forex-scalping.com

Writing down the nearest calendar event for every entry

A column I've found useful: for each trade, the nearest scheduled high or medium impact event on the currencies involved, and how many minutes before or after the entry it was.

Example rows:
EURUSD long at 09:12 London. Nearest event: German Ifo at 09:00, so 12 minutes after.
GBPUSD short at 13:20. Nearest event: US retail sales at 13:30, so 10 minutes before.
USDJPY long at 02:00. Nearest event: none within two hours.

The reason is simple. Results near events can be very different from results in quiet periods, and without the tag they all mix together. With it you can filter trades within 15 minutes before an event, within 15 minutes after, and the rest.

What I'd expect to see, though everyone's numbers will differ: trades opened just before events have more variance and worse fills. Trades opened 15 to 60 minutes after events might do well because there's movement and direction. Trades in quiet periods might have smaller ranges and lower win rates for breakout systems.

To fill it in, I keep the week's calendar in a sheet with times in London time. A lookup finds the nearest event to each entry time. It takes a few minutes to set up and then works automatically.

If your journal has a few hundred trades, adding this retrospectively is possible using historical calendars from most calendar websites.

A simple and useful column. One more detail is the type of event, since a central bank speech behaves differently from a data release. Speeches can move markets for a long time and without a clear start, while data hits at a fixed second. Grouping trades near each type separately can show which events to avoid and which matter less. For the Asian session, events like Chinese data or Bank of Japan meetings are worth including, since they are easy to miss on calendars set to European time.

Event impact ratings differ between calendar sites. One site's medium is another's high. Pick one source and stick to it, otherwise the tags won't be consistent. I'd also record the event name, not just the impact level, so you can later check specific events like US CPI against your results. Using one source also makes it easier to backfill old trades, since you'll know which calendar archive to use. Some calendars change impact ratings over time, so saving a copy of the week's calendar at the start of the week preserves what you saw at the time.

The surprise size matters as well as timing. A trade opened 20 minutes after a CPI release that matched forecasts is in a different market from one opened after a big miss. Adding actual minus forecast for the nearest event, when available, lets you separate these. It's more work, but it explains a lot of outlier trades. Actual minus forecast can be noisy, since forecasts vary between sources. Using the same calendar for both keeps it consistent. Over time, you might find that only big surprises matter for your trades, and small misses can be ignored.

For multi-currency trades, like EURJPY, the nearest event could be on either currency. I'd log the nearest on each side separately. A EURJPY trade near a German release and far from any Japanese one is mostly a euro trade, which might explain the result better than treating EURJPY as one unit. The two sides might also interact. A EURJPY trade near both a German and a Japanese release is a very different situation from one near only one of them. Logging both makes it possible to see this. It doesn't need to be complicated, just two columns instead of one.

This also works as a pre-trade check. If the column for a planned trade would read 5 minutes before US data, that's a prompt to reconsider. Having the calendar in the same sheet as the journal makes it natural to look before entering, not just afterwards. A checklist item before each trade, check the calendar column, keeps the habit going. It also means fewer trades placed carelessly just before data. Over a few weeks, it might reduce the number of news-affected trades without any change to the strategy itself. The habit is the real gain here.