Long, short and net: what they tell you about price
Net shows which way Non-Comm leans. Long and short show how it got there. We tested all three against price in 42 markets. None of them tells you on its own which way price goes next. Here is what they do tell you.
Net is the lean; long and short are how it got there
Net is Non-Comm long minus Non-Comm short. On Tuesday 29 September 2026, Gold's Non-Comm long was 249,736 contracts and short was 31,104, a net of 218,632 long.
| Gold, 29 Sep 2026 | Contracts | Change on the week |
|---|---|---|
| Non-Comm long | 249,736 | −4,246 |
| Non-Comm short | 31,104 | +2,975 |
| Net | 218,632 | −7,221 |
Net fell 7,221. The same fall could be funds taking profit (longs cut), fresh selling (shorts added), or, as here, a bit of both. Net alone can't tell them apart, so read all three together.
The weekly change describes the week just gone
Every Non-Comm report in 42 markets, September 2009 to June 2026, against Tuesday-to-Tuesday closes. Each figure is the share of weeks price was up.
| Non-Comm that week | Weeks | Same week | Next week | Next 4 weeks | Next 13 weeks |
|---|---|---|---|---|---|
| Net rose | 17,901 | 64% | 51% | 52% | 52% |
| Net fell | 17,919 | 39% | 52% | 53% | 53% |
| Longs added and shorts cut | 8,203 | 68% | 51% | 51% | 52% |
| Longs cut and shorts added | 8,187 | 35% | 52% | 53% | 53% |
| Every week | 35,867 | 51% | 51% | 52% | 53% |
Funds move with price in the week itself: net and price agree about two weeks in three. From the week after, every row is back to the normal rate. The same holds in each of the 10 core markets: the next-week figure stays within 3 points of normal. The report is context, not timing.
Extremes: no edge on average
Two ways to measure an extreme: each side's % of record (its own highest level so far; MarketsLens rings 80% or more), and net against its range so far. Both use only the history known at the time. Ranking against later years too makes results look far better than they were.
| All 42 markets | Weeks | Separate runs | Up 13 weeks later |
|---|---|---|---|
| Every week (normal rate) | 22,117 | — | 50% |
| Non-Comm long at 80%+ of record | 6,143 | 736 | 53% |
| Non-Comm short at 80%+ of record | 6,035 | 809 | 53% |
| Net in the top 10% of its range so far | 1,213 | 253 | 50% |
| Net in the bottom 10% of its range so far | 2,195 | 420 | 49% |
A run is a stretch of weeks in a row at the extreme. Weeks next to each other share most of the same price move, so runs are the separate examples that count.
The core markets
Net at the edge of its range so far, September 2012 to June 2026. "Up" means the futures price rose: for the Canadian dollar, Swiss franc and yen, the currency strengthening against the US dollar. "Normal" is the share of weeks price was higher 13 weeks later. "Then up" counts each run once (the first week net reached the edge), with the 95% range in brackets: with this few runs, the true rate could be anywhere in it. A result is coloured only when its whole range sits away from normal.
| Market | Normal | Most short: runs | Then up | Most long: runs | Then up |
|---|---|---|---|---|---|
| Australian dollar | 44% | 16 | 50% (28–72) | 5 | 20% (4–62) |
| New Zealand dollar | 46% | 16 | 62% (39–82) | 7 | 29% (8–64) |
| Swiss franc | 48% | 0 | — | 13 | 15% (4–42) |
| British pound | 51% | 11 | 64% (35–85) | 7 | 57% (25–84) |
| Canadian dollar | 60% | 1 | too few | 11 | 55% (28–79) |
| Gold | 58% | 9 | 44% (19–73) | 12 | 50% (25–75) |
| Euro | 47% | 2 | too few | 10 | 70% (40–89) |
| Japanese yen | 61% | 7 | 43% (16–75) | 18 | 50% (29–71) |
| S&P 500 | 77% | 2 | too few | 9 | 56% (27–81) |
| Nasdaq 100 | 77% | 1 | too few | 8 | 88% (53–98) |
Only one result is clearly outside chance: when Swiss franc net was at its most long, the franc was higher 13 weeks later in just 2 of 13 runs (15%, range 4–42%) against 48% normally. Everything else, including results that look large such as Nasdaq 100 at 88% or the Australian dollar at 20%, rests on so few runs that its range takes in the normal rate. Net at an extreme is worth knowing as context; on its own it is not a reliable guide to what price does next.
How to use it
- Which way Non-Comm leans (net)
- Whether that came from buying, selling or closing out (long and short)
- How crowded each side is (% of record)
- Which side has the most to unwind
- Which way price goes next week
- When a crowded side will close out
- Where to act: that comes from chart location
Past results, not a forecast or a trade signal. Data: CFTC Legacy Commitments of Traders, futures only; daily closes from futures and spot prices. Tested 8 October 2026.
Or read how positioning works first.