Methodology

Informational only — not financial advice. No forecast is certain or guaranteed.

What we forecast

We publish interval-valued forecasts for housing metrics (home values, rents, sale prices, supply) across Southern California — region, county, city, and neighborhood. Every forecast is shown with a prediction interval (a low–high range) and a confidence score. We never present a single number as a sure thing.

Data trust tiers

We tag every observation by source trust tier and weight them accordingly:

• Tier 1 — recorded transactions (county deeds, FHFA): ground truth. Weighted most heavily.
• Tier 2 — modeled estimates (e.g. Zillow ZHVI, Redfin): useful but smoothed and revised.
• Tier 3 — listing / asking / vendor data: the most biased, and where promotional "shill" signals appear.

Shill detection

We continuously compare each source against Tier-1 reality and measure its bias. Where listing/vendor data diverges from recorded transactions, we flag it and down-weight that source in the ensemble. Those divergences are surfaced on the homepage.

Intervals & confidence

Intervals are calibrated against historical outcomes (the realized share of outcomes that land inside the stated band is reported on the track record page). Confidence generally falls at finer geographies, where data is sparser.

Limitations

Forecasts can and do miss. Markets are affected by rates, policy, and shocks we cannot anticipate. Treat everything here as one input among many — not advice.