Every year produces a fresh crop of “best performing condo” lists, and most of them measure the wrong thing with confidence. They rank asking-price growth, or average PSF movements across a changing mix of units, or they relay a developer’s narrative with a chart attached. This guide sets out the method we consider honest, the one our production analysis is built on, so you can judge our future tables and everyone else’s lists by the same standard.
Why do most “best condo” lists mislead?
Three quiet substitutions do the damage. First, asking prices stand in for transacted prices, which measures seller optimism rather than buyer money. Second, project averages stand in for actual units: when this year’s sales skew toward high floors or larger types and last year’s did not, the “growth” between the two averages is partly a change in what sold, not in what anything is worth. Third, price growth stands in for owner outcome, ignoring when each owner bought, how long they held, and what the rest of the market did meanwhile. None of these substitutions requires bad faith. All of them flatter the seller’s side of the table, which is why a buyer-side method has to be stricter.
What is the honest unit of measurement?
The same unit, bought and then sold. A unit-level repeat sale pins down one owner’s actual entry price, actual exit price, and actual holding period, which yields a realized profit rather than an implied one. Nothing is averaged across unlike units, nothing is inferred from a neighbour’s listing, and the changing mix problem disappears because each unit is compared only with itself. This is the strictest evidence property data can offer, and it is the foundation of our production dataset of 547,000+ transactions. The cost of the strictness is coverage: a unit must have transacted twice to testify. We accept that trade deliberately, because a smaller set of true outcomes protects you better than a larger set of proxies.
Profit compared to what?
A raw profit number, on its own, mostly measures the market the owner happened to sail through. A project where owners made money in a strongly rising market may still have served them worse than the ordinary alternative next door. So each repeat sale is judged as an excess return versus its region’s market over the same holding window: what this unit returned, minus what the regional market did between the same purchase date and sale date. A positive excess return means the project beat the tide it floated on; a negative one means the tide deserves the credit the marketing will claim. This single choice removes the most common illusion in performance lists, the rising market wearing a project’s name.
Why is there no single league table?
Because one table forces false comparisons. A city-fringe two-bedder bought in one part of the cycle and a suburban family unit bought in another are answering different questions for different owners, and ranking them on one axis rewards whichever segment the cycle favoured. Our results are therefore segmented, by district, by unit size, and by entry period, so each project is measured against genuine peers: similar homes, bought under similar conditions. The practical payoff for you is relevance. An upgrader shopping for a three-bedroom home wants the track record of three-bedroom outcomes in that area and era, not a leaderboard headed by an unrepeatable penthouse.
What will the tables look like?
The production tables are not yet connected to this page. The layouts below show the shape of what ships, with every figure deliberately left blank.
Illustrative placeholder: production table computed from our unit-level repeat-sales dataset (547,000+ transactions).
| Segment (district, size, entry period) | Repeat-sale pairs | Median realized profit | Median excess return vs region | Median holding period |
|---|---|---|---|---|
| Segment A | To be computed | To be computed | To be computed | To be computed |
| Segment B | To be computed | To be computed | To be computed | To be computed |
| Segment C | To be computed | To be computed | To be computed | To be computed |
Illustrative placeholder: production table computed from our unit-level repeat-sales dataset (547,000+ transactions).
| Project within segment | Repeat-sale pairs | Share of units sold above region benchmark | Median excess return vs region |
|---|---|---|---|
| Project 1 | To be computed | To be computed | To be computed |
| Project 2 | To be computed | To be computed | To be computed |
A track record, not a forecast
Everything above describes what happened to real owners, and we present it strictly as that. A project with a strong realized record earned it under past supply, past pricing, past interest rates and past neighbourhoods, none of which is under contract to repeat. The honest use of a track record is as a filter and a question generator: it tells you which projects have historically rewarded owners like you, and prompts you to ask what drove that and whether the driver still exists. The dishonest use is as a promise, and we decline to make it, in our tables or in conversation.
The honest caveats
The tables on this page are placeholders; no live figures are published here yet, and nothing above should be read as a performance claim about any project. Repeat-sales evidence covers only units that sold twice, so young projects and long-held units are underrepresented by construction. Realized profit as published will state exactly what it includes, and costs such as taxes, interest and renovation will be handled explicitly when the tables go live. And even the cleanest track record describes owners who bought in a past that is gone. Use this method to sharpen your questions, then run your own purchase through the affordability and upgrade guides before any history, however flattering, moves your money.