Where the money is
positioned in gold.

The Commitment of Traders report breaks the gold futures market into the traders who take a view (large, non-commercial speculators) and the traders who hedge physical exposure (commercials). This week, non-commercials hold 357,494 long contracts against 71,060 short, for a net long position of +286,434 — equal to 0.5% of the 541,232 contracts of total open interest. Commercials sit on the other side of that trade, net short at -312,658, as they almost always are. The single most useful number, though, is not any of these absolutes — it is where this week ranks against its own past year.

Raw positioning is easy to misread. A net +286,434 long position sounds like a strong directional bet — but the number is meaningless without a scale. Gold's open interest, the number of large players, and the market's baseline size all drift over time, so a position that was extreme two years ago can be unremarkable today. This is the difference between absolute and relative positioning:

• Absolute positioning is the headline contract count — net +286,434 this week. It answers “how many?” but not “is that a lot?”
• Relative positioning asks the better question: compared to its own recent history, is this crowded or washed out? That is exactly what the 52-week COT Index measures.

The COT Index rescales the current net position against its own trailing-52-week range: a reading of 100 means this is the most long speculators have been all year, 0 means the most offside, and 50 is the midpoint. It converts a raw, unanchored contract count into a percentile you can actually reason about.

This week the index reads 84 out of 100 — positioning sits near the top of its trailing-year range — historically a crowded, one-sided long.

Two things are worth holding in mind at once. In absolute terms, speculators are net long by +286,434 contracts — a genuinely bullish headline. Yet in relative terms the index of 84 says that sits near the top of its trailing-year range — historically a crowded, one-sided long. Both statements are true; they simply answer different questions. Positioning extremes are best read as a measure of how one-sided — and therefore how vulnerable to a reversal — the crowd has become, not as a timing trigger.

Levels describe where positioning is; changes describe where it is going. Over the most recent week, the net non-commercial position moved +8,254 contracts. Over the trailing four weeks it moved -23,632. Our reading of that combination is: large speculators extremely long (crowded bullish).

A four-week build in the same direction as the level suggests conviction; a four-week move against the level suggests the crowd is already unwinding an extreme. Neither is a signal to act — it is context for the volatility and macro readings elsewhere on the site.

For every speculative contract, someone takes the other side. Commercials — miners, refiners, and other hedgers of physical gold — are net short at -312,658 this week. Because they are hedging rather than speculating, commercials are typically positioned opposite the speculative crowd, and their net tends to sit near the mirror image of the non-commercial net. Read the two together: the speculative side tells you about sentiment and crowding; the commercial side is its structural counterweight.

The record of net non-commercial positioning, the 52-week COT Index, and open interest in the weeks leading up to this report — newest first.

156 weeks of history available · source: CFTC Commitment of Traders (Legacy, Futures Only) — COMEX Gold

The Commitment of Traders report is a positioning tool, not a price tool. It tells you how crowded a side of the market has become; it does not tell you when — or whether — that crowding will unwind. Extremes can persist for months, and a “washed-out” reading can always get more washed out. Use it the way we do: as one honest input into a wider picture, weighed against volatility regime and macro context, never as a standalone trigger.

Everything on this page is a sentiment read derived transparently from public CFTC data. It is educational analytics — not financial advice, not a recommendation, and not a solicitation to trade. Read exactly how these numbers are built, and everything they are not, in our methodology.