| Series | First | Latest | Min | Max |
|---|---|---|---|---|
| Duration crowding | 27.88 | 49.35 | 19.52 | 98.35 |
| 2Y | 60.58 | 40.61 | 0.77 | 99.00 |
| 5Y | 6.73 | 14.56 | 0.00 | 99.00 |
| 10Y | 16.35 | 29.12 | 0.00 | 99.00 |
| Ultra-10Y | 70.19 | 64.37 | 1.53 | 99.00 |
| 30Y | 27.88 | 98.08 | 0.00 | 99.00 |
extremity_i = 2 × |trailingPercentile260(NetOI_i) − 50|; crowding = equal-weight mean of extremity_iEach of the five Treasury futures tenors (2Y, 5Y, 10Y, Ultra-10Y, 30Y) is scored by the extremity of its Leveraged Funds Net/OI: extremity = 2 × |trailingPercentile260(NetOI) − 50|, where 0 marks a median reading, 100 a historical extreme. Crowding is the equal-weight mean of the five extremity scores, emitted when at least four tenors have valid scores.
As of 15 September 2026, the latest reading is 49.35 index (0-100). That is down 28.0% over the past year and below its long-run median of 59.20 index (0-100).
A high crowding reading means rate positioning is extended across tenors, a consensus position that is vulnerable to reversal if the rate narrative shifts. The index is structurally similar to the cross-asset crowding page but confines itself to rates, so it is the specific read a rates desk needs: is the duration trade crowded, or is the crowding elsewhere?
Each tenor has a different contract size and open-interest scale, so raw Net/OI ranges differ. The extremity score puts every tenor on the same 0–100 scale, and the mean gives each an equal voice: the 2Y and the 30Y count equally in the overall crowding reading.
Crowding is a relative measure against each tenor's own history, so an absolute small position that is extreme for its contract can read as crowded. The five-year trailing window shifts as history ages out. Treat the chart as a rate-positioning breadth gauge, not investment advice.
When the index reaches high levels, rate positioning is one-sided across tenors, the setup that produces outsized reversals when macro data or Fed communications break the consensus.
As of 15 September 2026, the latest reading is 49.35 index (0-100). That is down 28.0% over the past year and below its long-run median of 59.20 index (0-100).
This benchmark is built on weekly data. The page is refreshed when the source publishes new observations; the freshness block below the chart shows the exact data-through date and when PIER20 last fetched the file.
The chart is built from Duration crowding, 2Y, 5Y, 10Y, Ultra-10Y and 30Y, sourced from the CFTC Commitments of Traders report.
Cross-asset crowding averages across equities, FX, rates and commodities. It answers whether speculation is broadly crowded. This page isolates rates, so a crowded duration trade in a quiet equity/fx complex shows up here but not there.
One tested idea, through the evidence stack. No migration of any kind.