Methodology

Every number on this site can be reproduced from public, official data. Here is exactly how.

Data sources

Holdings come from official issuer daily holdings files (e.g. State Street/SPDR publishes a complete holdings spreadsheet for every fund, updated daily) and SEC N-PORT filings, in which every US fund is legally required to disclose its full portfolio. We do not scrape third-party estimates. Each page shows its data as-of date.

Weight overlap (the headline number)

For every stock held by both funds, take the smaller of the two weights, then sum. Example: if Fund A has 3% in Apple and Fund B has 5%, Apple contributes 3 percentage points. The result answers a practical question: "what share of my money ends up in the same stocks either way?" This is the same approach used by professional portfolio analysis tools.

Portfolio lookthrough

For the homepage calculator, each of your funds is weighted by the dollar amount you enter; each fund's holdings weights are multiplied through and summed per stock. The output is your true underlying exposure — "you own 6.2% NVIDIA across everything," not "you own three ETFs." This is the Stock Intersection piece of the retired Morningstar Instant X-Ray, rebuilt free — see our Morningstar Instant X-Ray alternative guide.

Portfolio-aware what-if analysis

Portfolio Review starts with the funds and amounts entered in the calculator. For each what-if, we normalize each fund's included equity holdings to 100% of its analyzed equity sleeve. This prevents excluded cash or derivatives from making one plain-equity fund appear artificially less concentrated than another.

For each selected fund, we calculate its stock-level contribution to the portfolio. At every stock, the duplicated contribution is the smaller of (a) that fund's contribution and (b) the contribution from all other selected funds combined. Summing those values identifies the fund that repeats the most exposure at the user's current allocation. These fund-level amounts are diagnostic and are not additive across the portfolio.

Automatic candidates must sit in a manually reviewed comparable role, use a healthy holdings source, and have an as-of date within 62 days of the fund being tested. A candidate is shown only when overlap with the rest falls by at least 5 percentage points, top-five and largest stock concentration do not materially worsen, and either top-five concentration falls by at least 0.5 percentage points or effective stock count rises by at least 5%. Effective stock count is the inverse of the sum of squared stock weights.

Consolidating near-identical wrappers such as two funds tracking the same index can simplify a portfolio without diversifying its underlying stocks. We do not label that as a diversification result. Candidates are holdings-only ideas to investigate, not recommendations; fees, taxes, liquidity, tracking error, performance, account type, and personal goals are outside this model.

Verification

We test the engine against hand-computed results from official files. Example: SPYD vs SDY (two SPDR dividend ETFs, holdings as of 2026-07-09) — 22 shared holdings, 20.5% weight overlap, computed directly from State Street's published spreadsheets. Ask an AI chatbot the same question from memory and you'll get a plausible ballpark with wrong specifics — which stocks, and at what weights, is exactly what stale training data gets wrong. That is why this tool reads the actual filings.

Known limitations

• Mixed as-of dates. Issuers publish holdings on different schedules. When you combine funds whose files carry different dates, the result mixes those snapshots — enough to move a weight by a few basis points, not enough to change the picture. Every fund page and comparison shows its own as-of date.
• Identifier matching. Holdings are matched by security identifier and normalized name. Share classes, depositary receipts, and multi-listed lines are treated as the same underlying company where we can establish that; where we can't, the position appears separately rather than being merged on a guess.
• Non-equity and unclassified positions. Cash, futures, currency forwards, and securities we cannot map to a sector are reported as Unclassified rather than silently dropped or spread across sectors.
• Coverage. Only funds in our universe can be combined. Funds we don't cover simply aren't offered in the picker; nothing is estimated for them.
• Not advice, not real-time. Every figure is a description of published filings at a stated date. Nothing here is investment advice, a forecast, or a live quote.

Where the options metrics went

Until July 2026 this site also published nightly max pain and gamma-exposure dashboards. They moved to our sister site Quant Data — same engine, same nightly cadence, now covering individual stocks alongside ETFs, with the full methodology on its own methodology page. Old URLs under /options redirect there automatically.

What we don't do

No investment advice. No "urgency" scores designed to sell you something. No signup wall. No brokerage connection. The tool states facts from filings; what to do with them is your call (ideally with a professional).

Want the complete portfolio reviewed and kept under watch? The paid Portfolio Review + Monitoring service includes every qualifying what-if scenario in the supported universe, a saved dated baseline, and 12 monthly holdings checks. A human reviews the result when a material change is detected. It uses the same data, formulas, and disclosures, and does not promise a human-written report when nothing material changed.