Find the best buys

Existing condo

The Windsor

Ontario Avenue · Bishan · District 20

: Freehold · Completed 1988 (EdgeProp) · 150 homes

Recent sales, median
S$1,677
3 sales in 12 months
Sales in 12 months
3

PropGPT has not researched this condo the way it researches new launches: the figures here come from URA's sale and rental records, public map data and, for asking prices, EdgeProp. Visit, check the unit and the building, and ask for the management corporation's accounts before you offer.

Is it worth buying?

PropGPT financial score

53out of 100

Fair case on the numbers

Confidence: medium, some figures missing (67% of the weight known)

Adds up the measurable lines below: price against comparable homes, rental yield, the lease, how nearby prices moved, MRT distance and competing supply. 65 or more is a strong case on the numbers, 45 to 64 fair, under 45 weak. It leaves out what only you can judge (layout, school, the view) and is not a valuation or advice.

What makes up the score
Score parts: each part's mark out of 100, its share of the score, and why
PartMark
Price vs similar homes nearbynot countedNot known yet.Not known
Rental yield vs typical25% of the scoreRent of about 2.4% of the price a year; typical is 3.2%.0
Price vs other condos in the town17% of the score3% below the typical price of other condos in the town.61
Value lost to the lease in the next 10 years17% of the scoreFreehold or 999-year: no lease running down.100
Nearby homes that beat the market17% of the score5 of 9 nearby projects beat the market.59
This condo's own prices vs the marketnot countedNot known yet.Not known
Distance to the MRT17% of the score498 m to the nearest station.78
Competing homes finishing nearby8% of the score1,268 competing homes finishing nearby around the same time.37

How the score is worked out

What a property agent would weigh up, from public data. Each line gives a plain answer and what it means for you. Open “Show the calculation” on any line for the exact figures.

PropGPT compares private condos and apartments only (ECs with ECs). HDB flats and landed homes are never in these comparisons, and every price is a real sale recorded by URA, not an asking price.

In your favour 2Neutral 2Watch out 2Not known yet 2

Am I paying a fair price?

  • Price vs similar resale homes nearby · Not known yet

    Not known yet

    Fewer than 10 similar resale condos sold within 1.5 km in the last 12 months, or no price is known yet.

    Show the calculation

    What it compares

    This condo's median price per sq ft (its own 3 resales in the last 12 months) against resale condos and apartments within 1.5 km, sold in the last 12 months, any size. Each resale price is scaled to this condo's lease. Not included: HDB flats, landed homes, homes resold before completion (sub-sales), and this condo itself.

    How PropGPT works out every line

  • Price vs other existing condos in Bishan · Neutral

    About 3% below the typical price of other existing condos in Bishan

    Priced like most of the homes you are likely to compare it with.

    This condo
    S$1,677 psf
    Typical in Bishan
    S$1,735 psf

    Based on 22 condos · a rough guide only

    Show the calculation

    What it compares

    Each of this condo's own resales (its own 3 resales in the last 12 months) against resales in other existing condos in the same town (URA planning area) of a similar size (within 20%), sold in the same period, each scaled to this condo's lease, and to its building age (completed 1988) where the other condo's completion year is known.

    The figures

    • This condo: S$1,677 psf
    • 3 of its 3 own sales are each compared with the median of sales of a similar size in 22 other existing condos in Bishan. The gap is the median of these per-sale gaps; S$1,735 psf is the town price that gap implies.
    • Age adjustment: 22 of the 22 other condos are also scaled to this condo's age (completed 1988), about 0.9% per year of difference.
    • In your favour when 10% or more below the typical price; watch out when 10% or more above.

    How fair is this comparison?

    A rough guide only. Same town, lease and building age adjusted, each sale against homes of its own size; but condition and distances to the MRT differ within a town.

    What this check cannot see

    • Building age is adjusted at one average rate, about 0.9% a year (measured from 10,793 private resales in 1,531 condos, Oct 2025 to Oct 2026). Real prices fall faster in a building's first 15 years and slower after, faster for 99-year leases than freehold, and upkeep differs between buildings.

    How PropGPT works out every line

Can it earn or hold its value?

  • Rental yield · Watch out

    About 2.4% of the price a year in rent, before costs

    Less than the typical 3.2% for existing condos, so rent would cover less of your costs.

    This condo
    2.4%
    Typical for existing condos
    3.2%

    Based on 178 rental contracts, Q1 2026 to Q3 2026

    Show the calculation

    What it compares

    Rents of condos and apartments of similar size within 1.5 km (URA rental contracts, last 4 quarters), per sq ft, × 12, against this condo's price per sq ft.

    The figures

    • Nearby rent: S$3.38 per sq ft a month, median of 178 contracts (URA, Q1 2026 to Q3 2026)
    • Yield = S$3.38 × 12 ÷ S$1,677 psf = 2.4%. In your favour when 0.3 percentage points or more above the typical yield; watch out when 0.3 or more below.

    How fair is this comparison?

    Fair. Rents of similar homes nearby, not of this condo.

    What this check cannot see

    • Gross yield: property tax, maintenance fees, agent fees and empty months usually take 1 to 2 percentage points off.
    • URA records rented sizes as ranges (for example 900–1,000 sq ft); PropGPT uses the middle.
    • Building age is not adjusted: a 25-year-old condo and a 3-year-old one count the same, so part of any gap is simply new against old.

    How PropGPT works out every line

  • Lease · In your favour

    Freehold

    No lease runs down, so the home does not lose value as the years pass. Buyers usually pay more for this.

    Based on URA's title record

    Show the calculation

    What it compares

    The title's tenure, as recorded by URA.

    How fair is this comparison?

    Very fair. A fact about this home.

    What this check cannot see

    • Freehold land does not protect the building: it still ages, and an en bloc sale needs most owners to agree.

    How PropGPT works out every line

  • How nearby homes' prices moved · Neutral

    5 of 9 nearby projects rose faster than the whole market

    Homes around here have moved roughly with the market.

    Based on 9 nearby projects, sales from Oct 2021 · a rough guide only

    Show the calculation

    What it compares

    For each project within 1.5 km with enough sales: the median price of its first 12 months of sales in PropGPT's data against its resales in the last 12 months, at least 2 years apart, compared with URA's private non-landed price index over the same quarters.

    The figures

    • THREE 11 (141 m): +11% from 2022-Q2 to 2026-Q3; market +21%
    • THOMSON THREE (439 m): +24% from 2022-Q1 to 2026-Q2; market +25%
    • THOMSON IMPRESSIONS (473 m): +23% from 2022-Q2 to 2026-Q3; market +21%
    • BISHAN POINT (665 m): +45% from 2022-Q2 to 2026-Q2; market +21%
    • THOMSON GRAND (980 m): +18% from 2022-Q2 to 2026-Q1; market +21%
    • THE GARDENS AT BISHAN (982 m): +28% from 2022-Q2 to 2026-Q2; market +21%
    • In your favour when 60% or more of nearby projects beat the market; watch out at 40% or fewer.

    How fair is this comparison?

    A rough guide only. Each project against itself over time, so like for like within a project; but the projects are not this one.

    What this check cannot see

    • PropGPT's sales data starts in October 2021, so an older project's 'first year' is a resale year, not its launch.
    • A project's early sales may be small units and later ones large, or the other way round.
    • Past price moves do not predict future ones.

    How PropGPT works out every line

  • How this condo's prices moved · Not known yet

    Not known yet

    Fewer than 5 sales early and 5 sales recently, at least 2 years apart.

    Show the calculation

    What it compares

    This condo's median resale price per sq ft in its first 12 months in PropGPT's data (from October 2021) against its last 12 months, compared with URA's private non-landed price index over the same quarters.

    How PropGPT works out every line

  • MRT · In your favour

    498 m to Upper Thomson MRT Station

    Within an easy walk; homes near stations are in steady demand from buyers and tenants.

    Based on LTA station exits and URA Master Plan 2019, worked out 4 Nov 2026

    Show the calculation

    What it compares

    Straight-line distance from the project's map point to the nearest MRT or LRT exit in service (LTA), and to the nearest planned station (URA Master Plan 2019).

    The figures

    • Nearest planned station: Teck Ghee (planned), 2,080 m
    • In your favour within 500 m of a station in service or planned; watch out over 1 km from both.

    How fair is this comparison?

    Very fair. A fact about this location.

    What this check cannot see

    • Straight-line distance: walking routes are longer, often by a third.
    • Planned stations can be years away and dates change.

    How PropGPT works out every line

What could go wrong?

  • Competing homes finishing nearby · Watch out

    1,268 homes in 1 other project within 1.5 km due 2026–2030

    Many owners nearby may try to rent out or sell at the same time, which can hold prices and rents down.

    Based on URA's list of projects not yet completed, worked out 4 Nov 2026

    Show the calculation

    What it compares

    Homes in other private projects within 1.5 km that URA lists as not yet completed, due in the next few years.

    The figures

    • URA pipeline of private homes not yet completed.
    • In your favour when there are none; watch out at 1,000 homes or more.

    How fair is this comparison?

    Fair. Counts homes, not who will sell or rent them.

    What this check cannot see

    • Completion years in URA's pipeline often shift by a year or more.
    • Land sold but not yet in URA's pipeline is not counted.
    • 1.5 km in a straight line can reach across an expressway, river or park into a different market.

    How PropGPT works out every line

  • Cost to buy, hold and sell · Fact

    S$79,600 in stamp duty for the estimated price

    Profit only starts after these costs are covered.

    Based on today's IRAS stamp duty rates

    Show the calculation

    What it compares

    Stamp duties under the current IRAS rates for the price shown.

    The figures

    • Stamp duty on S$2.20M: S$79,600 (assumes a Singapore citizen buying a first home)
    • Selling within 4 years of buying costs seller's stamp duty of 16% falling to 4%; buying now, that ends around Nov 2030.
    • A completed home: the full loan starts at once, and a buyer's agent fee (often 1%) is usual if you use one.

    How fair is this comparison?

    Very fair. Worked out for this price.

    What this check cannot see

    • Assumes a Singapore citizen buying a first home until you add your details.

    How PropGPT works out every line

Good to know (not judged)

  • Sales in the last 12 months · Fact

    3 resales

    More sales mean a clearer price and an easier sale later. Few sales make every figure here less certain.

    Based on URA sales, Nov 2025 to Nov 2026

  • Completed · Fact

    1988

    Older buildings need more upkeep; very old ones may be en bloc candidates.

    Show the calculation

    The figures

    • From EdgeProp, checked 2026-10-22.

    How PropGPT works out every line

PropGPT calculations from URA and data.gov.sg, updated 3 Nov 2026. PropGPT is not a licensed estate agency and does not act for buyers or sellers. These figures are not a valuation, a forecast or financial advice, and past results nearby do not predict future returns. How each line is worked out

What it costs

Estimates from what homes here actually sold for. URA records each home's size, not its bedrooms, so the bedroom count is a guess from floor area.

Size (bedrooms, a guess)Typical sizeMedian priceRangeSales
3-bedroom 1,000–1,399 sq ft1,345 sq ftS$2.20MS$2.20M1 in 24 months
4-bedroom 1,400–1,899 sq ft1,620 sq ftS$2.68MS$2.48M to S$2.99M6 in 24 months
5+-bedroom 1,900 sq ft or more2,223 sq ftS$3.63MS$3.15M to S$4.12M2 in 24 months

Asking prices now

  • 3-bedroomS$2.55M S$1,823 psf
  • 3-bedroomS$2.76M S$1,745 psf
  • 4-bedroomS$4.50M S$1,883 psf
  • 5-bedroomS$5.60M S$1,499 psf

A sample of listings on EdgeProp, checked 2026-10-22. Asking prices are what sellers hope for; homes usually sell for less. Listings may have sold since.

See current listings on EdgeProp, 99.co or PropertyGuru.

Stamp duty on S$2,200,000: S$79,600 (assumes a Singapore citizen buying a first home). A completed home: the full loan starts at once.

Recent sales

Every resale and sub-sale recorded by URA, newest first. URA gives floors as a range of five.

MonthSizeFloorPricePSF
Aug 20261,658 sq ft01-05S$2,780,000S$1,677
Mar 20261,345 sq ft01-05S$2,200,000S$1,636
Jan 20262,454 sq ft01-05S$4,120,000S$1,679
Nov 20251,485 sq ft01-05S$2,480,000S$1,670
Nov 20252,497 sq ft01-05S$3,900,000S$1,562
Aug 20252,497 sq ft01-05S$3,820,000S$1,530
Jul 20251,991 sq ft01-05S$3,148,000S$1,581
Apr 20251,830 sq ft01-05S$2,990,000S$1,634
Feb 20251,582 sq ft01-05S$2,580,000S$1,631
Jan 20251,572 sq ft01-05S$2,535,000S$1,613
Jan 20251,798 sq ft01-05S$2,778,000S$1,545
Oct 20241,550 sq ft01-05S$2,450,000S$1,581
Sept 20241,798 sq ft01-05S$2,898,888S$1,612
Aug 20241,572 sq ft01-05S$2,500,000S$1,590
Jul 20241,625 sq ft01-05S$2,400,000S$1,477
Jul 20241,302 sq ft01-05S$1,950,000S$1,498
May 20241,378 sq ft01-05S$2,075,000S$1,506
Aug 20232,454 sq ft01-05S$3,410,000S$1,390
Mar 20232,013 sq ft01-05S$2,950,000S$1,465

Location

Nearest mrt station
Upper Thomson MRT Station
498 m · ~7 min walk
Nearest school
Ai Tong School
750 m · ~10 min walk
Nearest park
Venus Green
215 m · ~3 min walk
Nearest supermarket
Ntuc Fairprice Co-Operative Ltd (Thomson Plaza)
448 m · ~6 min walk
The WindsorMRT stationSchoolParkSupermarket

URA gives one map point per condo. Open data from LTA, NParks, MOE and SFA (data.gov.sg). Map: OpenFreeMap, OpenStreetMap contributors.

What next?

Compare The Windsor with other condos in Find the best buys. Before you make an offer, view the unit and check the price against the recent sales above.

Sign in to start buying a home hereEvery step from the option to purchase to completion, with deadlines and amounts. You do not need an agent of your own.