The Universe of Symbols

Percentile scores are always computed against the same cohort: ~700 names — 500 S&P 500 stocks, 10 Nasdaq-100 additions not in the SP500, and a curated set of ~200 ETFs. Together they represent about 95% of US investable equity. The point of a finite cohort instead of every listed symbol is signal hygiene: we're ranking equities against the names actually competing for capital, not against the long tail of microcaps and SPACs. Symbols outside the 700 — Russell 1000 mid-caps, for instance — can still be scored, but they don't join the reference cohort itself.

Stock ranking systems exclude funds because they're "different instruments." But the investor's dollar goes to whichever place offers the better risk-adjusted prospect. Forcing AAPL onto the same percentile axis as SPY, XLK, TLT, GLD, and IBIT makes every measurement answer the question: Did the stock beat the cheapest passive alternative? The ETF tier carries asset classes underrepresented in the SP500 — overseas equity, metals, long-duration Treasuries, thematic baskets. Capital rotates across those classes when regimes shift, and the cohort needs to register that rotation rather than treat it as outside the frame.

A 95th-percentile market-beat on a stock isn't "best of the large caps" — it's beat 95% of the places that dollar could have gone, including the index funds, the sector ETFs, the long-duration Treasuries, and the gold trust.

Together, the 703 names represent ~$73 trillion of capital exposure — roughly the entire US public-equity market plus a curated pointer at every other asset class a US investor can wrap a dollar in. The double-counting at the wrapper level is intentional. SPY holds a slice of the SP500; XLK holds AAPL; VTI overlaps with almost everything. We're not building a non-overlapping portfolio — we're building a denominator. An investor's dollar can land in AAPL, in SPY, in XLK, or in TLT, and the cohort needs all four destinations because the dollar moves between all four.

The ETF row is where the cohort breaks out of pure stock-picking — 31 broad-market wrappers, 50 sector ETFs, 36 international funds, 25 income vehicles, 19 bonds, 11 commodities, 3 crypto, 18 thematic. A separate post (ETFs) walks through how the 193 were chosen from ~4,600 listed ETFs and what each category is a bet on. A stock can be at the 80th percentile on Market Beat 1Y and still lose to a handful of sector ETFs and a long-bond fund — and the percentile reflects that. A stock-only cohort would call it a winner; this one tells more of the truth.

The Russell 1000 extended tier, with 503 mid-cap names that aren't already in the SP500, NDX, or ETF set, has an unusual status: scoreable but not scoring. Russell-tier names get a full SYMBOL page, but the denominator of those ranks is still the 703-name cohort, not the extended 1,206-name list. This asymmetry exists for one reason: adding the Russell tier to the denominator would change every existing percentile. AAPL's 87th-percentile reading would shift the moment a new midcap joined the index, even if AAPL hadn't done anything different. A Russell-tier midcap at the 70th percentile means "ranks above 70% of the curated 703," not "above 70% of the 1,206."

With 703 names, each percentile point represents about 7 names — granular enough to distinguish 87th from 88th, dense enough that one IPO doesn't visibly shift ranks. And the denominator turns over slowly: the SP500 swaps ~20 names a year, NDX adds a handful, the ETF watchlist is hand-edited. AAPL's percentile today and the same metric six months ago are computed against essentially the same set.

When the SYMBOL card says AAPL is at the 87th percentile for Market Beat 1Y, the implicit comparison is:

Over the past year, AAPL's excess return vs SPY beat 87% of the 703-name cohort — the 500 large caps, the 10 foreign large-caps, and the 193 ETFs spanning every major asset class a US investor can wrap a dollar in.

The dashboard does not measure stocks against the universe of public companies. It measures them against the universe of things an investor would actually buy. Everything else on the site — the percentile fingerprint, the LENS dispositions, the LEADERS leaderboard, the COMPARE table — follows from that choice.