Token-Based Valuation · SABCD
Eight pillars. Fifty indicators. One PSS score on a 0–100 scale.
The same scoring engine behind every report we issue — what each pillar measures, how the 50 indicators work, and how a 0–100 PSS resolves to an S–D tier against the full reference pool.
A tier is a discretisation of pool percentile. S is the top 15% of the pool — the "S (top 15%)" the homepage cites.
| Tier | Band | What it means |
|---|---|---|
| S | ≥ P85 | Premier strategic |
| A | P65–84 | High-value core |
| B | P35–64 | Practical technology |
| C | P15–34 | General defense |
| D | < P15 | Foundation layout |
And rolled up to a portfolio — each gauge is the mean PSS across a company's graded patents, resolved to one aggregate SABCD tier. The strip below shows how it distributes across S–D.
Strong · avg PSS
42 graded patents in the 183-patent semiconductor / IC design pool. SEP-dense, deep international families.
Mid · avg PSS
36 graded patents in the 124-patent image-sensor pool. Strong citation impact, narrower geographic reach.
Mid · avg PSS
28 graded patents in the 91-patent biotech pool. Solid core coverage; technically active, defensible.
Weak · avg PSS
19 graded patents in the 67-patent AI-accelerator pool. Narrow scope; largely defensive positioning.
Weak · avg PSS
23 graded patents in the 142-patent power-electronics pool. Broad but low-citation, ageing foundation layout.
Every patent is graded on 50 indicators across 8 pillars. Raw values are normalised inside the full reference pool, aggregated into a 0–100 PSS, and mapped to a SABCD tier — percentile and rank fall out of the same comparison.
| Step | What happens |
|---|---|
| 1 | Structured signals and NLP rubric scores are computed for all 50 indicators. |
| 2 | Each value is ranked against the full pool — a percentile, not an absolute, on one cross-industry benchmark. |
| 3 | Indicators roll up into eight pillars, each independently weighted. |
| 4 | The composite resolves to a 0–100 PSS and one SABCD tier. |
By the nature of each indicator.
| Method | How it works |
|---|---|
| Continuous percentile | Ranked against the pool by tie-aware percentile — most indicators. |
| Scarcity | Rarer is stronger — e.g. triadic (US–EP–JP) filings. |
| NLP rubric | A 1–10 rubric score, then normalised in-pool. |
| Disqualifying screen | A gate, not a score — legal status must be granted and active. |
The PSS is a weighted composite of eight pillars; each pillar decomposes into 4–8 indicators, every one traceable to its formula.
| Pillar | What it measures | Indicators |
|---|---|---|
| P1 · Citation Impact | How much later art leans on it — volume, velocity, quality. | 7 |
| P2 · Legal Strength | Protection breadth and durability — family, claims, remaining life. | 8 |
| P3 · Technical Value | Substance behind the grant — disclosure depth and NLP-graded novelty. | 7 |
| P4 · Market Coverage | Where it is enforceable — GDP-weighted, key-market footprint. | 5 |
| P5 · Text Quality | Claim clarity, support, hierarchy — the pillar only deep review used to reach. | 8 |
| P6 · Network Centrality | Position in the citation graph — PageRank, betweenness, hub/authority. | 4 |
| P7 · Licensing Potential | Whether it can be monetised — detectability, design-around, damages basis. | 6 |
| P8 · Forward-Looking Value | Future trajectory — maturity, citation trend, adoption, runway. | 5 |
Every score traces to a formula. Structured indicators show their raw inputs; NLP indicators carry a rubric score and a reasoning trail you can audit. The pool re-ranks weekly — rank and percentile can move as patents enter or lapse, even when a patent's own indicators don't.
All indicators draw from the Innovue 180M-patent index — 100+ jurisdictions, refreshed daily. Industry is descriptive metadata, not a comparison cohort — ranking is always against the full pool.