Editorial policy & model standards

GB Nomics adds calculations, comparisons and explanations to published UK economic and market data. This page explains what those additions can—and cannot—tell you.

Updated 7 October 2026.

Published observations and our interpretation

A source observation is a value reported by its publisher for a stated period. A calculated indicator transforms observations, for example by subtracting two yields or calculating a growth rate. A model combines or transforms several inputs. GB Nomics does not treat these three things as interchangeable: units, dates, formulas and source links accompany the charts.

The data guides are written for this dashboard and use illustrative worked examples. Example numbers are labelled as examples, rather than presented as current readings. The site’s explanatory content and code have been developed with AI assistance. This does not constitute independent professional review.

How to assess a chart

1. Check the observation date and frequency. A quarterly number and a daily market close do not describe the same period.
2. Read the units. Percentages, percentage points, index levels and pounds answer different questions.
3. Compare the actual data with its history before interpreting a standardised score.
4. Read the formula and source limitations. A relationship in a chart does not establish cause and effect.

What a Model Chart means

The standard Model Chart fits a straight-line trend over the selected window, subtracts that trend, then scales the residuals by their sample standard deviation. Zero means on the fitted trend. It does not necessarily mean zero growth, a neutral policy stance or fair value. A value of +2 means two standard deviations above the trend within that window.

The full selected window is used to estimate the trend. Changing the time range, adding a new observation or revising old data can therefore change earlier scores. These charts describe history; they should not be read as signals that were necessarily available in real time.

Forecasts require a different standard

Experimental GDP estimates are distinct from the descriptive models. Their cards identify the target period, training assumptions and historical errors. The Scorecard compares its selected weighted forecast with training-average growth and a no-growth benchmark. At its initial assessment, the historical test did not demonstrate an advantage over both benchmarks; consult the current card for recalculated results. A 0–100 score is a historical percentile, not a probability that GDP will grow.

Historical testing uses currently available, potentially revised source histories and assumed publication delays. It is not an archived record of forecasts made before each release. Changing the methodology after examining results can also overstate apparent performance. Uncertainty ranges describe past errors under the stated method, not guaranteed future bounds.

Revisions, missing data and refreshes

Source feeds are checked when the dashboard opens and every 24 hours while it stays open, subject to server caching. This is not a promise of an update at the instant a publisher releases data. Monthly and quarterly statistics naturally arrive after their reference periods.

Failed or stale inputs are withheld from current cards and dependent models. Missing observations are not silently converted to zero. Where a model uses a partial basket or a matching-date rule, its methodology explains that choice. A change in a historical chart can reflect a source revision rather than a new economic event.

Corrections and accountability

To investigate a discrepancy, compare the card’s source link, period, units and data table. Include the indicator or model name, the observation date and the source value when reporting an issue. Source revisions and errors in our calculations are different problems; both can affect the displayed history.

Send corrections to info@gbnomics.com.

Independence and scope

GB Nomics is an independent educational project, not an official government service, a statistical authority or an investment adviser. It does not offer personal financial advice or guaranteed trading signals. Advertising does not determine formulas or source values; any future advertisements must be visibly separate from editorial material and chart controls.

Access to a public data endpoint does not itself establish a licence to redistribute or commercially reuse every dataset. Rights and conditions remain with the relevant publishers. Source links identify where observations can be checked; they do not imply endorsement of GB Nomics.

Read the data guides · Chart methodology · Privacy & cookies