The 5 Metrics That Actually Predict Suburb Booms — And the Ones That Don’t

We started where everyone starts. The metrics investors talk about at meetups. The numbers buyer’s agents put in their slide decks.

This formula tried to predictbooms before they started. Infrastructure projects, population inflows, tightening supply — the theory was that leading indicators could spot a boom 12–18 months out.

A coin flip with seven extra steps. Infrastructure spending doesn’t predict suburb-level booms — it affects whole corridors over decades. Population growth is too slow-moving to time an entry. Vendor discount data barely exists at the suburb level for free.

The biggest lesson: prediction is glamorous. But at suburb granularity with publicly available data, it doesn’t work.

The pivot came from a simple question: what if we don’t try to predict booms before they happen? What if we just detect them once they’ve started?

At first this felt like cheating. Surely the whole point is to be early? But when we looked at actual boom trajectories across Australian suburbs, a pattern emerged: booms are multi-year events. Catching one 6–12 months after it starts still captures 60–85% of the total gains. With dramatically higher confidence.

We replaced prediction indicators with detectionsignals — metrics that answer “is this suburb booming right now?” instead of “will it boom?”

90% accuracy. Zero false positives on 10 known-outcome suburbs. The formula correctly identified Stones Corner (QLD), Armadale (WA), Elizabeth (SA), and flagged controls like Southbank (VIC) and Darwin CBD as stable.

Ten suburbs isn’t enough to trust. We expanded the backtest to 78 suburbs — 28 that boomed, 50 controls that didn’t — and rebuilt the formula with harder data.

The biggest change: we replaced the Absorption component (which needed monthly sales and listings data that doesn’t exist for free) with Growth Strength — scoring annual growth directly. We also added a hard $800K median price cap. Every single boom in the expanded dataset was led by suburbs priced well below the city median.

Accuracy dropped from 90% to 85.7% — but on 8× more data. The 20.2-point separation gap between real booms and false signals means the formula doesn’t just get the right answer. It gets it with conviction.

We also discovered the nastiest silent bug in the formula: missing data defaulting to zero. If a suburb has no days-on-market figure and your parser fills in 0, it scores as the tightest market possible. If median price is missing and defaults to $0, headroom scores perfect. Invisible. Deadly. We caught it only because backtest results on sparse suburbs were suspiciously good.

This is the version we’re most glad we killed.

Detection tells you whethera suburb is booming. It doesn’t tell you which booming suburb will outperform the others. So we built a forecaster — a model that predicts 3-year forward capital growth with confidence intervals for every suburb.

Six input features: annual growth rate, affordability headroom, boom pattern signal, 5-year momentum, national regime indicator, and acceleration ratio. Walk-forward backtest across 28,049 scored rows.

The headline number looked fantastic.

Then we did something most backtests never do. We split the correlation within each scoring period.

Negative 0.08. Worse than random.

The pooled 0.42 was a statistical illusion. Boom years had both higher predictions and higher realised returns. The model was ranking time periods, not suburbs. Within any given month, it couldn’t tell you whether Armadale would outperform Elizabeth or vice versa.

The confidence intervals were just as broken. An 80% interval should contain the real outcome 80% of the time. Ours hit 46%. At the top decile of model confidence — the rows the model was most sure about — coverage dropped to 20%.

We deleted the entire forecaster. Every line of code.

After the forecaster failed, we went back to first principles. What actually discriminates between suburbs after you cancel the market tide?

We tested every feature we had: growth rate, acceleration ratio, repeat-boomer history, days on market, vacancy, yield, rental trend. We computed tide-cancelled excess returns — each suburb’s growth minus the median growth across all peers in the same period — and checked which features predicted outperformance.

One signal survived.

Combine that with the detection formula’s timing signal — is this suburb early in a detected boom, measured by how much of its affordability gap has already been consumed — and you get a two-axis system: when to buy (timing) and what to buy (affordability).

Walk-forward backtest, 12,360 postcode-months, 2012–2026. No lookahead. Excess return = suburb 12-month growth minus market median growth. Full methodology →

Perfectly monotonic. Strong Signal outperforms Good Signal outperforms Fair Signal outperforms Weak Signal. The simplest version of the formula produced the strongest backtest.

Five versions. Here’s what survived backtesting and what didn’t.

You don’t need our formula. The two signals that matter are public knowledge:

That’s the honest version. Two signals. Both free to check. The hard part is doing it across 15,000 suburbs fortnightly, catching the booms within weeks of starting, and filtering out the noise from thin markets. That’s what we automate.

Full backtest methodology, the 78-suburb validation, and the walk-forward tier discrimination results are published on our proof page. No gating, no email required. Check the maths yourself.