Cot Report Gold Futures Positioning Data: Dataset Structure and Field Coverage

Every week the U.S. Commodity Futures Trading Commission counts who holds what in exchange-traded gold futures, then publishes the tally as the CFTC Commitments of Traders (COT) Reports. The snapshot reflects Tuesday's open interest and lands each Friday at 3:30 pm Eastern Time, shifted by holidays. One row covers one contract market per report week, split by maturity bucket and trader category.

Datadory catalogs this record under Diversified Capital Markets with a quality score of 9, one of only two official-API sources touching the gold pool's sentiment layer alongside SPDR's fund feed.

Four report families exist, and gold appears in more than one:

Each family generally ships a futures-only version and a combined futures-and-options version. If you want the cleanest read on speculative gold futures positioning, start Legacy futures-only; if you want to separate swap-dealer hedging from managed-money bets, use the Disaggregated combined file.

Field definitions are verified at confidence level "verified", so you can build parsers against them rather than guessing column meanings. The core columns:| Field | What it carries | | --- | --- | | market_and_exchange_names | Full contract and exchange name, e.g. "WHEAT-SRW - CHICAGO BOARD OF TRADE" | | report_date_as_yyyy_mm_dd | Date of the Tuesday open-interest snapshot | | cftc_contract_market_code / cftc_market_code | Six-character contract code plus exchange mnemonic (CBT, CME, NYM, IFL) | | open_interest_all | Total open interest across all maturities | | noncomm_positions_long_all / short_all / spread_all | Legacy: speculative longs, shorts and spreading | | comm_positions_long_all / short_all | Legacy: commercial hedger positions | | nonrept_positions_long_all / short_all | Positions below reporting thresholds | | change_in_open_interest_all | Week-over-week change in total open interest | | pct_of_open_interest_all | Each category's share of total open interest, in percent | | traders_tot_all | Distinct reporting traders; suppressed where fewer than four are active |

Long-format Legacy files repeat these categories across old-crop and other-maturity buckets (_old / _other column sets), which is why the long report runs wider than the short one. A real Legacy futures-only row from the 13 September 2022 report shows how the columns fit together: open_interest_all of 287,046 contracts against comm_positions_long_all of 119,219 and traders_tot_all of 315.

Weekly coverage starts 15 January 1986 for Legacy, giving nearly four decades of net speculative length in gold futures. The Disaggregated and TFF formats begin 13 June 2006, and the supplemental CIT series starts 3 January 2006. That makes COT one of the longest sentiment time series available for free anywhere in the metals space — longer than GLD's fund-level history and deeper than most retail charting platforms expose.

Pairing the two is the standard workflow: price tells you direction, COT tells you crowding.

Positioning reads best against corroborating flows. SPDR Gold Shares (GLD) Fund Data publishes daily NAV, tonnes held and premium/discount for GLD and GLDM — 5,676 daily rows since 18 November 2004 — so you can test whether speculative futures length moves with physical ETF tonnage. World Gold Council Goldhub adds monthly ETF holdings-and-flows workbooks for the industry-wide view beyond one fund.

For modelers, Kaggle - Gold Price Data offers a ready-made 2,290-row daily panel (GLD closes with SPX, USO, SLV and EUR/USD, 2 January 2008 to 16 May 2018) that accepts a weekly COT feature joined on date — though note the sample window predates the deepest COT history by design.

Quantitative researchers treating net speculative length as a sentiment factor top the list — Datadory tags investor-quants workflows for gold around exactly this kind of overlay. Market surveillance teams at exchanges and brokers use the same tables for their intended purpose: spotting concentration when few traders dominate a contract. Journalists cite the reports when gold rallies or sell-offs get attributed to "speculators."

The record scores 9 of 10 on Datadory's quality scale — the same score as SEC EDGAR mining filings and Goldhub — placing it in the top tier of the catalog, where the average dataset scores 7.81. Across all 1,744 datasets Datadory catalogs, 32.9% offer an official API; the COT record is one of them, which is why it anchors the programmatic side of the gold stack.

This page is the positioning chapter of Datadory's 17-dataset gold catalog. Continue with:

Cross-industry readers can see how the same CFTC record is framed over in diversified capital markets, including COT reports vs FINRA Data Portal and TRACE for a head-to-head access comparison.