Reading A CPI Release In The First Ten Minutes

CPI lands at a fixed minute. You already know the minute. What a lot of traders do with it is this: read the headline year-over-year figure, compare it against the one they half-remember from last month, decide inflation is hot or cool, and click. Ten minutes later the position is offside and the number they read was correct.

The problem is not the number. The problem is the order. A CPI release is not one figure. It is a small table published in a single instant, and the market reads the whole table — plus the forecast it was already holding, plus the revision to the month before — inside a few seconds. If you only read the headline, you are reacting to roughly a quarter of what moved the price.

This guide sets out a sequence: the data, the revision, the reaction. Read in that order, the first ten minutes become legible. Read out of order, they look random.

A standard US CPI release gives you four figures at once: headline month-over-month, headline year-over-year, core month-over-month and core year-over-year. Headline includes food and energy. Core strips them out.

The two timeframes describe different things. The month-over-month figure describes the pace of price change in the month just measured. The year-over-year figure carries eleven months of history inside it, which means it moves slowly and can rise or fall purely because of what dropped out of the back of the window. That is a base effect, and it is not new information about the current month.

None of the four is automatically the number. Which one dominates depends on the regime. When oil is elevated, the gap between headline and core becomes the story, because energy-led inflation still feeds inflation expectations and still pressures a central bank to stay restrictive, even when the core figure looks contained. When energy is quiet, core month-over-month tends to carry the weight, because it is the cleanest read on momentum in the sticky part of the basket.

Before the release, decide which of the four you will weight and why. Deciding afterwards is not analysis. It is justification.

The first read is not the level. It is the gap between the print and the consensus forecast.

The forecast was in the market days before the release. If consensus was 0.3 on core month-over-month and the print is 0.3, nothing has been learned, however high 0.3 may feel in absolute terms. Price moves on the difference between what was expected and what arrived, not on whether inflation is objectively high.

Two things sharpen that first read. The first is the size of the surprise relative to how much these figures normally miss by. A tenth of a point on core month-over-month is a larger statement than a tenth on headline year-over-year, because the monthly core series has less room in it. The second is direction against trend. A hot surprise that continues three hot prints says something different from a hot surprise that interrupts a run of cool ones.

Write the consensus down before the release. On paper, in your journal, wherever you will actually see it. If you have to look the forecast up after the number lands, you will anchor on the level instead of the surprise, and that is the habit this whole sequence is built to break.

The second read is the part most headline traders never get to, because by then they are already in a position.

Revisions to the prior month are published alongside the new print and reported almost nowhere. They matter because they change the trend the new print sits on top of. A hot print that arrives with a downward revision to the prior month is a different statement about the trend than a hot print stacked on a confirmed one. Average the two months before you decide anything has changed.

Then take the headline and core split. If the upside is concentrated in energy, the inflation runs through the expectations channel and the policy path rather than through the core basket directly. If core is firm on its own, the policy path shifts more directly and more durably.

This is also where a catalyst's narrative and its effect come apart. Inflation pulls a bullish narrative for gold, because gold is widely described as an inflation hedge. The actual effect can run the other way, if the market reads the print as forcing a central bank to stay restrictive and real yields climb as a result. Narrative pull and actual effect are two different fields, and on CPI day they often disagree.

The third read is where the market is expressing its answer.

The first candle after a CPI print is liquidity, not information. Spreads widen, resting orders get cleared out on both sides, and the chart prints a shape that reflects thin books rather than a settled interpretation. Reading that candle as a verdict is how a correct view on the data turns into a bad entry.

So look at rates first. The two-year yield, the ten-year, the real yield and the dollar. Real yields are gold's cleanest anti-driver: when the expected policy path rises, real yields rise, and a higher real yield raises the opportunity cost of holding a metal that pays no coupon. That relationship is the mechanism. Gold's candle is only the output of it.

The question between minute five and minute ten is simple. Did the yield move hold? If yields jumped and kept the move, the market has agreed on an interpretation of the print and gold is being repriced against it. If yields jumped and gave it all straight back, the first interpretation was rejected, and whatever gold did in the meantime was noise with a wick on it.

One more thing worth saying plainly. You do not need to be in the first ten minutes at all. The sequence above is a reading exercise. Participation is a separate decision, and sitting on your hands through a release while you read it properly is a legitimate outcome.

The fundamental desk publishes a daily read on XAU/USD. It gives a bias, a confidence level, a dominant driver, a short regime summary, and a list of active catalysts. Each catalyst carries two separate fields: its narrative pull and its actual effect — the exact distinction the second and third reads above are built on.

The use on CPI day is before the print, not after it. If you already know which driver the desk has marked as dominant, your third read has a frame. A regime summary built around elevated real yields tells you which channel to watch when the number lands, and which reaction would be consistent with it and which would not. You are no longer meeting the data cold.

Be clear about what the desk is. It is a structured description of the regime as it currently stands, with its reasoning shown. It does not forecast the print, it does not predict the reaction, and it does not tell you to do anything. The decision is yours, and so is the risk. The AI Trading Assistant in the member app is read-only for the same reason — it cannot place a trade, close one, or move money.

After the release, write down what you expected, what the four figures actually said, what the revision did, and where yields settled by minute ten. Do that for a few releases and you will see where your reading is leaking long before you see it anywhere else.

Trading carries risk of loss. Every decision is yours. Nothing on this page is financial advice.