How we calculate this

How to read our numbers

Every statistic in Verdiqt is shown with its context: the benchmark it is compared against (always labelled with the geography level), the source of the figure, and an "as at" date. Two kinds of figures appear in the app:

• Measured figures come from published government data and are labelled with their source (for example "NSW Valuer General" or "ABS"). They carry a sample size and, where a history exists, a small trend line.
• AI estimates are generated by our analysis model from its training knowledge. They are always labelled "AI estimate — verify before deciding", never carry an invented benchmark, and should be treated as a starting point — not a market fact.

We never invent, interpolate or zero-fill a benchmark. If no real benchmark exists for a figure, it says "benchmark unavailable".

Median price

Definition. The middle sale price of properties sold in a suburb over the most recent reporting period — half of sales were above it, half below. We use the median (not the average) because a handful of very expensive sales would otherwise drag the figure upward.

Source and update frequency. NSW Valuer General bulk property sales data (address-level sales records, published weekly; history back to 1990), aggregated per suburb and reporting period. Attribution: Contains property sales information © State of New South Wales (Valuer General), CC BY 4.0. Coverage is currently NSW only — suburbs outside NSW show an AI estimate, labelled as such.

Benchmark. The suburb median is compared against the most specific ABS mean price of residential dwellings figure we hold — currently the NSW state figure, because the ABS publishes catalogue 6432.0 (quarterly) at state level only. The ABS discontinued its free capital-city median series in 2022, so the benchmark is a mean, and we label both the statistic and its level exactly — "vs NSW (ABS mean dwelling value)" — rather than implying a median or a city-level comparison. Benchmarks at the city (Greater Sydney) and council (LGA) levels are planned; until the data supports them we show none rather than an approximation.

The delta. The percentage shown beside the benchmark is the simple difference between the suburb figure and the benchmark. A higher or lower median price is not inherently good or bad, so this delta is never coloured green or red.

Property value

Definition. An estimated market value range for a specific property.

Source. This is an AI estimate based on the model's knowledge of the suburb, property type and attributes — Verdiqt does not currently hold property-level sales feeds. It is labelled as an estimate wherever it appears and carries no benchmark. Verify against recent comparable sales before relying on it.

Capital growth

Definition. The change in a suburb's median sale price over a period, expressed as a percentage. The measured "12-month growth" figure compares the latest reporting period's median with the same period a year earlier, from the same NSW Valuer General dataset as the median price. Longer horizons (5-year, 10-year) shown in reports are AI estimates and are labelled as such.

Benchmark. We do not currently hold a published city-wide growth-rate series, so the growth figure says "benchmark unavailable" rather than comparing it against a number of a different kind.

Trend line. Where the suburb has enough recent reporting periods, a small trend line shows the recent median history. It is only drawn from real data points — if a suburb doesn't have enough history, no line is shown (we never draw a flat line to fill the space).

Rental yield

Definition. Gross rental yield: annualised rent as a percentage of the property price — (weekly rent × 52) ÷ price. "Gross" means before costs like strata levies, rates, insurance, and maintenance.

Source. Currently an AI estimate. Licensed vacancy/rent data feeds are under evaluation; until one is licensed for redistribution, yields carry the estimate label and no benchmark. We deliberately do not scrape third-party websites for this data.

Vacancy rate

Definition. The share of rental properties in a suburb sitting vacant and available for rent. Lower is generally better for landlords (stronger rental demand), so where a benchmark exists, a higher vacancy delta is shown red — the colour follows what the number means, not its sign.

Source. Currently an AI estimate, for the same licensing reason as rental yield.

Days on market

Definition. The typical number of days between a property being listed and sold in a suburb. Lower usually indicates stronger demand, so higher deltas are shown red.

Source. Currently an AI estimate — no licensed listing feed is held.

The low-sample badge

Any measured figure built from fewer than 10 observations (for example, a suburb median from fewer than 10 sales in the period) carries a "low sample" badge. Small samples swing hard on individual results — one unusual sale can move a small suburb's median dramatically — so treat badged figures as indicative only. The threshold is reviewed as the dataset grows.

How the AI analysis works

Verdiqt's reports are generated by an AI model. The AI reads what you give it (a suburb name, an address, or the text extracted from an uploaded document) and produces a structured report: scores, findings and plain-English explanations.

What that means in practice — consistent with our AI Analysis Disclaimer:

• General information only, not professional advice. AI output does not constitute financial, legal, building, valuation or investment advice.
• It can be wrong. AI output may be incomplete, inaccurate, out of date, or wrong. It may misinterpret or omit information in an uploaded document, rely on partial documents, produce plausible-sounding but incorrect statements ("hallucinations"), or fail to identify important issues.
• It only knows what it is given. The analysis reflects the information provided at the time — it does not account for a physical inspection, current market conditions, or facts not in the document.
• Document claims are labelled. In document analyses, figures are tagged by how they were obtained: Stated in document (quoted from the text), Calculated (arithmetic on stated figures), or Interpretation (the model's reading). Where a document's dates can be established, findings carry staleness labels — an old AGM or an expired certificate is flagged rather than treated as current, and "date not stated" is flagged too.
• Coverage is disclosed. If only part of a document could be read, the report says which pages were analysed and phrases gaps as "not found in the analysed portion", never "missing from the document".
• Your responsibility to verify. Independently verify any figure or finding before relying on it, and seek professional advice for decisions.