forex-scalping.com

Trading the release itself means fighting wide spreads and slippage. Waiting for the first pullback trades the same information with much better execution.

A common framework:

• Wait 3–10 minutes after the release until spreads return to near-normal levels.
• Measure the spike from the pre-release price to its extreme.
• Look for a pullback that holds somewhere between roughly 38% and 62% of the spike.
• Enter in the spike's direction on a break of the pullback's structure, stop beyond the pullback extreme.

Where it fails:

• Mixed data, where the first move reverses completely.
• A second release or a speech shortly after.
• Spikes caused by a single number that is revised or contradicted in the details.

Checking the bond market reaction helps: if yields moved in line with the currency, continuation is more likely.

How long do you wait after a release before looking for an entry?

FXS Prop Desk wrote: Wait 3–10 minutes after the release until spreads return to near-normal levels.

The 3 to 10 minute range is wide enough that it's worth tying to something observable rather than the clock. Spread is one, but you could also wait for the first M1 candle after the spike that closes inside the previous candle, as a rough sign the first burst of orders has run out. The other thing is that the 38 to 62% zone can get tagged within the first minute if the opening tick overshoots and snaps back, so measuring the spike from the extreme of the first full minute rather than the very first print gives cleaner levels. Mixed data is the real killer though, as you say. If the details contradict the headline, standing aside usually beats guessing which part the market will pick.

The part I would add is a time limit on the pullback. If price is still drifting in the 38 to 62 percent zone twenty minutes after the release, the news has been absorbed and I stop treating it as a news trade.
The chart is a made up EURUSD M1 example of the version I do take. Price sits at 1.1650 before a 13:30 release, the spike tops at 1.1688 in the second minute, and the pullback holds at 1.1669, half of the 38 pip move.

Entry is the break of the small lower high inside the pullback at 1.1675, with the stop under the pullback low at 1.1666. That risks 9 pips for a retest of 1.1688, only about 1.4R, so I would not take it if the pullback had gone past 62 percent.

Fairman's inside candle idea fits well here, since the first M1 close inside the prior candle often marks the start of the pullback. In this example it came at 13:32, two minutes after the print.

On the yields check, I watch the 2 year rather than the 10 year on US data. A EURUSD spike with a flat 2 year is the one I distrust most.