HOMEH Research
Methodology: how HOMEH Research computes every figure
Every number we publish is computed in code from closed sales, frozen with the edition, and checked against that frozen dataset before anything goes live. This page explains each step.
How each figure is computed
All figures describe closed sales in one calendar month, grouped by neighbourhood (the MLS® community). Each is a median, so one unusual sale cannot move it far.
- Median sold price
- The middle closing price of every sale in the neighbourhood that month.
- Median days on market
- The middle of each sale's days on market, as reported on its listing.
- Median sale-to-list ratio
- The middle of each sale's closing price divided by its asking price. Above 100% means homes typically sold over asking.
- Over or under asking
- A property of the neighbourhood's median sale, not of every sale: a neighbourhood counts as over asking when its median sale-to-list ratio is above 100%, under asking when below. We write it as "the typical home sold under asking in X% of qualifying neighbourhoods", where typical means the median: the number of such neighbourhoods divided by all qualifying neighbourhoods.
- Share sold over asking
- Within one neighbourhood, the sales that closed above their asking price divided by the sales that had a usable asking price (the same sales the sale-to-list ratio is measured on). A sale with no usable asking price is left out of both counts, never counted as "not over".
- Year-over-year change
- Like-for-like: the change in the median sold price of one property type (detached, semi-detached, freehold townhouse, condo townhouse or condo apartment) against the same type in the same month one year earlier. Both months need at least 10 sales of that type; where several types qualify, the one with the most sales is used and named. A change larger than 40% either way is treated as an outlier and left out unless both months have at least 25 sales; every exclusion is listed in the dataset. An all-homes median is never compared across years, because it moves when the mix of homes sold changes.
- Rankings
- The top and bottom 5 neighbourhoods on each measure. Ties are broken alphabetically.
Figures are rounded the same way everywhere: prices under $1M to the nearest thousand ($845K), from $1M to two decimals ($1.23M), percentages to one decimal (12.3%), and days and counts as whole numbers.
The 5-sale floor
A neighbourhood appears in an edition only if it had at least 5 closed sales in the month. A median of two or three sales says more about those homes than about the neighbourhood, so smaller samples are left out rather than reported with a caveat. The same floor applies to the condo and detached splits in the dataset. Year-over-year changes need more: at least 10 sales of one property type in each of the two months.
Provisional months
Some sales are reported to the board days or weeks after they close. The most recent month in any edition is therefore labelled provisional: its figures may move slightly as late sales arrive. A published edition never changes silently. If a later revision changes a published figure, we issue a correction (below).
Geography
The GTA here means the Toronto, York, Peel, Durham and Halton regions, plus Barrie and Innisfil on the Simcoe edge. Neighbourhood names are the board's own community names, shown with their city, so "Annex" in Toronto is never confused with a same-named place elsewhere. Within Toronto, North York, Scarborough and Etobicoke are named as their own cities.
How every edition is checked
- Frozen dataset. The month's figures are computed once and stored with a checksum. The page, the charts and the CSV are all read from that one record.
- Freshness gate. If the sales feed is behind, or the month's data has not been rebuilt since the month closed, the edition waits.
- Number check. Every number in the headline, summary, text, charts and quotes must match a value in the dataset. A single unmatched number holds the edition.
- Editorial check. Compliance notices, truthful headlines that match the market's direction, no individual address or sale price, and correct attribution of every quote.
Only an edition that passes every check is published. Individual addresses and individual sold prices are never published.
Sources and notices
Our figures are computed from MLS® System data. Every edition carries this notice, with the exact dates of its period:
Based in whole or in part on MLS® System data owned by the Toronto Regional Real Estate Board covering the period the first day of the period through the last day of the period. Information deemed reliable but not guaranteed accurate.
The trademarks REALTOR®, REALTORS® and the REALTOR® logo are controlled by The Canadian Real Estate Association (CREA) and identify real estate professionals who are members of CREA. The trademarks MLS®, Multiple Listing Service® and the associated logos are owned by CREA and identify the quality of services provided by real estate professionals who are members of CREA. Used under licence.
Corrections policy
If a published figure is wrong, we correct it on the same page and say so at the top: what changed and the date. The original dataset is kept on file next to the corrected one, so every version that was ever public can be produced. Report a possible error to [email protected].
How to cite
You may quote, chart and republish any figure with attribution to HOMEH Research and, online, a link to the edition. Each edition page gives its exact citation. The general form:
HOMEH Research. (2026, November 6). GTA Neighbourhood Report: October 2026. HOMEH. https://homeh.com/research/gta-neighbourhood-report/october-2026
For a chart or a broadcast, "Source: HOMEH Research" with the report name and month is enough.
Publishing schedule
- GTA Neighbourhood Report monthly, early in the month after the period closes