The model
How we predict when a homeowner will sell
We score every home in the county on how likely it is to sell in the next year, and we show our work. Here is what the score means, how we test it, and why you can trust the order it puts your database in.
4 in 10
Of the homes we mark Hot sell within a year
That is about 9 times the rate of an average home.
1 in 23
Homes in Cuyahoga County sell in a typical year
That is what an agent working an unranked database is up against.
9x
More likely to sell than the average home
Hot means a home scored at least 5 times the county rate, the same rule the app uses.
500,000+
Homes in the county, every one of them scored
Tested on 1.4 million homes the model had never seen.
We trained the model on years of county history, then turned it loose on the next year's worth of homes it had never seen — 1.4 million of them. These are the results, not a projection. Separately, in 16 quarterly tests since 2021, the top of our list beat the county rate by more than 3 times every single time. Results are for Cuyahoga County, Ohio, where the model has the deepest history.
The accuracy gap
They flag half the market to catch 72% of sales. We catch nearly half the sales flagging 15% of it.
The best-known accuracy claim in this industry is “72%.” By its own published definition, that means flagging half the homes in a market and catching 72% of the sales that follow — barely better than splitting the phone book. Across roughly thirty vendors selling seller predictions, not one publishes how their model was tested, and exactly one independent test has ever been run: commissioned by the vendor itself, in 2019.
Here is ours, measured on homes the model had never seen:
- Marked HotAbout 4 in 10 of the homes we mark Hot sell within a year.
- Every timeIn 16 separate quarterly tests since 2021, the top of our list sold at more than 3 times the county rate every single time.
- Everyone elseIn a typical year only about 1 in 23 homes in Cuyahoga County sells.
And the scores mean what they say. When we say a home has a 4-in-10 chance, about 4 in 10 of those homes sell. Where the property records are thin we round down rather than up — we would rather under-promise on a home than send you after one that was never moving.
Trained on history, not hindsight
The training set is built from quarterly snapshots of every property in the county. For each snapshot, features are computed looking strictly backward — only transfers, liens, and valuations recorded before that date — while the label looks strictly forward, asking whether the property actually sold in the following year.
That boundary is what makes the accuracy number meaningful. A model that can see the future during training will always look excellent in testing and fail in production.
Only real sales count
Public records are noisy. Properties move between family members, into trusts, and through foreclosure — none of which is a listing an agent could have won.
The model counts only arm's-length market sales, filtering out quit-claim transfers, sheriff's deeds, and nominal transactions. What it predicts is a real transaction, which is the only kind worth calling about.
Every score explains itself
Each prediction returns the five factors that moved it most, expressed as how much each one shifted the probability. An agent opening a contact does not just see 74% — they see that the owner has held the property eight years, that equity crossed 60%, and that homes on the street have started turning over.
A score without reasoning tells you who to call. A score with reasoning tells you how to open the call.
Score tiers
Every tier is measured against the county's own rate — the roughly 1-in-23 chance that any given home sells in a year. Hot means a home scored at least 5 times that rate, the same rule the app uses, so “Hot” means the same thing in a busy market and a slow one.
Where it runs today
Results are for Cuyahoga County, Ohio, where the model has the deepest history — enough deed, mortgage, and assessment records to see a homeowner's whole story. New markets are tested the same way before agents see a score.
We would rather tell you a market is not ready than ship you a number we cannot stand behind.
Ask about your market