RBOB
AT EXTREME · CROWDED LONG · WEEK 3FREE · NO ACCOUNT- 3y percentile
- 96.8
- COT index
- 68.7
- z-score (156w)
- 0.90
- weeks in zone
- 3
- report as of
- 2026-08-04
THIS WEEK’S REPORT, EXPLAINED
The crowd is stretched: managed money is long here, a bigger bet than in 97% of the last 3 years, week 3 in the crowded zone. This week they cut exposure (-0.0% of open interest).
in contracts: specs cut 6,247 longs · covered 2,285 shorts this week
COT INDEX · 3-YEAR WINDOW
0 = most short in 3 years · 100 = most long
69 / 100
Where the speculative net sits between its 3-year low and high. This one is counted in contracts, not in % of open interest: the low and the high come from this market’s own last 3 years, so you are only ever comparing this market against itself. Most sites publish a 26-week version, so this reads differently from theirs by design · the longer window is harder to reach and moves less on one loud week.
Where the speculators’ net position sits between its own 3-year low and high. Only this market against itself, never against another one.
managed money · net % of OI
23.0%
94,331 long · 36,024 short
3y percentile 97 · change -0.0% · the crowd this file grades
commercials · net % of OI
-23.1%
113,289 long · 183,588 short
change -0.4% · usually mirror the specs
small traders · net % of OI
3.9%
25,242 long · 13,250 short
change +0.7% · the rest of the market
WHO IS ON THE OTHER SIDE · DISAGGREGATED REPORT
Same week, same contracts, split finer. managed money is the number at the top of this page · these are the traders it is trading against.
producers & merchants
-31.1%
-1.2 wk · 1 of 100 in 3y
the physical trade, hedging real barrels and bushels
swap dealers
+8.0%
+0.7 wk · 91 of 100 in 3y
banks laying off the swaps they sold someone else
other reportables
-3.8%
-0.2 wk · 8 of 100 in 3y
large traders who are none of the above
small traders
+3.9%
+0.7 wk · 94 of 100 in 3y
everyone below the CFTC’s reporting threshold
same row as above · unchanged between the two reports
every long is somebody’s short, so these nets cancel out across the whole report · a display lens: the graded record for this market runs on managed money, from the CFTC’s disaggregated report · what these groups mean
Data is as of Tuesday, published Friday. The percentile compares this week’s speculative net position against every week of the last three years: above 95 or below 5 counts as crowded. The COT index above is scored the same way, against this market’s own history and never against another market.
Data is as of Tuesday, published Friday. Net % of OI = a group’s longs minus its shorts, as a share of all open contracts in the market. The long / short lines are raw contract counts from the futures-only report: the story. The percentile is the number measured on that %-of-OI scale, because a share of open interest survives being compared across markets and decades and a contract count does not. It compares today’s net position against every week of the last three years: above 95 or below 5 counts as crowded. The COT index and the 156-week z-score run on the contract count instead: each is scored against this market’s own 3-year history, so neither is ever compared to another market.
THE CHART. PRICE AND WHO HOLDS WHAT, SINCE 2003
Stretched: crowded long · week 3
The full file: turn odds, winner markers, and what marked the past turns.
The backtest: every past extreme, numbered on the chartFREE
What happened every time this market got crowded, the turns, the timing, what they paid.
For information and education only. Nothing on this site is investment, trading, legal or tax advice.
The Friday Letter
Every Friday 21:30 CET the new report lands and we regrade all 1,206 cases overnight. Confirm and your crowd-map snapshot arrives right away. Then every Friday: what moved, where the crowd is stretched, and which fades are running in public. And the moment we open one, you get the ping. Free.
Every Friday: where the crowd is stretched, what is running, what it cost. Free.
Confirmation by e-mail, one-click unsubscribe in every issue. Your address is never shared.