forex-scalping.com

The Commitment of Traders report shows futures positioning as of Tuesday, but it is released on Friday afternoon. By the time most traders act on it on Monday, the data is almost a week old.

What remains useful despite the lag:

• Extreme positioning: when speculators hold record long or short positions, a reversal or squeeze becomes more likely over weeks.
• Trends in positioning: steady increases or decreases over several weeks.
• Divergence: price moving one way while positioning moves the other.

Limitations:

• Futures are only part of the FX market.
• Positioning can stay extreme for a long time.
• A week of lag can miss a sharp reversal.

For scalpers, COT is background context, not a trigger.

Do you use COT data, and how much weight do you give it?

FXS Prop Desk wrote: Futures are only part of the FX market.

On top of that, the categories don't mean the same thing across the legacy report and the TFF report, so two people quoting COT positioning can be reading different tables. For extremes, it helps to use a percentile or z-score of net positioning over the last two or three years rather than raw contracts, since open interest in each currency future changes over time. And dollar positioning is spread across several contracts, so a big net long in 6E doesn't automatically mean a big net short USD overall. As background for a scalper I agree, it's more of a which side is crowded if something surprises read than anything timing related.

Fairman's percentile point is the one I lean on. I keep a three year percentile of leveraged fund net positioning from the TFF table for yen, euro and sterling, and I only pay attention when a reading sits above the 90th or below the 10th.

Even then it changes nothing about entries. What it changes is which surprise I respect: if funds are crowded short yen and a hawkish headline hits, I will not fade the first yen spike on USDJPY.

On the lag itself, I check how far price travelled between the Tuesday snapshot and the Friday release. If the pair has moved more than about one and a half daily ATRs net in those three days, I assume positioning has already shifted and mostly ignore the figure.

The release schedule is not guaranteed either. During last autumn's US government shutdown the reports stopped and then arrived as a backlog, which showed me how little my week actually depended on them.

So for me it is a Sunday background note, two lines at most. Does anyone here use the week on week change rather than the level?