About this site, its data and its method

What this is

A calendar, a rate table and a record of reactions for 8 central banks: Federal Reserve, European Central Bank, Bank of England, Bank of Japan, Bank of Canada, Reserve Bank of Australia, Reserve Bank of New Zealand, Swiss National Bank. For each decision it shows when it lands, what economists expected, the result, and, since January 2025, what the bank's main currency pair did in the minutes around it. Forexia built it as the public record of the read it teaches: that a decision is a burst of liquidity at one minute, and that the first spike is often not the move.

Where the data comes from

• Dates, consensus and results: FXStreet's economic calendar. The "expected" figure is the consensus of the economists it surveys.
• Prices: Dukascopy's bid candles, at one minute around each decision and at 15 minutes for the standing highs and lows. Bid prices, so a pip here can differ by the spread from a broker's chart.
• Times: the central banks' own sites are the authority. Our times come from the calendar, and the Bank of Japan does not announce a fixed hour, so its times are estimates and are marked. Check the bank before you trade.
• Built: the data on these pages was last refreshed Oct 28, 2026. Every page is built from it, and nothing is typed by hand.

How a reaction is measured

The rules were set before any decision was counted, and they are the same for every bank. They are measured on one-minute candles of the bank's main pair.

The box

the high and low of the 60 minutes before the announcement.

The first side

the box edge that price trades through first, within 30 minutes of the announcement.

The close back

a five-minute candle (counted from the announcement) closing back inside the box within 60 minutes of it.

The other side

after the close back, price trading through the opposite edge within four hours.

That is the read the site teaches: the decision brings a burst of orders at one minute, the first spike runs the stops beyond one side, and when it closes back the move usually belongs to the other side. The counts say how often that happened, including how often it did not. A decision whose candles are incomplete is left out rather than estimated.

Two details of the measure. When one minute goes through both edges, the bigger run is the first side. "Median first move" is the larger of the up and down move from the last price before the announcement, within 15 minutes.

How the projection is drawn

Before a decision each page shows a map of where stops are likely resting, worked out from real 15-minute candles, never typed:

Standing levels

the highs and lows of the current day so far, of the two trading days before it (a day starts at 17:00 New York), and of this week and last week.

Play A

the spike runs the nearest high, a candle closes back below it, and the projection is the nearest low, then the next one.

Play B

the mirror: the nearest low is run, closed back above, and the high is next.

Either, both or neither may happen. The reaction counts say how often the first spike was closed back; the map only says where to look. It is not a prediction and not a trade signal.

How votes are counted

Before each decision its page asks two questions: what the bank will do (hike, hold or cut) and, when the map has both plays, which one happens on the pair (A, B or neither). Voting closes at the announcement. Each browser has one vote on each question, and one network at most three, so that a household or an office can vote but one machine cannot fill the ballot. The split is shown to you once you have voted, so that earlier votes do not steer later ones, and to everyone once voting has closed.

What happened is read from the data, the same way every time: the rate from the result, and the play from 15-minute candles, for 24 hours from the announcement.

A played out

price trades above A's high, a 15-minute candle closes back below it, and a later candle trades at A's first target.

B played out

the mirror: below B's low, a close back above it, then B's first target.

Both

the first to reach its target. If both reach it in the same candle, neither is counted: the candle does not say which came first.

Neither

24 hours pass and neither play has reached its target.

The map settled is the one drawn from every candle before the announcement; before then its levels move with each refresh of the data. The votes are what visitors thought, not a forecast, and how a crowd voted says nothing about what the next decision will do.

Using the data

The figures here describe 105 measured decisions. You may quote them, in an article, a video or a post, with a link back to the page you took them from, for example How to trade interest rate decisions. Please say they are counts of past decisions measured with the rules above, and do not present them as a forecast. The underlying calendar and prices belong to FXStreet and Dukascopy.

Corrections

If a date, time or figure is wrong, tell us on the Forexia Telegram channel. We have no email address for this site yet. Say which page, what it shows, and what the bank's own site says. Corrections to a time are the most useful.

Not advice

Nothing here is investment advice or a recommendation to trade. Trading currencies on margin carries a high level of risk and can lose more than you put in. The figures are about what happened. Forexia teaches the method in full, and its own policies are at forexiapro.com/privacy.