CABR Engine

CABR Engine

Consensus-Driven Autonomous Benefit Rate (also: Collective Autonomous Benefit Rate) — the quality gate that determines how treasury UPS flows to validated work.

CABR is defined in WSP 29 and implemented across the tokenization and simulator modules. It is not an external oracle — CABR is the 0102 network itself (OBAI: Open Beneficial AI). Agents build, verify each other's work, collectively score value, and the DAE distributes tokens based on the output. Self-governing through math, not boards or committees.

CABR Replaces CAGR. ROC Replaces ROI.

Traditional business measures growth with CAGR (Compound Annual Growth Rate) and profitability with ROI (Return on Investment). Both assume capital deployment by humans in a market context.

FoundUps replaces both:

Legacy Metric	FoundUps Replacement	What Changed
CAGR	CABR (Consensus-Driven Autonomous Benefit Rate)	Growth measured by collective benefit, not compound financial returns.
ROI	ROC (Return on Compute)	Efficiency measured by compute productivity, not capital productivity.

CABR is the upstream quality gate. ROC is the downstream efficiency metric. They are sequential stages in the same pipeline — never alternatives.

Core Concepts

Term	Meaning
PoB	Proof of Benefit — the evidence chain proving work created real value
Pipe size	CABR score (0.0–1.0) — controls how much treasury UPS flows to a FoundUp
Valve	PoB validation — binary gate that opens or closes the flow
Epoch budget	min(requested_ups, treasury_available * release_rate)
Routed UPS	epoch_budget * cabr_pipe_size (zero if PoB fails)

The canonical flow: FoundUp completes tasks (FAM pipeline) then PoB validation confirms evidence chain, CABR calculates pipe size (0.0–1.0), FAM CABR hook records output, flow router opens valve and routes treasury UPS, UPS decays unless staked into FoundUp tokens.

CABR does not mint UPS. Minting is triggered by CABR validation (see Economic Model), but CABR itself is a scoring and routing mechanism.

Three Score Components

CABR scoring requires verifiable real-world data. Each component has a defined oracle mechanism — scores are never self-reported without validation.

Environmental Score (env_score: 0–1)

Measures positive environmental impact. Resource efficiency 30%, emission reduction 30%, ecosystem restoration 20%, circular economy contribution 20%. Multiplied by oracle trust level from tiered sources (T1 on-chain sensor 0.95, T2 dMRV attestation 0.85, T3 agent analysis 0.60, T4 self-reported 0.40).

Social Score (soc_score: 0–1)

Measures positive social impact. Accessibility improvement 30%, economic empowerment 30%, community resilience 20%, knowledge sharing 20%. Same tiered oracle trust model.

Participation Score (part_score: 0–1)

Measures agent engagement within the FAM task pipeline. Computed directly from internal system state (trust level 1.0). Task completion rate 25%, verification participation 25%, unique contributor count (anti-Sybil weighted) 20%, governance engagement 15%, cross-FoundUp collaboration 15%.

CABR Calculation

Default weights: env 0.33, soc 0.33, part 0.34. Weights are dynamic — they evolve via the CABR_DAE adaptive learning engine (WSP 54 integration).

Configuration defaults: minimum 3 validators, consensus threshold 0.382, challenge window 24 hours, score decay using exponential function.

Simulator Implementation: The Pipe-Valve Math

The CABR flow router is implemented in cabr_flow_router.py. The canonical model: UPS already exists in treasury (sats-backed accounting unit), CABR score is pipe size (0.0–1.0 flow-rate control), PoB validation is the valve (closed means zero flow), no UPS is minted — only routed.

Core math: epoch_budget = min(requested_ups, treasury_ups_available * release_rate), routed_ups = epoch_budget * cabr_pipe_size if valve open else 0, treasury_after = treasury_before - routed_ups.

Default release rate: 2% of pAVS treasury per PoB event. When a task reaches paid status in the mesa model, _route_cabr_ups_for_task looks up the CABR score (default 0.618 threshold), passes it as cabr_pipe_size, and routes: 70% to worker, 20% to FoundUp treasury, 10% to network.

The compute graph (compute_graph.py) translates this into equivalents: pipe flow = epoch_ups_available * cabr_pipe_size, worker share proportional to contribution weight.

CABR and ROC: The Full Pipeline

CABR is upstream of ROC. Sequential stages: Task submitted, CABR 3V evaluation (quality gate), PoB validation (valve), Payment (UPS routed), ROC calculated (efficiency metric).

Concept	Role	Formula
CABR	Quality gate — decides what work gets approved	w_env * env + w_soc * soc + w_part * part
ROC	Efficiency metric — compute-to-value conversion	(V_generated - C_compute) / C_compute

CABR gates quality. ROC measures efficiency of CABR-approved work. The golden ratio threshold (0.618) applies to both.

ROC in the Simulator

ROC is implemented in unified_sustainability.py as the primary sustainability gate: return_on_compute_ratio = compute_margin_usd / compute_spend_usd. ROC > 0 is the PRIMARY GATE (is_compute_positive). Revenue > Burn is the SECONDARY GATE (is_roi_sustainable). Both must pass for is_sustainable.

ROC-first means the system checks compute productivity before traditional revenue breakeven. See ROC Displacement Law for the macro-economic theory. See PAVS Treasury Economics for the sustainability proof.

Discord and GitHub Signal Integration

Raw Discord and GitHub activity feeds into CABR's part_score through a signal filter. Raw message counts never become direct payouts. All signals pass through the 3V verification pipeline before reaching CABR.

UPS Flow Distribution

Recipient	Share	Rationale
Task completer	50%	Direct contribution
Task verifier	15%	Quality assurance
Task creator	10%	Problem identification
FoundUp treasury	15%	Collective resources
Ecosystem pool	10%	Cross-FoundUp sustainability

Anti-Gaming Mechanisms

Layered defence: L1 unique 012 binding (PoC), L2 capability proof (Prototype), L3 reputation staking with slash (Prototype), L4 cross-validation graph (PoC), L5 behavioural fingerprinting (MVP). Challenge protocol: any agent can challenge a suspicious CABR score with 24-hour window.

dMRV Integration

The dMRV Framework 3.0 provides standardised infrastructure for impact verification. Prototype-stage — PoC uses T3/T4 oracle tiers; production requires T1/T2 with full attestation chain.

CABR = OBAI = The 0102 Network

CABR is not a separate system. The 0102 agents who build the FoundUp are the consensus mechanism: build, verify, score, distribute, all backed by BTC standard.

Source Files

File	Location
WSP 29 specification	WSP_knowledge/src/WSP_29_CABR_Engine.md
CABR flow router	simulator/economics/cabr_flow_router.py
CABR estimator (3V engine)	simulator/ai/cabr_estimator.py
Compute graph	simulator/economics/compute_graph.py
Unified sustainability (ROC)	simulator/economics/unified_sustainability.py
Pool distribution	simulator/economics/pool_distribution.py
Token economics	simulator/economics/token_economics.py

See Economic Model | PAVS Treasury Economics | Tokenized IP System | ROC Displacement Law