Why the First Candle After a Release Lies

Everyone who has traded a news release knows the feeling. The number lands, gold rips forty dollars in fifteen seconds, and the move looks so decisive that not being in it feels like a mistake. Ten minutes later price is back where it started, your stop was taken on the way, and the candle that convinced you is still sitting on the chart looking perfectly reasonable.

That first candle is the single most misread object in news trading. Not because it's fake — the prints are real — but because people read it as a conclusion when it is mostly a description of who happened to be standing there when the number hit. This piece is about what it's actually made of, and how to find out what it does on your instrument rather than taking anyone's word for it.

The book empties before the number lands

The important thing happens a few seconds before the release, and it doesn't show up as a candle at all. Market makers and anyone else quoting two-sided prices pull their orders. Nobody wants to be the resting bid when a number they haven't seen is about to print. Liquidity thins out — sometimes dramatically — and the spread widens.

So when the number arrives, the first orders to hit are trading into a book that has far less depth than normal. The same size that would move price a few tenths on a quiet afternoon can move it many multiples of that. The distance travelled in those first seconds is telling you about the thinness of the book at least as much as it's telling you about the market's opinion of the data.

This is why the first candle so often has a long wick in both directions. It isn't a battle between conviction buyers and conviction sellers. It's price traversing an empty zone, finding almost nothing, then traversing back when real size finally shows up and quotes normalise.

The headline number is not the news

The second reason the candle misleads: the instant reaction is triggered by whatever can be parsed instantly, which is the headline figure against the consensus. But the thing that actually matters to positioning is usually somewhere in the detail — the composition of the print, the revisions to previous months, the internals that inform what the central bank will care about.

A payrolls number can beat expectations on the headline while the prior two months are revised down by more than the beat. A CPI print can land exactly on consensus while the core reading and the shelter component say something quite different about the trend. Those details take a human, or a slower model, some seconds to minutes to digest.

So the sequence you're watching is frequently: instant algorithmic reaction to the headline, then a partial or complete unwind as the detail gets priced, then whatever the actual conclusion turns out to be. Trading the first candle means trading step one of a three-step process and hoping steps two and three agree with it.

The chart is not the price you'd have got

There's a third gap, and it's the one that turns a theoretically fine idea into a losing one. Your chart draws a mid price, or your broker's bid, on a clean continuous line. What you would actually have transacted at, in that window, is a different number.

• The spread widens, often substantially, and often before the release rather than after. A position opened into that pays the wide spread immediately.
• Market orders slip. In a thin book, the price you clicked and the price you got can be meaningfully apart, and the gap is worst exactly when the move is biggest.
• Stops become market orders when triggered, so they suffer the same slippage — but against you, and at the worst moment.
• Pending orders can fill at gaps rather than at your level, which quietly changes the risk on the trade you thought you'd defined.
• Some brokers restrict or reject orders entirely in the seconds around a major release, so the plan you rehearsed may not even be executable.

None of this appears on the candle. Backtesting a news strategy on chart prices, or judging it from a screenshot afterwards, systematically flatters it — every one of the effects above pushes results in the same direction, and the direction is not yours.

Measure it yourself — the release log

Here's the part that actually changes anything. You do not need anyone's statistics about how often the first move reverses, and you should be sceptical of any that get quoted at you without a stated instrument, broker, sample and period. Those conditions are exactly what determines the answer, so a number without them is decoration. What you need is a record of what happens on your instrument, with your broker, over the releases you actually intend to trade.

Build it by hand. It takes a few minutes per release and it will teach you more than a year of reading about it. For each high-impact release you care about, log:

• The release, the date, the consensus and the actual figure — plus any revision to the previous period.
• The spread you observed about thirty seconds before the release, and again ten seconds after. Write both down; the difference is the cost of being early.
• Price immediately before the release.
• The high and the low of the first minute, and which came first. The order matters more than the range.
• Price at five minutes, fifteen minutes, one hour, and at the session close.
• Whether the first-minute direction was still the direction at each of those checkpoints.

After ten or fifteen releases you will have something no listicle can give you: the actual behaviour of your instrument, in your conditions. Look at how often the first minute's direction survived to the one-hour mark. Look at whether the surviving cases had anything in common — a particularly large surprise versus consensus, agreement between headline and revisions, a trend already in place. Look at how much of the eventual move was still available fifteen minutes in, once the spread had normalised and you could actually see what you were doing.

That last one is usually the finding that changes people's behaviour. If most of the move is still there at fifteen minutes, the entire argument for trading the first candle collapses — you were accepting the worst execution of the day to capture a fraction you didn't need.

What to do instead, while you're gathering the data

You don't have to sit out releases to run this. You have to stop treating the first candle as the signal. A workable stance, and the one the log will usually validate:

• Treat the release as a volatility event with a known time, not as a directional forecast. You know when it lands; that's the edge, and it's about preparation rather than prediction.
• Be flat or deliberately sized down through the print unless you have a specific reason not to be. Being flat is a position, and in a thin book it's frequently the correct one.
• Define in advance what would make you act afterwards — a level reclaimed and held, a retest that fails, a close beyond a boundary you drew before the number. Write it down before the release, because afterwards you will find a reason for whatever price just did.
• Let the spread normalise before you transact. Watching it is a better entry timer than any indicator, and it costs nothing to check.
• Judge yourself on whether you followed the plan, not on whether the candle went your way. Over ten releases, the process is the only thing you control.

The releases worth building this habit around are the ones that reliably move gold and the dollar — the Fed decision and its press conference, CPI, and payrolls. If you want the calendar-first routine those sit inside, that's covered separately; this piece is only about the sixty seconds everyone gets wrong.

FAQ

Does the first candle always reverse?

No, and anyone who tells you a fixed percentage is inventing precision. Sometimes the initial move is the move, and those are usually the cases where the headline, the internals and the revisions all point the same way and the market was positioned the other way. The point isn't that the first candle is always wrong; it's that at the moment you'd have to act on it, you cannot yet tell which kind you're looking at — and your execution is at its worst precisely then.

What time frame should I actually watch?

For the log, the one-minute is fine for recording the initial range, but don't make decisions on it. The useful habit is to record the first minute and then decide on something slower, because the slower frame naturally waits out the liquidity vacuum. If you find yourself needing a five-second chart to justify an entry, that's a reliable sign the trade depends on execution you don't have.

My broker's spread barely widens. Doesn't that solve it?

It helps, and it's worth knowing — which is precisely why the log records the spread rather than assuming it. But spread is only one of the three problems. The thin book and the headline-versus-detail lag are properties of the market, not of your broker, so they apply regardless of how good your pricing is. A tight spread makes participating cheaper; it doesn't make the first candle informative.

Can I just use a wider stop to survive the whipsaw?

You can, and it changes the problem rather than removing it. A wider stop means either a smaller position for the same risk — which shrinks the reward you were chasing — or the same position with more risk, which is how a single bad release becomes a genuinely damaging day. It also doesn't help with slippage, since the fill quality is unrelated to where you placed the stop. If the honest answer is that the trade only works with a stop wide enough to make the position trivial, the trade wasn't there.

How many releases before the log tells me anything?

You'll notice the execution facts — spread behaviour, slippage, whether your orders even go through — within two or three, and those alone are worth the exercise. The directional question needs more, and honestly it needs more than you'll gather in a season. That's fine: treat the log as a description of conditions rather than a prediction engine. Knowing what the first sixty seconds costs you is useful even if you never learn to forecast the direction, because it tells you what you're declining to pay for.