How we calculate our numbers

Why we report the median, not the average

Almost every headline price on haho.ca is a median: the midpoint, where half the listings cost more and half cost less. We use it because an average is easily distorted. A single very expensive property, or a handful of commercial or land listings mixed into a residential set, can pull an average well above what a typical home in that area actually costs. The median barely moves.

Where you see an average instead, it is labelled as one. We never quietly swap the two, because they answer different questions.

Market statistics and trends

Market figures are calculated in advance, checked, and then published. Each one has to clear the same set of tests before it appears anywhere on the site:

• A minimum sample. A median calculated from a handful of sales is not a market rate, it is an anecdote. Months without enough sales behind them are withheld rather than drawn on a chart.
• A sanity range. A figure that falls outside the range a real housing market can produce is treated as a data problem, not a discovery.
• A change threshold. When a figure moves sharply against its own recent history, it is held for review instead of published automatically. The previous verified value stays up in the meantime.

A deliberate consequence: a city page will sometimes show no market summary at all. That is the system working. We would rather show nothing than show a median built on three sales.

Which date a sale belongs to

A property sells on one date and legally closes on another, typically one to two months later. We count a sale in the month the deal was agreed, not the month the keys changed hands. Counting by the closing date pushes every sale forward by weeks, which makes the most recent months on any trend chart permanently wrong no matter how much data sits behind them.

Home value and rent estimates

Our estimate is an automated valuation. It compares your property against recent nearby sales and adjusts each one for the ways it differs. It is a starting point for a conversation, not an appraisal, and it is not a commitment from any lender.

How it works:

• Comparable sales are selected from the same area, preferring the same postal area before widening to the wider city.
• Each comparable is adjusted for time. A sale from ten months ago happened in a different market, so it is moved forward using that city's own measured price trend. This is usually the single largest correction.
• Then for the property itself: living area, bedrooms, bathrooms and covered parking. Homes of the same architectural style are preferred, so a bungalow is valued against bungalows.
• Comparables are weighted by how recent, how similar in size, and how close they are. The most similar properties carry the most weight.
• If the property itself has sold before and we can match that history confidently, that prior sale is projected forward and used as an anchor.

The range shown around the estimate is not a fixed percentage. It is calculated from how much the comparable sales disagree with each other. Tightly clustered comparables produce a narrow range; scattered ones produce a wide one, which is itself useful information. Every estimate carries a confidence indicator based on how many comparable sales were available and how closely they agreed.

Numbers we decline to publish

Some figures can be calculated but cannot be trusted, and we leave those blank rather than fill the space:

• A price per square foot from an open-ended size. Some listings state a size only as a band, such as five thousand square feet or more. Dividing a price by a lower bound produces a confident-looking number that can be far off, so we show the size and omit the rate.
• Rates derived from mixed units. Commercial lease amounts are quoted per square foot by some brokerages and per month by others under the same label. We never guess which from the size of the number, so no blended commercial rent figure appears on this site.
• A monthly fee of zero. When a listing states no condominium fee, we show nothing. Showing a fee of zero would be a confident claim about a real carrying cost that the listing never actually made.
• Causes. We describe what changed. We do not infer from a single price movement why it changed, and market descriptions are written only where the sample behind them supports the statement.

How current the numbers are

Active listings refresh from the feed several times a day, and a weekly reconciliation removes listings the source no longer carries. Sold data updates nightly. Where a page shows a market figure, it also shows the period that figure covers. See data sources for where each figure originates and our editorial policy for how written market descriptions are produced and corrected.

Found something that looks wrong?

Tell us and we will check it. Listing data originates with the brokerage that filed it, so some corrections have to be made at the source, but we would rather know. Contact details are on our about page.