Not every data reaction continues. Some fade within minutes, others turn into trends for the rest of the day. A few signs help tell them apart.
Signs of a sustained move:
Bond yields moving in the same direction as the currency.
Related assets confirming (e.g. equities and gold reacting consistently).
Price holding beyond the first spike after 10–15 minutes.
Other data released at the same time pointing the same way.
Signs of a fading move:
Price returning inside the pre-release range.
Mixed data components or large revisions.
No reaction in yields.
A useful rule: judge the move after the first 15 minutes rather than the first 15 seconds. Spreads are back to normal, the initial algorithmic reaction is done, and the market's interpretation is clearer.
How long do you wait before deciding whether a data move will continue?
FXS Prop Desk wrote: judge the move after the first 15 minutes rather than the first 15 seconds.
Makes sense as a default, with one exception worth flagging. On releases where other data lands at the same time, like Canadian and US jobs together, or a Fed speaker scheduled half an hour later, the 15 minute read can get overwritten before it means anything. So the window should probably end before the next scheduled event rather than at a fixed time. On the yields point, it helps to match the right tenor. For data that changes near term rate expectations, the 2-year is the better confirmation, and the 10-year can stay flat while the front end moves a lot. Equities are noisier as a check because they react to growth and rates at the same time.
FXS Prop Desk, waiting fifteen minutes before calling a data move sustained is my default too. The first fifteen seconds are mostly spreads and algorithms, and I have faded too many spikes that were already dead by minute ten.
Fairman's exception on stacked releases matters. When Canadian and US jobs land together, or a Fed speaker follows inside the half hour, I end the read before the next event rather than at a fixed clock.
Yield confirmation for me means the two year when the print is about near term rates. The ten year can sit still while the front end and the dollar move hard, so I stopped using it as a veto.
I also want price holding outside the pre release range after that wait. A return inside is enough for me to treat the move as a fade candidate or a skip.
Related assets help only when they agree cleanly. How often do you still take the first spike, or have you moved fully to the delayed read?
