An economic calendar lists scheduled releases and policy events with time, prior values and, where available, forecasts and actual readings. Ratings are provider judgments, not guarantees that preparation prevents loss.

What an economic calendar actually is

An economic calendar organises scheduled macro releases, policy decisions and speeches. Depending on the provider and event, it can display time, currency or country, reference period, forecast, previous and actual values. Some events have no numerical consensus. Provider stars or colours describe expected importance, not a common volatility standard. For the relationship between releases and a trading hypothesis, see the fundamental analysis section in our library.

A calendar is preparation context, not a directional trade trigger or a complete risk map. Unscheduled news and operational failures can arise without an entry. Ordinary stops do not guarantee execution price, and a fixed delay after a release does not guarantee normal conditions. Record how the strategy handles open positions and pending orders; the wider workflow is in how to use the economic calendar.

Four free tools that actually deserve a tab

Compare the current interface, coverage, units, time-zone controls and access terms. Forex Factory shows a weekly table and export formats; Investing.com offers a Polish interface and country/category/importance filters; FXStreet links event rows to descriptions, history and commentary; Myfxbook documents filters, countdowns and email reminders. These features do not establish a speed ranking or a trading advantage. Test the view on the device you use.

IG states that DailyFX closed on 4 September 2024, with trading content available through IG. Do not treat its old calendar recommendation as a separate current service. Keep an agency’s or central bank’s official schedule alongside the aggregator and recheck dates rather than relying on a remembered recurring rule.

The filters that actually matter

Filter for relevant currencies and events, but consider overnight positions, cross-market effects and releases outside your active session. A low provider rating does not guarantee an event is irrelevant.

Use filters to make the schedule manageable without hiding relevant risk. Neither a six-to-ten-event count nor a tighter stop follows automatically from a rating filter.

How to fold the calendar into session prep

Record the event name, date, publisher, reference period, unit and original time zone, then verify the displayed local time. USA and Poland change clocks on different dates. In 2026, New York starts daylight time on 8 March while Warsaw changes on 29 March. A conversion example, not an event schedule: 08:30 New York is 13:30 Warsaw on 13 March 2026 but 14:30 on 24 September. Likewise 14:00 converts to 19:00 and 20:00, and 14:30 to 19:30 and 20:30. BLS’s July 2026 schedule places the June Employment Situation on Thursday 2 July at 08:30 ET, rather than a first Friday. The Fed’s September schedule separates the 16 September meeting communication at 14:00 from the press conference at 14:30. Recheck each actual schedule; these hours are not a rule for every speech.

If you use a reminder, choose its lead time and test the device’s time zone, permissions and delivery. A reminder before publication differs from a notification of new data. Neither executes a position decision, and no universal twenty-minute interval creates safety. Predeclare how to handle entries, pending orders and open exposure; moving a stop or changing size also changes the trade’s rules. Record the current quotes and action rather than assuming a calm or guaranteed fill. See weekly market preparation.

Same monthly change, different annual change

Identify whether a calendar percentage compares consecutive months or the same month in two years. A monthly change alone does not determine an annual one.

Scroll sideways to see the full diagram

Same monthly change, different annual changeTwo fictional versions of the same non-seasonally-adjusted price index both read 100 exactly 12 months earlier. A reads 100 in the previous month and 100.2 now. B reads 102 in the previous month and 102.204 now. Both monthly changes are 0.2%; annual changes are 0.2% for A and 2.204% for B. The shared linear index axis is 99.5–102.5. The time axis runs from −12 to 0 months with proportionally spaced points at −12, −1 and 0. Only three readings are shown, with no line reconstructing unknown history. These are invented data, not CPI releases or exchange-rate forecasts.CHECK PERIOD AND COMPARISON BASEFictional price index • same units99.5100101102102.5−12−10Months relative to nowA: 100 →100 →100.2B: 100 →102 →102.204Both m/m: +0.2%A y/y: +0.2% • B y/y: +2.204%Three readings, not a complete index path.

Assumptions: Two fictional versions of one index, with index units and the same base. Country, coverage and population are unchanged; both versions are non-seasonally adjusted. Times −12, −1 and 0 are exact months relative to the current reading. Other months are unknown.

Both fictional versions read 100 exactly 12 months ago. A reads 100 in the previous month and 100.2 now; B reads 102 and 102.204. Monthly changes are (100.2 /100 − 1) × 100 and (102.204 /102 − 1) × 100: both 0.2%.

Annual change compares now with 12 months ago: A gives 0.2%, B gives 2.204%. Record country, index coverage, change period and seasonal-adjustment status when reading a calendar. Our example isolates only the period: both versions have the same coverage and no seasonal adjustment. We do not mix monthly SA with annual NSA data.

BLS — calculating index percentage changes explains monthly and twelve-month changes and the requirement for the same index base. Our numbers are assigned, not statistical releases. This comparison establishes neither currency direction nor a market reaction.

Monthly 0.2% does not determine an annual change without the correct base.

Read surprise and revisions together

A forecast is not a directional trading instruction. The surprise, revisions, other details and positioning can all matter; matching consensus does not guarantee a quiet market or week.

Hypothetically, an actual jobs change of 95,000 versus forecast 180,000 is a negative surprise of 85,000 in that measure. A separate two-month exercise uses thousands of jobs: prior published 200, current forecast 200, revised prior 170 and current actual 210. Current surprise is +10 and prior revision −30. Against an explicit no-revision baseline of 400, the new sum 380 is −20. If a revision to 180 was already expected, the baseline is 380 and the sum difference zero. This is reconciliation of like periods and units, not a universal report-strength score or currency signal. Do not mechanically sum rates, indices or differently timed series.

What the calendar will not tell you

A schedule alone does not show the full policy statement, positioning or the sequence of information. Follow the underlying releases; no fixed fifty-pip effect from one word is established.

A policy decision and subsequent questions can contain different information. Follow the underlying statement and press-conference material rather than assuming the first reaction captures everything; a calendar does not prescribe that a live stream must always be watched. For historical research, retain the first available release, later revisions and the time at which consensus was collected. Today’s revised history is not necessarily what a trader could see at the earlier decision time.

What to do tomorrow

  1. Choose a calendar, verify its time zone and relevant currencies, and review filters regularly. Do not permanently hide all lower-rated or out-of-session events relevant to open exposure.
  2. List relevant releases over the possible holding period, including overnight events. Record the unit, reference period and official time zone; update the plan on the decision day rather than imposing a two-to-three-event limit.
  3. Test any reminder’s time zone, delivery and permissions. Choose its lead time for the task; no fixed number of minutes guarantees calm decisions or safe fills.
  4. Record responses to plausible release outcomes, revisions and execution changes. A matching headline can still accompany a price move; the plan should include when evidence is insufficient to act.
  5. After the event, compare the information actually available at each decision, the action and the fill. Preserve discrepancies and costs; a one-week comparison does not isolate the calendar’s causal effect on net returns.