# Evolution 16 — valuation methodology

**Everything in this file and in `ip_register_delta.md` is an internal analytical scenario. It is not an appraisal, not
a financing valuation, not a market price, not a legal conclusion and not a guarantee.**

## Method (unchanged from register v46)

1. The previous register's totals are taken as they are. Evolution 16 does not recompute, re-tier or re-discount any
   earlier family. `CAIN_IP_VALUE_RECALC_V46.json` and the v41 company record are byte-identical before and after
   (checked by hash in this session).
2. Each newly implemented and tested mechanism family is assigned a tier from the register's existing table
   (`scripts/ip/recalculate_ip_value.py`, `TIER`):

   | Tier | LOW | BASE | HIGH |
   |---|---|---|---|
   | NOVEL | 40,000 | 90,000 | 130,000 |
   | UNUSUAL | 20,000 | 45,000 | 85,000 |
   | ENGINEERING | 5,000 | 12,000 | 20,000 |
   | INTEGRATION | 3,000 | 8,000 | 15,000 |

3. The deployed discounts are applied to the gross: 22% LOW, 12% BASE, 5% HIGH.
4. Net is added to the previous total.

## How tiers were assigned in Evolution 16

- No family was given NOVEL or UNUSUAL. The research found close published work for every mechanism (Atomix, Cordon,
  SagaLLM, Temporal, AP2, event sourcing) and overlapping families already in the register. With that overlap, and
  with no full-text reading, no patent search and no legal review, a higher tier is not supported.
- ENGINEERING: a mechanism with a rule the prior art, as far as it was read, does not state. Three families.
- INTEGRATION: an existing CAIN mechanism bound to the new journal, or a known technique applied to this runtime.
  Three families.
- Each family is counted once. Where one mechanism serves two purposes (delegation re-check at dispatch is both
  "delegation" and "dispatch") it is counted under one family only.
- Remediations (F16-01, F16-02), parameters (the evidence validity window) and capabilities that already existed
  (self-improvement lifecycle, delegation attenuation, drift classification) are counted as zero.

## Scenarios A-E requested by the specification

| Scenario | Treatment |
|---|---|
| A. Standalone technical IP | computed: `ip_register_delta.md` |
| B. Product portfolio commercialization | not recomputed. No product was added to the Top-100 registry and no price was set, so there is no arithmetic to change |
| C. Revenue-supported business | not recomputed. Revenue evidence is the founder attestation already in the repository; Evolution 16 produced none |
| D. Acquisition or strategic value | carried through by the existing company formula only (IP base + the existing premiums); no new assumption. At the owner's instruction of 2026-10-12 the company and founder records were rolled forward to v42 and the two canonical founder files were updated |
| E. Independent professional appraisal | **none has been commissioned or completed** |

## What the number is sensitive to

The tier table and the discounts dominate: the increment is six table lookups. The table is the register's convention,
not a market observation. Nothing in Evolution 16 measured willingness to pay, licensing interest or replacement
cost. If counsel finds that the three ENGINEERING families are anticipated by the cited prior art, they would be
INTEGRATION under this same method.

## What would raise the evidence level

E4 today (deterministic, adversarial and regression tests executed by the author). E5 needs someone else to run
`EVOLUTION16_REPRODUCE.sh` and the verifier on their own machine and report. E6 needs an external technical review or
a real deployment. E7 needs a customer.