Details
What the report contains
Every week the CFTC publishes positions as of Tuesday, usually released on Friday afternoon US Eastern time. For each market it shows total open interest and the long and short positions of each trader category.
There are three formats, and they slice traders differently:
- Legacy: commercial and noncommercial (plus nonreportable small traders).
- Disaggregated (commodities): producer/merchant, swap dealers, managed money, other reportables.
- Traders in Financial Futures (TFF) (equity indices, rates, currencies): dealer/intermediary, asset manager/institutional, leveraged funds, other reportables.
For an equity index future the TFF version is the useful one. Asset managers are pension funds and mutual funds, often long to hedge or replicate an index. Leveraged funds are hedge funds and CTAs, the group most traders watch for speculative positioning.
Worked example with real rows
These three rows are from our free CFTC sample, E-mini S&P 500, the last three weekly reports:
| symbol | date | open_interest | asset_manager_long | asset_manager_short | leveraged_funds_long | leveraged_funds_short |
|---|---|---|---|---|---|---|
ES |
2026-10-27 | 1955711 | 1111055 | 187606 | 102079 | 502683 |
ES |
2026-11-03 | 1986532 | 1111721 | 183137 | 125269 | 497587 |
ES |
2026-11-10 | 2017687 | 1115756 | 201626 | 145755 | 448815 |
Source: CFTC
Net position is long minus short. For leveraged funds that gives -400,604 contracts on 27 October, -372,318 on 3 November and -303,060 on 10 November. They are still heavily net short, but they cut that short by about 97,500 contracts in two weeks. Asset managers sat near +923,000 to +928,000 net long and then dipped to +914,130.
What does that mean? A big net short from leveraged funds in ES is often the other side of basis trades against cash equities, not a pure bearish bet. So the level matters less than the change, and the change matters most when it reaches an extreme compared with the past few years.
Details
Compute net positions and a 3-year percentile
"""Example: net positioning from TFF rows."""
import pandas as pd
cot = pd.read_csv("cftc-traders-in-financial-futures.csv", parse_dates=["date"])
es = cot[cot["symbol"] == "ES"].sort_values("date").copy()
es["lev_net"] = es["leveraged_funds_long"] - es["leveraged_funds_short"]
es["am_net"] = es["asset_manager_long"] - es["asset_manager_short"]
es["lev_net_pct_oi"] = 100 * es["lev_net"] / es["open_interest"]
es["lev_net_change"] = es["lev_net"].diff()
print(es[["date", "lev_net", "lev_net_pct_oi", "lev_net_change", "am_net"]])
if len(es) >= 156: # needs 3 years of weekly reports
es["lev_pctile_3y"] = es["lev_net"].rolling(156).apply(
lambda x: (x[:-1] < x[-1]).mean() * 100, raw=True)
On the sample, lev_net_pct_oi moves from about -20.5% to -15.0% of open interest. The free sample covers only three reports, so the percentile block is skipped. With full history you would read a value near 0 or 100 as an extreme, and that is where COT signals have the most to say.
Our FUT-POSITIONING schema includes leveraged_funds_net and net_position_percentile_3y as derived fields, so you can skip this step once the API is available.
Watch out
Common mistakes
Treating it as live data. Positions are as of Tuesday and released Friday. By Monday's open the snapshot is six days old. Use it for weekly or slower decisions.
Reading the level instead of the change. Some categories are structurally long or short because of hedging. A permanent net short is not a signal; a sharp move toward a multi-year extreme can be.
Mixing report formats. "Noncommercial" in the legacy report and "leveraged funds" in TFF are not the same group. Pick one format per market and stick to it across your whole history.
Ignoring contract size. ES and MES (Micro E-mini S&P 500) are separate rows with different multipliers. Adding their contract counts together overstates the micro positions tenfold.
Details
Get the data
Free CFTC positioning sample. 294 rows: the last 3 weekly Traders in Financial Futures reports for every financial future, in our schema with
symbol,cftc_codeand every category's long and short.
- Download the CFTC TFF sample (CSV and JSON)
- Source: CFTC
- Full history for legacy, disaggregated and TFF reports: COT report data, planned for the Free plan in early access.
FAQ
Frequently asked questions
How do you read the COT report?
Pick the right format for the market (TFF for financial futures, disaggregated for commodities), compute each group's net position as long minus short, and watch the weekly change. Compare the current net to the last few years to see whether it is extreme. Remember the data is as of the previous Tuesday.
When is the COT report released?
The CFTC normally releases it on Friday afternoon, Eastern time, with positions as of the Tuesday of that week. Holiday weeks can shift the release by a day or more.
Which group should I watch in the COT report?
For financial futures, most traders watch leveraged funds, since they include hedge funds and CTAs. In commodities, managed money plays the same role in the disaggregated report. Commercial hedgers are useful too, but their positions mostly reflect hedging needs.
Is COT data free?
Yes, the CFTC publishes it as public data. What you pay vendors for is clean history in one format, consistent symbols and derived fields like net positions and percentiles.
Keep reading