Post-news reversion scalps: rules I actually share
Posted: Tue Oct 06, 2026 4:41 pm
People asked for my post-news reversion rules after a thread last month, so here they are, as written on my card.
1. Only on releases where my own log shows the first spike reverting more often than not. For me that is a short list of second-tier data, not NFP or CPI.
2. The spike must reach a level marked before the release: a session high or low, a previous day's level, or a round number that mattered earlier.
3. Wait for spread to return to near normal. Usually 30 seconds to two minutes.
4. Entry needs a rejection candle at the level on M1, closing back inside.
5. Stop beyond the spike extreme plus a buffer. If that stop is more than twice my normal scalp stop, no trade.
6. Target is the pre-release price or halfway back, whichever is closer.
7. One attempt per release. If stopped out, I do not try again in either direction.
8. Half size if the figure was a large surprise, or skip completely if it was extreme.
The rule I break most is number seven, wanting to take a second shot because "now it really has to revert". The log says second attempts lose more often than first attempts.
If you try these, change the list of releases in rule one to match your own data. Mine will not fit your pairs and hours.
1. Only on releases where my own log shows the first spike reverting more often than not. For me that is a short list of second-tier data, not NFP or CPI.
2. The spike must reach a level marked before the release: a session high or low, a previous day's level, or a round number that mattered earlier.
3. Wait for spread to return to near normal. Usually 30 seconds to two minutes.
4. Entry needs a rejection candle at the level on M1, closing back inside.
5. Stop beyond the spike extreme plus a buffer. If that stop is more than twice my normal scalp stop, no trade.
6. Target is the pre-release price or halfway back, whichever is closer.
7. One attempt per release. If stopped out, I do not try again in either direction.
8. Half size if the figure was a large surprise, or skip completely if it was extreme.
The rule I break most is number seven, wanting to take a second shot because "now it really has to revert". The log says second attempts lose more often than first attempts.
If you try these, change the list of releases in rule one to match your own data. Mine will not fit your pairs and hours.