Legal & Methodology

Methodology

Sparkline scores every fund on five composite factors built from underlying holdings. This page describes how those scores are constructed.

Funds Covered
28,254
Funds Scored
16,000
Metrics per Fund
568
Peer Groups
134
Holdings as of
2026-08-31

The Five Composite Factors

Each fund's score is the holdings-weighted aggregate of its underlying companies' scores, ranked within the clean equity fund universe (one entry per fund series); 0 = lowest, 100 = highest, and the middle of the range is denser than the ends.

Intangible Value
Composite of five sub-pillars: intellectual property, brand equity, human capital, network effects, and tangible capital. Each of these sub-pillars are themselves calculated as the composite of various metrics (e.g., R&D/market cap, trademarks/market cap, PhDs/market cap, Book value/market cap).
Traditional Value
Composite of book/price, sales/price, earnings/price, fcf/price, opcf/price, and ebitda/ev, using both trailing twelve month figures and consensus next twelve month estimates.
Small-Cap
Inverse company market capitalization (smaller firms get higher scores).
Quality
Return on equity over a trailing twelve month period.
Momentum
Trailing twelve month return.

About Sparkline Capital

Sparkline Capital is an investment management firm applying state-of-the-art machine learning and computing to uncover alpha in large, unstructured data sets. Much of our research concerns “intangible value” — the idea that a company's intellectual property, brand equity, human capital, and network effects drive real economic value while remaining largely absent from its balance sheet, and therefore from traditional value measures.

The factor framework described here aims to address that limitation, drawing on dozens of research papers published in full on Sparkline's website.

Read the published research(opens sparklinecapital.com in a new tab)

Fund Coverage Rules

Our data is not applicable to all funds. Below are cases where funds fall outside our coverage universe.

  • Holdings Analytics: Registered US Funds Only. Everything we derive from holdings — factor scores, size and style placement, sector, country and thematic exposure, look-through and concentration — is built from Form N-PORT filings, which only investment companies registered under the Investment Company Act of 1940 file: mutual funds, ETFs, closed-end funds, and unit investment trusts such as SPY and QQQ. A listed set of products organized under the Securities Act of 1933 instead — physically backed metals and spot crypto trusts, commodity futures pools and currency trusts, among them GLD, SLV, IBIT and USO — is carried at price level, with the full return history, risk statistics and factor regression, and none of the holdings-derived figures above. Each such page says so on its face. Funds listed outside the United States are out of scope entirely.
  • Common Equity Only. Factor scores are computed from a fund's common-equity holdings. Bonds, cash, derivatives and other assets are excluded from the factor calculation.
  • 50% Coverage Floor. When less than 50% of a fund's total net assets can be scored, its factor scores are deemed unreliable and thus blanked.
  • One Share Class per Series. Peer groups and percentile ranks use a single representative share class per SEC series, so a fund with six share classes counts once rather than six times.
  • Small Peer Groups Suppressed. Percentile ranks require at least 10 reporting members; a fund whose group falls short is ranked against a broader tier instead.

Data Sources

Holdings
Form N-PORT filings to the SEC — public regulatory data. Each fund page states the filing date its holdings-derived figures come from. Because N-PORT is filed with a lag, a fund's current book may differ from what is shown.
Company Fundamentals
Sparkline Capital's own research dataset, which supplies the per-company factor scores that fund scores aggregate.
Fees, Flows and Fund Facts
SEC filings — prospectuses (485BPOS) and shareholder reports (N-CSR) — plus reported net assets.

Peer Group Taxonomy

Every fund is placed in three nested tiers. A fund is ranked against its peer group where that group has enough members to be meaningful, and against a broader tier where it does not.

Peer Groups (134)

The narrowest tier, and the default pool for percentile ranks

Peer Regions (5)

The middle tier, which only applies to diversified equity funds (i.e. excludes leveraged, inverse, derivative, and sector funds)