How is this different from a Zestimate?
The Zestimate is an automated valuation. It estimates what the house is worth. That's a useful number, but it doesn't tell you whether the seller is likely to negotiate, or by how much. The Tilt Score is about the seller's position, not the house's value — different question, different answer. Two properties at the same asking price with identical Zestimates can have completely different negotiation realities. One owner is underwater on a mortgage and can't come down without short-saling; the other inherited the property, lives in another state, and has already cut the price twice. Zestimate doesn't see that. PriceTilt does.
Are you scraping?
No. PriceTilt uses licensed data from ATTOM Data Solutions and public government data from FEMA. Address resolution uses AWS Location Service. No website scraping is involved. The methodology is deliberate about legal posture because PriceTilt is being built to be acquired — clean data sourcing matters to acquirers, and it matters to us.
What if my agent disagrees with the Tilt Score?
The Tilt Score is one input among several. A good buyer's agent has local market knowledge, knows specific sellers and listing agents, and may have context that public records can't capture. PriceTilt is built to complement an agent's work, not replace it — and a buyer's agent who uses PriceTilt with their clients can demonstrate the data trail behind their recommendations. If your agent disagrees with the score, ask why specifically. That conversation often surfaces the assumption worth understanding.
Why don't you show me the formula?
The methodology is internal IP — the asset PriceTilt is being built around. We show you the inputs (the public-records facts), the directional labels (what's contributing positively or negatively), and a plain-language explanation. We don't show the numeric weights, because that's the part that's defensible against duplication. Same posture as Carfax, FICO, or any other proprietary score.
What does a low Tilt Score mean?
It means the asking price is well-supported by the data — the seller likely has limited room to come down. That isn't bad news; it's actionable information. With a low Tilt Score, your negotiation leverage lives in terms (close date, inspection contingencies, repair credits, inclusions) rather than price. PriceTilt's recommendation for a low score is to focus negotiation on those terms instead of trying to push down the price.
Where does the data come from?
ATTOM Data Solutions — deed records, sale history, tax assessments, owner records, automated valuations, building permits, preforeclosure data. FEMA National Flood Hazard Layer — official US flood zone designations. AWS Location Service — address standardization. We're explicit about every data source because acquirers care, and so should buyers.
Is this financial or legal advice?
No. PriceTilt is a research tool. It surfaces what public records and commercial property data suggest about a property's negotiation context. Decisions about whether to buy, what to offer, or how to structure an offer are between you and your buyer's agent, financial advisor, and attorney.
How accurate is the Tilt Score?
We're early. The methodology is grounded in first principles — what we believe the public records suggest about seller flexibility, based on standard real-estate negotiation patterns. As PriceTilt accumulates scored properties and observes actual sale outcomes via ATTOM's deed-record updates, the methodology will be calibrated against real data. Every score is timestamped with its methodology version, so the audit trail is durable. The honest answer in 2026 is: directionally useful, calibration evolving.
Why don't you factor in race, income demographics, or neighborhood "fit"?
Federal Fair Housing law prohibits using race, national origin, religion, sex, familial status, and disability in real-estate decisions or in services that facilitate them. PriceTilt does not pull these signals, even where commercial APIs offer them. The system has explicit guardrails — both at the prompt level and at the AWS Bedrock guardrail layer — preventing them from entering the methodology or the chat responses. We factor in property-level and seller-level signals (tenure, equity, distress, market dynamics) — never protected-class signals about who lives where.
Can I trust the chatbot?
The PriceTilt chatbot has access to the same data your Tilt Score is built from, plus general real-estate knowledge. We commit to grounded answers about your specific property — if the data doesn't support a claim, the chatbot will say so rather than guess. The chatbot will not disclose the methodology's numeric weights or formula structure (that's protected IP), but it will explain directionally what's driving your score in plain English.