BoomAU Past-Performance and Backtest Disclaimer

BoomAU presents two distinct historical results. They measure different things and are not additive.

1. Detection backtest (85.7%).

The “85.7% detection accuracy” claim refers to a v2.3 backtest of 78 Australian suburbs, measuring the proportion of observed boom episodes the model flagged prior to onset within the tested window, alongside 100% specificity and a 20.2-point separation gap against the non-boom control group. The 78-suburb set is a defined research sample, not a forecast of the national market. Accuracy on out-of-sample suburbs, in different macroeconomic regimes, or over different time windows may differ materially.

2. Tier-discrimination backtest (12,360 postcode-months).

The “Strong Signal +7.5pp excess return, 71% beat-market, monotonically down to Weak Signal -6.4pp / 28%” claim refers to a tier-discrimination study over 12,360 postcode-months. “Excess return” is measured against a national benchmark over the matched holding period; “beat-market” is the proportion of postcode-months in the tier that outperformed that benchmark. Internally, BoomAU uses the engine labels Strong Buy, Buy, Watch, and Pass for its research pipeline; these are mapped to the user-facing Strong, Good, Fair, and Weak signal ratings for presentation. The backtest was performed on the engine labels.

Past performance, whether real or backtested, is not a reliable indicator of future performance.

Backtest limitations

Backtest limitations include, without limitation:

• Data lag and revision. Historical real-estate data is frequently revised; backtests use the best vintage available at the time of the study.
• Regime shift. Interest-rate cycles, tax-policy changes, migration shocks, planning reform, and structural changes in credit availability can render historical relationships unreliable.
• Survivorship and selection. Suburb definitions change over time. Postcodes have been amalgamated, split, and rezoned.
• Model evolution. BoomAU's model has been revised over time. Historical results may reflect versions of the model that differ from the current live version.
• Transaction costs. Backtest returns are gross of stamp duty, legal costs, agent fees, finance costs, holding costs, and tax. Real-world net returns for a buyer will be materially lower.
• Non-executability. Backtests assume clean entry and exit at observed price points; real transactions take weeks to months and clear at negotiated prices.

BoomAU does not represent that any subscriber, at any price point, will achieve the backtested results. BoomAU does not represent that the model will detect every boom, avoid every downturn, or correctly rate every suburb it covers.