How it works · ts-v1.2.0
How the TetherScore works
A short, honest description. The model is a set of reviewed rules, not a validated forecast.
What the number means
A 0–100 index of how strong your current-job position looks, plus your outside options, with a planning focus of roughly the next year. Higher is stronger. It is not a percentage chance of keeping or losing your job, and it has not been validated against real outcomes.
Five factors
- Employer health: public, dated evidence about your employer, such as layoff notices, hiring and financial reports where they exist.
- Your position: the answers you gave about your own role, taken as you gave them.
- Role outlook: demand for your occupation, including what AI exposure means for its tasks.
- Industry & local market: conditions in your industry and your area.
- Rebound strength: how well your confirmed skills and credentials match open roles.
Each factor is scored from evidence and combined into the one number. Tenure, role function, work arrangement and employment type are context only and never add or remove points.
Important events
Some events matter more than any average: a layoff notice that matches your employer and site, a current performance plan you confirmed, or a contract ending soon without a written extension. Each puts a ceiling on the score, and the result says so. A confirmed notice that your own job is ending means no number is shown at all, because any number would mislead.
Missing evidence
Unknown evidence is never filled in with a favourable guess. When something is unknown, the result is labeled Provisional and shows the likely range. When too much is unknown, we say we can't give a number yet, and why. Private employers without public reports are assessed with a declared private-employer method and a lower evidence confidence.
Evidence confidence
High, Medium or Low describes how complete, fresh and well matched the evidence is. It is not the accuracy of a prediction.
AI & your job
The AI risk level combines how exposed your occupation's tasks are, what your employer has announced, and demand for your role, read from both recent job postings and the long-term projection. When those point in opposite directions it says "mixed signals" and shows both. It is a rating based on current evidence, not a prediction, and we do not estimate a personal chance of replacement.
What never counts
Age, race, ethnicity, sex or gender, pregnancy, disability, religion, national origin, marital status, visa status, school or employer prestige, employment gaps, CV polish, courses completed, what other users do, or whether you pay. Checking an action off never changes the score.