A property decision engine that publishes its track record.
97.2% direction accuracy on 144 high-confidence Australian predictions, backtested across 9,663 historical anchor points — and the same investor-framework engine, with transparent pass/fail factor breakdowns, now decoding India (Goa, Bengaluru, Hyderabad, Pune) and Dubai. Built for the regional-over-central investor.
Not investment advice — a data tool. Australia · India · Dubai
New free tool — the 2026-27 Budget CGT impact calculatorPublished evidence
97.2% direction
Live forecasts and audit metrics stay visible so the conversion promise is backed by a public scoreboard.
Published evidence
25pp median error
Live forecasts and audit metrics stay visible so the conversion promise is backed by a public scoreboard.
Published evidence
0.973 calibration score
Live forecasts and audit metrics stay visible so the conversion promise is backed by a public scoreboard.
Published evidence
9,663 backtests
Live forecasts and audit metrics stay visible so the conversion promise is backed by a public scoreboard.
See the current top sub-$400k picks, filtered live from the backend.
The cards below call the existing FastAPI engine. Filter by state to inspect the shortlist the same way a SaaS prospect would.
Investor strategy
Live feed unavailable
Backend warming up — first call can take 60s. Refresh in a moment.
Four steps from raw suburb history to a nightly investment signal.
Step 1
Match
Find 20 historical (postcode, year) lookalikes via 13-feature cosine similarity.
Step 2
Predict
Build the forecast from the similarity-weighted average of lookalike suburb growth.
Step 3
Calibrate
Replay every confidence tier in backtests so HIGH only stays HIGH if the hit-rate survives.
Step 4
Alert
Re-score watched suburbs nightly and flag any move greater than 5 percentage points.
Every prediction can expose the 13-dimensional vector behind it.
The landing page doesn't hide behind vague AI copy. It shows the shape of the model, the calibration discipline, and the audit trail a serious buyer's agent would expect.
“Tap any prediction in the app, see the math behind it.”