Trader bets
This condition measures how heavily traders are already betting on gold. Big speculators in the US futures market report their bets every week to the government. When their long bets are unusually crowded, the market often gets fragile. When they have been washed out, one source of pressure to sell is gone. We read the extremes here, not the everyday drift.
- Share of the score
- 10%
- Share used today
- 11.1%
- Source
- US regulator's weekly trader-bets report (CFTC), professional bets as a share of all open bets
- How often it updates
- weekly
- Date of the newest figure
- 2026-09-15
- Newest figure
- 23.7352 % of all open trader bets
- How fresh it is
- Up to date. The newest figure is 5 days old, and we stop counting this part once it passes 10 days. An old figure left counting would quietly drag the score toward whatever was true last month.
- How strong the evidence is
- weak to moderate. It says most when nearly every trader bets the same way
What we look at
The weekly trader-bets report (Commitments of Traders, detailed version) from the US regulator, the CFTC. We use the professional funds' bets that the price rises, as a share of all open bets.
Why it matters if you're buying gold
Most of the trading that sets the gold price is not people buying coins. It is professionals placing bets, and a United States regulator publishes a count of those bets every week. When nearly all of them are betting the price will rise, something awkward follows. Everyone who wanted to buy has already bought. Nobody is left to push the price higher, and one small piece of bad news sends that same crowd for the exit at once. When those bets have been cleared out instead, a source of selling pressure is simply gone. A boat is steady while people stand all over it. It rolls over when they crowd to one side. We read this only at the extremes. In between it tells you very little, which is why it carries a small weight.
What this cannot tell you
Weak to moderate. This condition is deliberately small in weight and we read it at the extremes. Between them, its signal is weak. We include it because crowded trader bets change the risk you take, not because they predict the price. We publish the lore caveat: the academic version of this rule reads the miners and dealers, not the professional funds.
Today's reading in plain words
Today this condition reads 25/100, below the historical middle.
Notes: The bet-against-the-crowd read is applied across the whole rank, but the actual signal concentrates at the extremes.
How the score is worked out
We rank the professionals' rise-bets, as a share of all open bets, against their own past, from 0 to 100. Then we flip it: washed-out bets give a high score, crowded bets give a low score. The real signal sits at the extremes.
Which way it points. A low crowding rank gives a high score. The signal concentrates at the tails. The middle carries little information.
What the evidence says
- Sanders, Irwin and Merrin (2009): speculator bets in commodity futures do not forecast price on average. Speculators tend to follow trends, not fight them.
- Wang (2003): the classic bet-against-the-crowd signal in this data sits with the miners and dealers laying off risk, not with the professional funds. Practitioner lore has flipped this. We use the professional funds because their data is the cleanest and deepest, and we disclose the academic caveat.
- Chen and Mo (2023): professional bets on gold futures switch between market phases. The paper explains the mechanism. It does not prove forecasting power on years the model never saw.
If you are selling
If you are selling, this one flips cleanly. When speculator bets are unusually crowded, sharp drops have often followed. It is the condition with the strongest sell-side reading at extremes.
Source: US Commodity Futures Trading Commission, Commitments of Traders (detailed version). See the methodology page for the full formula, and the provenance page for where each number comes from and how fresh it is.