The same CRS for every viewer
One additive score from the same eligible evidence and formula version: a profile foundation plus every eligible review. A recruiter cannot privately rewrite someone’s canonical score.
Career Rank Score / CRS v5
One score.
Many lenses.
CRS adds eligible career evidence into one transparent, viewer-independent score. Open Algorithm lets you decide how to filter and sort that evidence—without changing the person’s CRS.
Deterministic, versioned, explainable. Decision support, never an automatic hiring decision.
Score active
crs-v5
726points · no ceiling
The separation that matters
A measurement is not a decision.
CRS, the evidence beneath it, and your search preferences are three distinct things. Keeping them separate makes the system easier to understand, challenge, and improve.
Counts stay visible
Eligible review count, distinct reviewers, foundation completion, last evidence date, and native 0–5 averages sit beside the score, so a large total is never read as quality without context.
Your policy, clearly labeled
Open Algorithm filters and sorts for your use case. The policy is deterministic, saved by scope and version, and shown beside its result.
How the score is built
Career evidence, added in the open.
The current CRS formula adds a profile foundation—dated work, completed education, and skills or certificates—to the raw sum of all eligible submitted four-dimensional reviews. The score is unbounded above, with no renormalization: a missing component adds nothing rather than being scaled up.
Each 4D review
Every eligible submitted review adds its raw Rating + Integrity + Intelligence + Energy, each on the native 0–5 scale. Private and anonymous complete reviews still count. No review means unrated—not automatically poor.
Dated work
Work-experience records with usable dates, merged so overlapping roles never count twice; reaches 150 points at 15 merged years. Tenure is evidence context, not account age or a proxy for a protected trait.
Education
The strongest completed degree level supplies most of this component. When a supported grading scale is usable, normalized GPA supplies the remainder; a missing GPA is neutral. School prestige is not scored.
Skills & certificates
Declared skill and certificate evidence up to fixed saturation thresholds, with verification coverage exposed separately from the canonical score.
Current published formula: crs-v5 (page updated 2026-10-13). Canonical CRS = profile foundation (up to 300 points) + the sum of every eligible review (up to 20 points each), rounded to a whole point and unbounded above. No multiplier for relationship, reviewer CRS, verification, recency, account tier, employer, or institution changes those points. Until the first eligible review, the foundation appears as a faded potential CRS. Education uses degree level for 80% of its component and normalized GPA for 20% when a 4, 5, 10, 20, or 100-point scale is usable. Historical crs-v4 snapshots (0–1000, percentage weights) remain readable in score history and are never relabeled as v5. Formula versions and score snapshots preserve reproducibility.
Open Algorithm / Build your own algorithm
A filter you can explain.
“Build your own algorithm” sounds complex. In practice, it is a saved set of evidence choices, filters, thresholds, weights, and a sort metric. It does not need a language model or AI credits, and it does not create a second CRS.
Choose a preset lens
Watch the same fictional profiles reorder while their canonical scores stay fixed.
Canonical CRS order
Same viewer-independent, unbounded score for everyone.
Profile Aillustrative data
726canonical CRS
296verified-only
60.5per verified yr
Profile Billustrative data
655canonical CRS
—verified-only
131.0per verified yr
Profile Cillustrative data
588canonical CRS
402verified-only
84.0per verified yr
Profile Dillustrative data
471canonical CRS
344verified-only
117.8per verified yr
Illustration only; no real candidate data. A profile with no booking-verified review drops out of the verified view rather than scoring zero. “Score per verified year” is one optional ordering lens, not a claim that a higher rate means a better person or hire.
What a saved policy can contain
Useful controls, with provenance.
A lens can apply to you, a company, or a job. Server-side authorization controls who can create or change it, and immutable versions preserve what was used.
Use all eligible evidence or verified-only evidence (reviews backed by a paid Market booking), and optionally ignore reviews by a stated reviewer relationship, such as friends. Verification and relationship are lens choices, never hidden multipliers in the canonical score.
Minimum or maximum CRS, confidence, verification coverage, and verified work-year thresholds.
Canonical score, explicit Rating, Integrity, Intelligence, Energy, and foundation weights with total or average review aggregation, or a derived score-per-verified-year ordering.
See canonical CRS, the lens value, review counts, policy version, and explanation for each ranked result.
Human-decision boundary
Signal in. Judgment stays human.
CRS can help people inspect and organize evidence. It cannot establish suitability for a role by itself, replace due diligence, or make an employment decision.
Designed to do
- Show the point budgets, evidence components, and formula version behind a score
- Expose review and distinct-reviewer counts, verification coverage, freshness, and missing inputs
- Apply reversible filters and explain why a result appears
- Let a person correct source profile data through Sorcrr controls
Not designed to do
- Use age, account age, salary, referrals, or institution prestige in CRS
- Use protected traits as Open Algorithm filters or weights
- Automatically reject, hire, fire, or otherwise decide for a human
- Hide a viewer’s ranking policy inside the canonical score
- Rank Market providers; Market shelves show native per-industry ratings, not CRS
Questions worth asking
CRS, without the black box.
For operational details, authenticated Sorcrr users and internal agents can use the canonical wiki article linked below.
Is CRS an AI score?
No. CRS is calculated by a deterministic, versioned formula. SAI may explain a score or help a user configure a lens, but the score and ranking tool do not require a language model or AI credits.
Can a hiring team change my CRS?
No. A viewer can save a ranking policy for their own authorized scope, but that policy does not rewrite your canonical CRS. The original score remains visible alongside any derived ranking value.
Why might a score be unavailable or stale?
CRS exposes availability states such as available, unrated, stale, or unavailable. Until the first eligible complete review, a profile shows its foundation as a faded potential CRS and stays unrated. A profile may also have evidence awaiting recomputation or be outside an authorized view.
Is there a maximum?
No. CRS is unbounded above. The profile foundation is capped at 300 points (150 work, 90 education, 60 skills and certificates), and each eligible review adds up to 20 points, so the score keeps building with every eligible review. A valid low review adds fewer points; it never subtracts earned evidence. Because a large total alone is not quality, CRS is always shown with its eligible review count and distinct reviewer count.
What changed from v4?
crs-v4 was bounded to 0–1000, combined percentage weights (70% reviews, 15% dated work, 9% education, 6% skills), and renormalized missing components. crs-v5 adds fixed point budgets to the raw sum of every eligible review, with no ceiling, no renormalization, and no multipliers. Historical v4 snapshots remain readable in score history under their own formula version.
What should an employer do with a ranking?
Review the evidence, job requirements, accommodations, and relevant context; keep a qualified human in control; and meet any notice, audit, assessment, or appeal duties that apply in the employer’s jurisdiction.
Transparent by construction
Bring your lens. Keep the evidence visible.
Use CRS to understand a career evidence snapshot, then use Open Algorithm to make your own priorities explicit—not to hide them.