How the catalogue grades and computes.

1 · The verifiability grade (A/B/C)

The grade answers one question: who computed this record, from what? It grades the record's format, never its quality, and never its future.

A — platform-verified

The venue computes the record from a real account or certificate: real fills, platform math, a regulated operator. wikifolio certificates and Darwinex DARWINs live here. A− flags a record with an imported pre-listing segment or a partially gated risk panel.

B — timestamped or wrapped

One layer is verified, one is not. Collective2 timestamps every signal platform-side but fills are hypothetical by design. CTA databases sit inside a real regulatory wrapper (NFA), but the public figures are manager-reported and the full series is registration-gated.

C — public calls, no composite

Analyst targets and strategist calls are public and datable, but nobody computes an outcome record over them, revisions are free, and no standard scoring convention exists. We grade the format honestly and score the headline calls one by one in the Famous Calls Record.

One consequence worth stating: a fund with a daily NAV (an ETF, say) is more verifiable than any leaderboard — which is why a regulated fund can out-grade a star trader even in a drawdown. Verifiability and performance are different axes. This catalogue never confuses them.

2 · The computed battery (v1)

Where our series pipeline covers a subject, we compute the metric battery ourselves from daily data — the platform's own numbers stay quoted separately, as shown and dated. Battery v1:

• Returns: daily simple returns from adjusted closes; CAGR over the stated window; total window return; trailing 1-year.
• Risk: annualised volatility (stdev × √252); maximum drawdown from the running peak, with dates; downside deviation (daily, annualised, target 0).
• Ratios: Sharpe and Sortino with rf = 0 (labelled on every page; a risk-free layer joins in a later version), Calmar (CAGR / |max drawdown|).
• Profile: monthly win rate, best and worst month, from month-end closes.

Every battery states its window, its data source and its computation date. A window that misses a famous drawdown says so on the page — the window is part of the number.

3 · The Closelook Rating (v1 — live)

The rating is a 0–100 quick view computed from the battery, with the full battery as its second level. Formula v1, in full:

• Evidence (0–25): 25 × √(years / 10) — a record earns points by existing through time; ten verified years earn the maximum.
• Risk-adjusted return (0–35): 35 × Sortino / 2.5, capped — a Sortino of 2.5 earns the maximum.
• Drawdown control (0–20): 20 × Calmar / 1.5, capped.
• Consistency (0–20): 10 × clamp((winning months − 40%) / 30%) plus 10 × clamp(1 − |worst month| / 15%).
• Short-window dampener: the three performance components are multiplied by min(1, √(years / 3)) — a hot eighteen months cannot out-score four verified years on heat alone.
• Verifiability cap: a NAV-verified or platform-real series can reach 100; timestamped-with-modelled-fills is capped at 75; a self-reported series at 55. A record nobody can verify cannot out-score a record everyone can.

Bands: ≥75 Exceptional · 60–74 Strong · 45–59 Solid · 30–44 Mixed · <30 Thin. Two fixed rules besides the formula: the rating scores the record, not the person — on subjects whose verifiable series is one slice of a broader public record (a strategist's ETF, say), the page names the slice. And where no series exists, the page says “insufficient data” — never a low score.

The battery also reports benchmark context — beta, correlation and excess CAGR against the S&P 500 (SPY) over the same window — which does not enter the v1 score. Planned for v2: regime-conditioned performance against our own market series, outlier-month dependency, and uniqueness against the rest of the catalogue. Version changes are dated here.

4 · Sourcing and restraint

Figures quoted “as shown” come from each platform's public pages, dated on the page and linked at the source. Series come from our own data lake and from public endpoints the platforms themselves serve to every visitor; login-gated data stays out of the catalogue until a platform or subject provides it — several records say exactly that in their grade basis. Refresh cadence is deliberately low: once per day after US close at most.

Subjects have not contributed to these pages. A listed subject who wants a figure corrected, a gated series verified, or their record claimed can reach us via the contact on the About page — corrections to factual errors are made and dated, and a claimed record upgrades its grade basis.

5 · What this is not

Closelooknet documents public records as part of an investment diary. Nothing in this catalogue is investment advice, a recommendation to follow, copy or invest in anyone listed, or a solicitation of any kind. Grades and ratings are our published-method opinion over sourced facts — the same genre as a fund star rating or an analyst scorecard, and readable the same way: a lens on the record, not a forecast.