 0–25 components drawn from ROIC vs WACC spread, Return on Equity, Debt/Equity, and Net Margin, the 0–100 average behind the letter, and what the grade is explicitly not claiming (it is a profitability+leverage screen, not a governance audit).
Ask AI
What sits behind the Ask AI assistant: the provider model, the retrieval pipeline that grounds answers in the filings corpus, the system-prompt disclaimer, the prompt-injection defenses, and the Socratic-mode default. Written to be auditable without signing up.
ETF Lookthrough (True Exposure)
How the True-Exposure panel on your portfolio Risk tab is computed — the 13F-HR holdings source, the weight-blended sector and geographic aggregation, why levered and inverse ETFs are flagged but not expanded, why commodity and bond funds are counted at the headline ticker, and what the coverage percentage actually measures.
Source code and corrections
Each page names its authoritative source files and is reviewed on the date shown at the top of that page. How we keep every figure honest — the contract tests, freshness reviews, and public incident log behind the data — is documented in the Trust Dossier.
Corrections: oxfordledge@gmail.com.