Friday afternoon, US Eastern time, the Commodity Futures Trading Commission publishes a census of open positions in the American futures market. The positions it counts were held at Tuesday's close. That gap is the first thing to understand about the Commitments of Traders report: by the time you read it, the market has had three sessions to move against every number on the page.
Traders still read it, and they are right to. Nothing else in public data tells you how a category of large accounts was leaning, contract by contract, week after week, going back decades. The skill is in knowing what the categories mean and what the report cannot see.
What the report actually counts
COT covers open interest in US listed futures and options on futures. For currencies that means the CME contracts on the euro, sterling, yen, Swiss franc, Australian dollar and Canadian dollar, each quoted against the US dollar. For metals it means COMEX, for energy NYMEX, and for equity index exposure the CME index complex.
Spot forex is absent, and it cannot be otherwise. Spot is an over the counter market with no central clearing house counting anyone's book, so no regulator can publish a position census for it. When a report says speculators are net long the euro, it means net long euro futures, a market far smaller than spot but populated by many of the same macro accounts. It is a proxy. Treat it as one.
Only accounts holding more than a contract specific reporting level get broken out. Everything below sits in a single non reportable bucket, which is where most retail futures activity lands. So the report is a picture of the large end of the market by construction.
Three reports, three sets of categories
The naming trips people up because the CFTC publishes several versions of the same underlying data.
| Report | Covers | Main categories |
|---|---|---|
| Legacy | All markets, the oldest format | Commercial, non commercial, non reportable |
| Disaggregated | Physical commodities, including gold and oil | Producer or merchant or processor or user, swap dealers, managed money, other reportables |
| Traders in Financial Futures | Currencies, rates, equity indices | Dealer or intermediary, asset manager or institutional, leveraged funds, other reportables |
For currency work, the financial futures version is the one worth opening. Leveraged funds is the closest thing to a hedge fund and managed futures bucket, the accounts that trade macro views and stop out when they are wrong. Asset manager or institutional is slower money: pension funds, insurers, mutual funds hedging or holding foreign assets. Dealer or intermediary is the sell side, warehousing risk from client flow rather than expressing a view.
For gold, the disaggregated report is the useful one, and managed money is the speculative line most people track against price.
Reading the numbers without a crystal ball
The headline figure is the net position: longs minus shorts for a category, in contracts. On its own the raw number means little, because open interest grows over the years and a net long of 100,000 contracts in one decade is not the same crowd as in another. Two habits fix that.
First, watch the weekly change rather than the level. A category adding 20 percent to a net short in one week is a different signal from a category sitting quietly on the same book for two months.
Second, normalise. The common approach is a positioning index that ranks this week's net position against its own range over the past two or three years, so you get a reading between 0 and 100 instead of a contract count. That is how COT gets used as a contrarian sentiment measure: when leveraged funds are at the top of their own three year range, the pool of buyers left to join them is thin, and any shock has more stops to run through. It says nothing about timing. Positioning can sit at an extreme for months while price keeps trending, which is exactly why it belongs alongside other sentiment tools rather than in place of price analysis.
The single most damaging misreading is the idea that commercials are the smart money and speculators are the dumb money. Commercial and dealer accounts hold offsetting exposure in a physical inventory or a client book that the report never shows. Their futures leg can look catastrophically wrong while the whole position is flat.
Where the report breaks
The three day lag matters most in the weeks that matter most. A Wednesday rate decision, an emergency statement, a surprise inflation print: any of these can turn Tuesday's book into history before you see it. That is why COT reads best against a slow calendar and worst around central bank meetings, when the fastest accounts reposition in hours.
Other limits worth holding in mind. Spread positions, where a trader is long one delivery month and short another, are reported separately and are not a directional view. The futures only and the combined futures and options versions of each report will not agree, so pick one and stay with it. Concentration ratios published alongside the main table sometimes reveal that a category's whole net position is a handful of accounts. And publication has been suspended during US government shutdowns, leaving multi week gaps that a positioning index cannot interpolate away.
A workflow that fits a normal week
Pull the file once a week, always on the same day, and record two things per instrument: the net position for the speculative category and where it sits in its own multi year range. Put both next to the week's price behaviour in your trading journal. After three months you will have your own history of how positioning extremes resolved on the instruments you actually trade, which is worth more than any generic rule about what an extreme means.
Use it to set expectations, not entries. If leveraged funds are stretched short sterling and price stops making new lows, that is a reason to take a short setup less seriously and to watch for a squeeze. It is not a reason to buy. The entry still has to come from structure and risk, and the position still has to be sized as though the view is wrong, because leveraged trading carries a high risk of loss whatever the positioning data shows.
"COT is a weekly photograph, not a live feed. If your trade only works while Tuesday's book is still intact, you are not trading positioning, you are guessing."
— Alex Onta, Executive Director, SINGUARD
Key Takeaways
- The report lands on Friday but shows Tuesday's positions, so treat every figure as three sessions old.
- Use the Traders in Financial Futures version for currencies and the disaggregated version for gold and oil.
- Rank net positions against their own multi year range instead of reading raw contract counts.
- Positioning extremes describe crowding, never timing, and can persist for months while price trends.
Frequently Asked Questions
When is the COT report released?
The CFTC publishes the Commitments of Traders report on Friday afternoon, US Eastern time, covering open positions as of the close of the previous Tuesday. Publication has been delayed in the past during US government shutdowns, and holiday weeks can shift the schedule.
Does the COT report cover spot forex?
No. It counts positions in US listed futures and options on futures, mainly CME currency contracts, COMEX metals and NYMEX energy. Spot forex is traded over the counter with no central clearing house counting positions, so futures data is used as a proxy for the wider market.
Are commercials always right and speculators always wrong?
That is a myth. Commercial accounts hold offsetting exposure in the physical or banking book that the report never shows, so their futures leg can look wrong while the combined position is flat. Treat every category as one input, never as a verdict on price direction.