Methodology

This page documents exactly how every number on this site is produced. If a figure appears anywhere on the site, its derivation is described here. Data as of Jul 17, 2026.

What the percentile means

For each asset except the benchmark, we take its daily adjusted closing price and divide it by the S&P 500 ETF (SPY) price on the same day. That gives a daily relative-price ratio. We then ask: out of every trading day in the asset's trailing 10-year window, on what share of those days was the ratio at or below today's ratio? That share is the percentile.

A reading of “Near lows (5%)” means the asset's price relative to the S&P 500 is lower than it was on 95% of the trading days in its own past decade. “Near highs (95%)” is the opposite. Neither means the asset is cheap, expensive, undervalued, or a buy — only that its price sits low or high within its own historical range.

Why we never compare valuations across assets

A natural question is: “Europe's PE is 18 and the US's is 27, so isn't Europe cheaper?” We deliberately never make that kind of claim, because it is misleading. Europe has traded at a lower PE than the US for most of the last two decades, largely because its market is weighted toward banks and industrials rather than the high-multiple technology companies that dominate the US index. A gap that has been present continuously for fifteen years is a structural feature, not a signal.

The same logic applies across sectors: technology has almost always carried a higher PE than energy or financials, and it always will, because the businesses are different. Comparing one asset's absolute valuation to another's tells you about their composition, not about which is a better value today.

Every percentile on this site compares an asset only to its own history. That is the one comparison that isolates change from structure.

Why relative price, not PE, is the primary metric

Trailing PE has two problems for this purpose. First, its history is not freely available — data vendors charge for multi-year PE time series, so we cannot compute a PE percentile going back a decade today. Second, PE is unreliable for cyclical sectors: energy earnings peak and trough with commodity prices, so a “low” energy PE often appears right before earnings fall, not when the sector is a bargain.

Relative price sidesteps both problems. Price history is long and free, and a price ratio does not get distorted by a single quarter of cyclical earnings. We still show each asset's current trailing PE as a plain snapshot, with its as-of date — but we make no historical percentile claim about it.

We do record each asset's trailing PE every day going forward. Over time this builds a proprietary PE history that will eventually support a PE percentile — but that series only began at this site's launch, so it is not yet long enough to use.

The 10-year window and short-history assets

The percentile window is exactly ten calendar years, cut by date — not by counting a fixed number of trading days, which would silently stretch the window because markets trade roughly 252 days a year, not 365.

• Assets with 10 or more years of history use the full 10-year window.
• Assets with 3 to 10 years of history show a percentile against their entire available history, clearly labeled “since inception” with the number of years, rather than claiming a 10-year figure.
• Assets with less than 3 years of history show no percentile at all — the sample is too small to be meaningful.

Data source and update cadence

Prices and trailing PE come from Yahoo Finance. Every day, an automated job fetches the latest daily closing prices and trailing PE for all assets, recomputes the ratios and percentiles, and commits the result. The site rebuilds from that committed data. Figures reflect the previous session's close; every number on the site carries its own as-of date.

This site is an educational data tool. It publishes observations about historical price ranges. It does not publish forecasts, recommendations, or personalized advice. See the disclaimer.