What really happens in the opening moment of a data release

What an embargo actually is

Official statistics are not invented at the moment of publication. They are compiled in advance, checked, and held under embargo until a stated release time. Journalists and sometimes analysts may be given access beforehand under lock up conditions, with communication restricted so that nothing leaves the room before the clock. The purpose is fairness: everyone receives the number simultaneously rather than in an order that favours whoever sits closest to the source.

At the release instant the data goes out through several channels at once. A web page is published. A machine readable feed carries the figures in a structured format designed to be parsed rather than read. Commercial distributors relay it onward. Those channels do not all arrive at the same microsecond, and the difference between the fastest and the slowest is enough for the market to have moved substantially. So the honest framing is not that information is released at a moment, but that it propagates across a brief interval, and participants sit at different points along that interval.

The book before the print

Watch the order book in the minute before a scheduled release and you can see the market prepare. Resting orders are pulled. Quoted spreads widen. Depth at and around the price thins out. Market makers are not being awkward; they are declining to hold a position against a number they cannot see, because a quote left in the book is an offer to be run over by whoever reads the data first.

The practical effect is that the market you trade into at the release moment is not the market you were looking at a minute earlier. The displayed price may be roughly the same. The quantity available at that price is a fraction of what it was. This is also why the widening spread before a release is useful information rather than an annoyance: it is a direct reading of how much protection participants think they need. A release expected to be decisive produces more withdrawal than a minor one, which means the book tells you something about the kind of event you are facing before the number even lands.

The first prints land wherever the orders were

In the first fraction of a second after the embargo lifts, price does not move to a considered valuation. It moves to wherever the remaining orders happen to sit. With the book hollowed out, a modest amount of aggressive buying or selling walks through several levels, and the prints that result are a record of where resting interest was left rather than of what the data means.

This is the origin of the long wick that so often appears on a release candle and then gets retraced. It is not a conspiracy and it is not usually a stop hunt in any deliberate sense. It is the mechanical consequence of an empty book meeting urgent orders. Once the initial flow is absorbed, quoting resumes, depth rebuilds, and price settles somewhere that reflects an actual interpretation. The figure sets out that sequence as a timeline, with the spread band widening before the release and narrowing afterwards. Treating the extreme of the first spike as a meaningful level is reading information into an artefact of thin depth.

Headline first, meaning afterwards

The reaction arrives in two stages with different authors. The first stage is automated and responds to the headline figure against the expected figure, in milliseconds. The second stage is slower and human, and it responds to the detail: the composition of the number, revisions to previous months, the internals that change its interpretation, and how the release fits with other recent data.

Those two stages can point in opposite directions. A headline that looks strong can contain weak internals or a large downward revision to prior months, and the slower reading reverses the fast one. This is why the move twenty minutes after a release is often more informative than the move two seconds after it, and why a trader who enters on the spike frequently ends up on the wrong side of the considered interpretation. It is also where a discretionary trader has any advantage at all, since reading context is a human skill while parsing a number is not. The practical consequences for gold are set out in trading gold around news.

Expectation is the baseline, positioning is the multiplier

A release does not move price because the number is good or bad. It moves price because the number differs from what was already priced in. The baseline is the expectation, and the reaction is to the surprise against that baseline. A weak figure that was widely anticipated can produce almost nothing. A mildly unexpected one can produce a large move.

Positioning then multiplies or mutes the result. If a large share of participants are already leaning one way, a surprise that confirms their view leaves little left to buy, while a surprise against them forces an unwind that moves price much further than the data alone would justify. This is why the response to an inflation print can look wildly disproportionate in one month and barely visible in another with a similar surprise. It also explains a pattern that frustrates newcomers: the market sometimes moves in the direction the data appears to contradict, because the dominant flow is people getting out rather than people responding. The specifics for one of the most watched releases are covered in how CPI affects gold.

Why the race is already lost, and why that is fine

Being honest about this saves a lot of money. Participants whose business is trading releases have direct feeds, systems positioned physically close to the matching engines, and parsing measured in microseconds. Their orders are in the market before a retail platform has finished redrawing the chart. There is no router, no connection and no reaction time that closes that gap, because the gap is structural rather than a matter of effort or equipment.

The conclusion is liberating rather than depressing. If you cannot win the race, stop entering it. Everything actually available to a discretionary trader happens after the first seconds: the interpretation, the structure that forms once depth rebuilds, the level that gets retested, the failure of a move that went too far. Those require patience and judgement rather than speed. Deciding in advance that you will not place an order in the opening seconds removes a whole category of expensive outcome, and the way price behaves immediately afterwards is worth studying on its own terms, as described in reading a fast market.

What your fill really looks like in the first seconds

What happens to your own order in that window is worth stating bluntly. A market order sent into a hollow book fills at the first available price, which can be far from the one on your screen. A stop already in the market becomes a market order the instant its level trades, so a spike that wicks through your stop and comes straight back can remove you at the worst price of the day. A limit order either does not fill at all, or fills because price kept going. A requote or an outright rejection is common.

Position size is the only variable fully under your control here, and it has to be set for the worst fill rather than the intended one. If a fill noticeably beyond your stop would be a serious problem, the position was too large for an event window. The behaviour of both spread and slippage around releases is detailed in spread and slippage on news. The figure below shows the same stop level producing two very different exits depending on whether depth was present when it triggered.

A routine that respects the mechanism

A routine that fits the mechanism looks roughly like this.

• Know the release calendar for the week before the week begins, and which items have historically been capable of repricing gold.
• Decide in advance whether each one is a stand aside or a trade, and make that decision while nothing is at stake.
• If a position is open through a release, size it for a bad fill and have protective orders already in place rather than planning to react in the moment.
• Place no new market orders in the opening seconds.
• Let depth rebuild, then assess structure: where the move stalled, what was swept, whether a level was reclaimed.

The caveat is real and should not be glossed over. Waiting does not guarantee a cleaner opportunity. Some releases produce a move that never retraces, and patience in those cases simply means missing it entirely. The trade off is accepting a lower frequency of entries in exchange for avoiding the worst executions the market offers, which is a reasonable bargain but not a free one. If you want to see the mechanism rather than read about it, watch a scheduled release land on the live gold chart with no order in the market and pay attention to the spread rather than to the candle.