Daily Cross-Asset Market Outlier Report

This free daily market screener ranks all 73 instruments across stocks, FX, crypto, precious metals, energy and the US Treasury curve by how unusual each one’s latest reading is versus its own history. Every market measures “unusual” differently, so each instrument is placed on one comparable scale — shown in two cadence lanes, daily price signals and weekly positioning — so that today’s most statistically unusual market readings across asset classes as of 2026-10-20 stand out at a glance. Unlike a “biggest movers” list — dominated by whatever asset is most volatile, usually crypto — this ranks each market against its own history, so a rare move in a quiet currency pair can outrank a routine swing in Bitcoin. 23 were unusual enough to flag that session. It is a discovery tool for deciding what is worth a closer look — not a signal generator, forecast, or recommendation.

Standout in each asset class

Today’s board: The highest outlier is GM (Stocks), at a score of 100 out of 100. 22 of 72 daily-board instruments cleared their flag threshold today (1 FX, 2 Metals, 17 Stocks, 1 Crypto, 1 Energy).

Each reading is the instrument’s single most extreme signal in the latest session, placed on a single 0–100 extremeness scale. That scale is fed two ways, depending on the signal: RSI, 20-day return, volatility and relative volume are scored by their percentile within the instrument’s own history; the range signals — 52-week position and CFTC positioning — are scored by where today’s value sits within its band (0 at the bottom, 100 at the top), a current-location reading rather than a historical percentile. Both map onto the same distance-from-middle axis, so all signals stay directly comparable. How to read a row: a reading at the 1st percentile is lower than this instrument on about 99% of days in its recorded history — a statistical low extreme; the 99th percentile is the opposite. A live example from today’s board: EUR/GBP sits at the 1st percentile on RSI(14) — its latest reading is lower than on roughly 99% of all trading days in its recorded history since 1999 — which is what places it towards the top of the board. The further a reading sits from the middle of its scale, the higher its outlier score, which is how the table is ranked. The score is simply twice the reading’s distance from the middle of its own scale — 0 at the midpoint, 100 at either extreme — a simple rescaling of that distance, not a separate statistical model. Folding both tails onto one axis is deliberate: it lets a low extreme (near the bottom of its scale) and a high extreme (near the top) rank as equally notable, so a rare sell-off and a rare melt-up sit side by side — ranking by the raw reading instead would bury every low one at the bottom of the board and surface only the highs. Rows can share a score because percentiles are rounded to whole numbers, and ties are ordered deterministically and carry no extra meaning. Because each instrument contributes only its single most extreme signal, the very top of the board normally sits near 100 on almost any day — even a quiet one — so a high top score is expected and is not by itself a sign of market stress; what varies day to day is how many rows clear the flag threshold below. Rows marked FLAGGED clear the flagging threshold used in that instrument’s own daily report — a cutoff that generally falls near the tails of its historical distribution but is not defined by one universal percentile, so a reading may sit close to the tail yet not clear it. (A one-day price move alone never flags a row; only a signal that clears the threshold above — an RSI, range, return, volatility or positioning reading at a genuine extreme — does.)

What a percentile does and does not say: it measures where today’s reading sits within this instrument’s own recorded history — nothing more. It does not adjust for the shape of that distribution (at the same percentile, a move can be far larger in a fat-tailed asset than a calm one), nor for the fact that a market’s behaviour shifts across decades, so a long history spans regimes that no longer exist. Read it as historical standing, not a risk-adjusted or predictive measure. Each percentile is measured against the instrument’s entire recorded history — the “since YYYY (N obs)” span shown on each row — not a fixed rolling window, so a longer-lived instrument is judged against more data.

Markets covered

• 11 forex (FX) pairs
• 4 precious metals
• 48 US large-cap stocks
• 4 cryptocurrencies
• 5 energy futures
• 5 US Treasury maturities, ranked as one curve row

Updated once every trading day, after all covered markets have reported.

The set is fixed, not filtered: a deliberately chosen cross-asset benchmark — major FX pairs, sector-diversified US large-caps, the main cryptocurrencies, the four precious metals, the core energy futures and the US Treasury curve — so the same instruments are scored every day and nothing is added or dropped based on what happens to look extreme. It is a benchmark monitor, not a filter over a large equity universe.

Weekly-cadence signals

CFTC positioning updates weekly (Tuesday snapshot, Friday release), so it is ranked in its own lane rather than against the daily price signals above. Extremeness is scored the same way — distance from the instrument’s own history; only the update frequency differs. Weekly positioning signals are ranked only against other weekly positioning signals.

By asset class

This board ranks every market together on one comparable scale. To focus on a single asset class — forex, stocks, crypto, precious metals, energy or government bonds — open its own daily report:

• Forex daily report — 11 instruments
• Precious metals daily report — 4 instruments
• US stocks daily report — 48 instruments
• Crypto daily report — 4 instruments
• Energy daily report — 5 instruments
• Government bonds daily report — 5 instruments