Gold signals during high-impact news: our rules for CPI, NFP and FOMC days

By Emanuele Calcina · Published 12 September 2026 · Examples from the 4 September NFP and the 10 and 11 September PPI and CPI releases

One hour on Friday 11 September covered $100 of XAUUSD, from 4333.0 to 4432.6 on the front-month hourly bar (Yahoo Finance, GC=F), and the CPI print that caused it matched the forecast to the decimal: 0.4% on the month, 3.4% on the year (CNBC). The number was not a surprise. The candle was. That gap between the headline and the tape is the entire problem with trading gold through high-impact news, and it is why the channel runs a fixed set of rules on those days instead of opinions.

Here are the rules. They apply to the free channel and the VIP channel alike, and they are the same rules every week, so you can check them against the channel history instead of taking our word for it.

Which releases count

Three US releases move gold more than everything else combined, and one of them is a meeting rather than a number.

Everything else, from retail sales to the Bank of England, gets the lighter version of the rules: we avoid opening a position in the minutes around it, and that is usually enough. The four above get the full set.

Rule 1: the calendar is read before the chart

Every session starts with the release times for that day written down in New York time, London time and the trader's own time. This sounds too basic to be a rule. It is a rule because the most expensive mistake in gold is a good setup entered nine minutes before a number the trader forgot was coming. The setup was fine. The timing killed it. If you follow signals from any provider, know the calendar yourself; do not outsource that part.

Rule 2: no new entries in the run-up window

The channel does not open fresh positions in the last stretch before a tier-one release. For CPI, NFP and PPI that window runs into the 8:30 ET print; for FOMC it runs into the 2:00 PM statement. The logic is that a position opened in that window has no edge, only exposure. The chart before a number is a chart of people waiting, and a level that holds at 8:15 means nothing at 8:31.

The exception is a trade already running with a stop already outside the event range. That trade stays. It was taken on structure, and structure does not stop existing because a number is due. It gets managed, which is Rule 5.

Rule 3: the first move is not the move

Friday's CPI hour is the example we will be using for a while. Print at 8:30, gold drops to 4333.0, reverses, prints 4432.6, closes the hour at 4430.3. Anyone who sold the first candle was stopped out by 8:50 and anyone who bought it was right for the wrong reason. The first move after a release is the algorithms reading one figure, usually core month on month. The second move is the market deciding whether that figure was already priced. The second move is the one with information in it.

So the channel does not chase the first candle. It waits for the first 15-minute candle after the print to close, reads where price sits relative to the pre-release range, and only then decides whether there is a signal. Sometimes there is not, and the channel says nothing for an hour. That silence is a decision, not an absence.

Rule 4: stops outside the level, size cut to match

A stop inside the expected event range is a fee, not a stop. On an ordinary hour a $15 stop on gold might be reasonable. On a CPI hour that just covered $100 it is a certain exit at the worst price. So when the expected range is wider, the stop goes outside the level that would actually prove the trade wrong, and the position size drops so the dollar risk stays the same as on any other day. Wider stop, smaller size, same risk. It is the only risk rule the channel has, and event days are when it becomes visible.

This is also where spreads matter. Brokers widen the XAUUSD spread in the seconds around a release, and a stop sitting a few dollars from price can be triggered by the spread alone before the market has moved. A stop outside the level has room for that.

Rule 5: open positions get managed, not abandoned

If a trade is running into a release, the channel posts what to do with it before the number, not after. That is usually one of: move the stop to a level that survives the range, take part of the position off, or close it flat because the reward left does not justify the event risk. What it never is: hold and hope. Hope is not a management instruction and the channel does not post it.

Rule 6: the reasoning is written down

Since 9 September every VIP entry carries a two or three sentence note under it explaining why it is a buy or a sell, in terms of the structure on the daily, 4-hour and 1-hour charts and the levels in play. On a news day that note does real work. "A long inside a range, capped under the 4-hour 50 EMA" tells you the target is the range top and the trade dies on a close through the bottom. "A breakout retest" tells you the opposite. Two trades with the same entry price get managed differently through the same release, and the note is how you know which one you are in.

Rule 7: the losers stay in the channel

Event days produce losers. Any provider who says otherwise has not traded many. The channel history is unedited, so every stopped-out call from every CPI, NFP and FOMC day is still there to read, and the weekly figures on the results page include the weeks those days made ugly. The method behind each number is on the numbers page; the lifetime figure across the tracked record is 84%. If you are evaluating any signal service, ask to see their worst news day. If they cannot show it, that is your answer.

Putting the rules on a clock

Next week has all of this in one place: FOMC with a dot plot on Wednesday 16 September, the Bank of England on Thursday, the Bank of Japan on Friday. The full calendar and the levels we are carrying into it are in the week-ahead piece. The easiest way to see whether these rules are real is to sit in the free channel on Wednesday afternoon and watch what gets posted, and what does not.

More on sessions, stop placement and how to read a signal in the guides hub.

Questions people ask

Should you trade gold signals during CPI or NFP?

You can follow them, but the rules change. No new entry in the run-up window, no chasing the first candle, stops outside the level with a smaller position, and the first move after the print is treated as noise until the next candle confirms it. On 11 September the CPI hour ran from 4333 to 4432 on an in-line print; a normal-sized stop inside that range was a donation.

How long before news should you stop opening trades?

We stop opening fresh positions in the last stretch before a tier-one release and do not resume until the first move has finished, which for CPI and NFP usually means the first 15-minute candle after 8:30 ET has closed. For FOMC the window runs from the 2:00 PM statement through the end of the 2:30 press conference.

Does the free channel post signals on news days?

Yes, the same one or two calls a day with entry, stop loss and take profit, and it follows the same timing rules as VIP. It is the best place to watch the process on a live event day before paying for anything.

What does the VIP channel add on news days?

Around five signals a day instead of one or two, and since 9 September a short note under every entry explaining why it is a buy or a sell, in terms of the chart structure and levels. On an event day that note tells you whether a trade is a range play or a breakout attempt, which decides how it gets managed when the number drops.

Where are the results for news weeks?

On the results page, week by week, losing weeks included, with the counting method on the numbers page. The lifetime figure across the tracked record is 84%.

Related reading

• Gold this week, 7 to 11 Sep 2026: what moved XAUUSD
• Gold week ahead, 14 to 18 Sep 2026: FOMC, BoE, BoJ and how we trade it
• XAUUSD trading hours and the best time to trade gold
• How gold signals perform in volatile markets
• All gold trading guides

Risk disclaimer. Trading gold on margin carries a high risk of loss and is not suitable for everyone. Past performance, including any figure on this page, does not predict future results. Nothing here is financial advice. Market data and third-party figures were read on the dates stated and may have been revised since.