How a NewsPips reading is made

Every reading NewsPips publishes is a claim about the news, and a claim you cannot check is one you should not trust. This page explains how a reading is made: where the news comes from, how it becomes a directional bias for one market, why no confidence number sits beside it, and where the whole thing can go wrong. NewsPips is built on the assumption that you will check its reasoning rather than take it on faith, so the method is written out here in full.

1. Watch

   hundreds of sources

2. Group

   one story, not five copies

3. Judge

   direction + impact

4. Cite

   every claim links to sources

Per market, refreshed as stories land. Context, not advice.

From hundreds of sources to one story

NewsPips watches hundreds of news sources around the clock: the wire services, the financial press, official releases from central banks and statistics offices, and the specialist outlets that cover one market closely. It watches the economic calendar alongside them, so a scheduled release is read against the forecast it beat or missed. Breadth is the point. Market-moving news rarely breaks in one place first, and no single publication catches everything early.

Breadth creates a problem of its own. One event produces dozens of near-identical headlines as outlet after outlet picks up the same report, and a raw feed buries the event under its own repetition. So NewsPips gathers the coverage of one event into a single story. You see the event once, not fifty times, and the story carries what a pile of headlines cannot show you at a glance: how many independent outlets are reporting it, and how the coverage has built since the first report.

Everything NewsPips writes starts from stories, not headlines. When a reading says the market is leaning on a particular development, it is pointing at a story you can open, with the source articles behind it.

One reading per market

NewsPips does not judge articles one at a time, and it does not publish a single "market mood" number. Each market it covers gets its own reading: the major dollar pairs, gold and silver, the dollar index, Brent and WTI crude, bitcoin and ether. A reading is built from the stories that matter to that market and from the calendar releases that have landed recently.

A reading has two parts. The first is a directional bias: the way the news currently leans for that market. The second is the written case for it, and the case is the point, not the decoration. Every material claim in it links to the story and the source articles that produced it, so you can trace any sentence back to the coverage and judge the sourcing yourself. If a reading says a hawkish surprise supports the dollar, the link shows you which release, reported where. The reading is written with the help of AI, which is one more reason every claim in it is linked: you are never asked to take the writing on trust.

The same event routinely points different ways for different markets. A hawkish line from the Federal Reserve tends to pull the dollar one way and gold the other, and the reading for each market weighs it separately. That asymmetry is why a reading is made per market rather than for "the market" as a whole.

A reading is refreshed as new stories land. When the lean on a market changes, NewsPips sends a bias-flip alert with the stories that caused the change, so you find out when the picture moves rather than when you next look.

Why there's no confidence score

You will not find a confidence or certainty number beside a reading, and that is deliberate. A figure like "72% confident" looks like a measurement. A reading of the news is a judgement about what the coverage implies, and a percentage attached to a judgement is false precision: it makes two readings look comparable when neither number can be checked against anything, and it invites you to lean on a decimal point that nobody can stand behind. Direction is not something anyone can honestly put a percentage on, so NewsPips does not print one.

What you get instead is what a number would have hidden: the direction, the stories behind it, and the sources to check. If the case holds up when you open the sources, that is your confidence, and it is worth more than a figure someone else assigned. If the case does not hold up, no number would have saved it.

NewsPips does put a number on the one thing it can measure. Inside the app, the chart carries an expected range: a band price has stayed inside about 80% of the time over the next half hour to an hour. That is a statement about how far price typically travels in the window after a reading, not about which way it goes. Range can be measured at short horizons and reported with how often it holds. Certainty about direction cannot.

Where a reading can be wrong

An honest methodology page lists where it fails. These are ours.

• Fast-moving stories. A reading describes the coverage at the time it was written. In a rapidly developing situation it can lag the latest headline, and the first reports of a breaking event are often partly wrong in ways that get corrected over the following hours.
• Ambiguous events. Some releases genuinely cut both ways: data that is strong for growth but hawkish for rates, for example. The reading comes down on one side of which channel dominates, and that judgement can be wrong.
• The writing itself. Like any written analysis, a reading can misweigh evidence, over-read thin coverage, or misstate a detail. The links exist partly as a check on exactly this. A claim you can verify is a claim you can catch.
• What is not in the news. A reading is built from public coverage. Order flow, positioning, and anything else that moves a market without producing a story sit outside its view.
• Already-priced news. A reading assesses what the news implies, not how much of that implication the market has already traded. A correct reading of stale news is not a useful one.

What a reading is not

To leave no room for doubt:

• A reading is not investment advice, and it is not personalised to anyone's situation, account, or appetite for risk.
• A directional bias is not a recommendation to buy or sell anything. It is a summary of what the current news implies for one market, published with its sources.
• No reading carries any assurance of accuracy or of future performance. Trading involves a substantial risk of loss, news-driven trading especially, and no analysis removes that risk.

NewsPips never tells you what to trade. Whether to act, in what size, and with what protection are your decisions, and nothing on the page makes them for you.

Reading NewsPips well

The readers who get the most from NewsPips treat it as a briefing beside their own process, not a verdict to follow. In practice that means letting the calendar tell you what is scheduled, so a release never takes you by surprise; reading each market's page as you would a morning briefing (what happened, how well sourced is it, which way does it lean); opening the sources on anything that would change your view; and reading the expected range as a gauge of typical short-term movement, not a forecast of direction. A bias-flip alert is a reason to re-read, not an instruction. If something on screen is unclear, NewsPilot will answer questions about it.

That habit assumes some grounding in how news moves prices: why a surprise matters more than a headline, and what happens inside a release window. If you are building that foundation, start with how to trade news events and the economic calendar, explained. To see the method at work first, the markets pages show each market's latest reading and the stories behind it, on a 15-minute delay, and the pricing page says what each plan includes.