Institutional benchmark OS for crypto fund ICs

Reference

Methodology

UV Terminal publishes the underwriting ontology used to structure deal evidence, valuation context, provenance, and audit records. Calibration data, factor weights, and synthesis prompts are not public.

1. Structural decomposition

Every deal is decomposed into six fields before scoring or memo generation.

Roles

Identifies who supplies work, validates state, governs the system, receives value, and carries execution risk.

Assets

Classifies the token, treasury, collateral, data, infrastructure, and rights that can create or absorb value.

Actions

Maps what users, validators, operators, market makers, and token holders can do inside the system.

Constraints

Records custody, liquidity, transferability, regulatory, unlock, and execution limits before recommendation quality improves.

Privileges

Separates ordinary user rights from privileged control, upgrade authority, routing power, and discretionary allocation rights.

Incentive Links

Tests whether value creation, value capture, token demand, dilution, and participant behavior are linked by observable evidence.

2. Four gating layers

A recommendation can improve only when each institutional gate has enough evidence for review.

Discovery

Evidence satisfies this gate when the asset, venue set, data sources, liquidity rails, and source coverage can be identified and repeated.

Understanding

Evidence satisfies this gate when token rights, value routing, governance authority, and economic exposure can be translated into committee language.

Permissioning

Evidence satisfies this gate when custody, transfer, collateral, policy, and mandate eligibility are known before allocation sizing.

Execution

Evidence satisfies this gate when liquidity depth, slippage, trading schedule, stress loss, and operational path support the proposed action.

3. Seven valuation models

The valuation suite produces a range from complementary models. This page names the models without publishing weights or thresholds.

Metcalfe

Uses network activity and comparable network value; fails when address activity is not economically meaningful.

Discounted Cash Flow

Uses protocol revenue, claim percentage, growth, and discount assumptions; fails when token holders have no credible cash-flow claim.

MV=PQ

Uses transaction economy, market share, and token velocity; fails when the token is not required for settlement or utility.

NVT

Uses on-chain transaction volume and a target network-value-to-transaction relationship; fails when volume is wash, circular, or non-economic.

Stock-to-Flow

Uses issuance scarcity; fails for most non-monetary assets and is primarily relevant to scarcity-driven monetary tokens.

TVL Multiple

Uses locked value, sector multiple, and protocol quality; fails when TVL is subsidized, mercenary, or not linked to token value.

Token Velocity

Uses economic throughput, velocity, holding behavior, and utility premium; fails when holding demand is speculative rather than functional.

4. Provenance and audit

Memo provenance uses provided, computed, and generated labels. Provided fields come from intake or linked evidence. Computed fields come from server-side engines. Generated fields come from memo or narrative synthesis. Decision logs retain the company, action, score, DAQ score, benchmark score, recommendation, actor, and timestamp for export.

Weights, calibration data, and synthesis prompts are not published.