StockResearch Is Now a Valuation Research Platform

StockResearch started as a cross-listing comparison tool: two tickers, one company, automatic FX and ratio normalization. That tool is still here and still free. But over the past month we built the thing we always wanted next to it, and today it is live: a valuation research platform that charts any covered stock against its own history.

The one measurement everything hangs on

Most "cheap or expensive" arguments quietly compare a stock to other companies. We think the more useful first question is self-relative: compared to what this stock has cost over its own last five years, what does it cost today?

So that is the number we compute. For every covered stock we build trailing P/E, P/S, EV/Sales, and EV/EBITDA as weekly series, going back up to ten years, from SEC EDGAR XBRL filings and split-adjusted prices. Then we rank today's multiple inside the stock's own trailing five-year distribution. A stock at the 5th percentile trades near the bottom of its own range; at the 95th, near the top. No look-ahead: a quarter only enters the series once it was actually filed. The details are in our guide to what a valuation percentile is and how we compute it.

What went live

The screener. The valuation screener ranks every covered stock by that percentile and lets you filter by sector, multiple, and revenue-growth trend. Three presets encode the questions we kept asking ourselves in July: AI infrastructure trading at bottom-quartile valuations, multiple compression without revenue deceleration, and stocks below their own five-year median multiple. Next to the table there is a chart view, valuation percentile plotted against revenue growth: cheap-and-growing lands in one corner, expensive-and-slowing in the opposite one, and the interesting arguments live in between. Ticker hubs. Every covered stock now has a research hub. NVDA's valuation history shows the full band chart: the weekly multiple, the five-year median, the middle-half band, and where today sits in it. NVDA versus its peers puts the same multiples side by side with sector peers, and NVDA's implied price page answers the question people reach for a calculator for: what would the stock trade at back at its own median multiple, or at its quartiles? Swap in any covered ticker. The market itself. The S&P 500 valuation history page applies the same math to the index: the median P/E across S&P 500 constituents, computed weekly since 2016. As of this week the median sits at 24.9x, the 64th percentile of its ten-year range. Above the middle, nowhere near the top. You can now check that sentence yourself instead of taking a stranger's word for it, which is the point of the page.

What we mean by coverage, honestly

Coverage today is the S&P 500 plus a curated AI-infrastructure and international watchlist. That is real coverage for most research questions, and it is also nowhere near "every stock," so we say so. Every number carries provenance: which filings it came from and when it was computed. We cross-check computed multiples against independent snapshots, and when a number diverges we flag the row instead of publishing it silently. Foreign filers with annual-only filings get labeled as partial coverage rather than given fake percentiles, and if a percentile would rest on under two years of history, we do not show one.

We compute the dataset in-house because derived numbers you cannot audit are not research. They are vibes with decimals.

Cross-listing compare is not going anywhere

The cross-listing comparison tool stays free and first-class: pick a dual-listed stock and see the premium or discount between its listings, with FX and ADR/CDI ratios handled automatically. The two pillars feed each other. Cross-listing tells you which line of the same business is cheaper; valuation history tells you whether the business is cheap at all.

Proof it works: we used it first

We published our first flagship research piece today: which AI stocks actually got cheap in the July selloff. It ranks 50 AI-infrastructure names against their own five-year ranges, and its headline finding (only five names actually got cheap; about a third of the complex still trades near five-year-high multiples) came straight out of the same screener you can run yourself.

What's next

Three things are in progress:

• Accounts, so your screens and settings survive a browser change.
• Saved baskets: pin a set of tickers and track their percentile moves as one view.
• A monthly valuation report: the state of the market's valuations, in your inbox, built from the same dataset.

We will announce each of these right here on the blog. Everything above is free. Type a ticker, see what the market pays for it now versus the last five years, and argue with the number instead of the narrative.

---

StockResearch provides data and research tools, not recommendations. Nothing on this site is investment advice; always do your own research before making investment decisions.