Positioning
COT extremes: why percentiles beat raw net positions
Position sizes drift over the years. Comparing today’s net position with its own range tells you far more than the raw number.
Raw numbers are not comparable
Open interest in currency futures changes over time, and each currency has its own typical size. A net long that was huge in one year may be ordinary a few years later. Comparing raw contract counts across time or across currencies is therefore misleading.
Percentile: where today sits in its own range
A percentile ranks today’s net position against a lookback window of the same currency. 90 means the position is longer than in 90 % of the weeks in that window; 10 means shorter than in 90 %. The scale is the same for every currency, so extremes become comparable.
- Choose the window deliberately. A short window reacts quickly but calls many readings extreme. A long window is stable but slow to recognise a new normal.
- Watch the change, too. An extreme that is being unwound week after week is a different situation from one still being built.
What an extreme does and does not say
An extreme says the trade is crowded and the market is sensitive to news against it. It does not say when that news will arrive. Many extremes simply fade sideways.
TakeawayPercentiles make extremes comparable. They measure vulnerability, not timing.
In Drift Engine
The COT Index uses the same idea in a simpler form. It places the net position between the lowest (0) and the highest (100) value in the chosen window — a range scale, not a percentile rank. One extreme week in the window therefore compresses everything else, which is why the window is yours to choose. The free COT data pages show the percentile.
Educational content. Not investment advice and not a recommendation to buy or sell any financial instrument. Trading currencies involves a high risk of loss.