# CR141

> CR141 verifies the account-level consequences of DeFi governance parameter changes
> before execution. Validated by exact replay of 28 executed Compound governance
> proposals carrying 141 collateral-factor changes: 89,645 analyzed accounts,
> 474 predicted and 474 observed account state transitions, difference 0.

Category: Governance Impact Verification / DeFi Risk Modeling
Status: audit CLOSED — 28/28 proposals exact, 474/474 transitions, 0 difference
Build: cr141-v1.3.2-openapi at commit 412d627

## Core facts

- Proposals replayed: 28
- Collateral factor changes: 141
- Analyzed accounts: 89,645
- Predicted account state transitions: 474
- Observed account state transitions: 474
- Difference: 0
- Proposals with exact set equality: 28/28
- Proposals producing a non-empty transition set: 14/28
- Transitions attributable to more than one market: 92
- Transitions where no single market alone flips the account: 14

## Proposal index

Every proposal in the corpus. Predicted and observed are account state transitions;
"exact" means the two sets are identical. This is the same table the dashboard renders,
written as text so it can be read without JavaScript.

pid | exec block | markets | targets | predicted | observed | exact
44 | 12370865 | 1 (cREP) | 1000 | 6 | 6 | yes
56 | 13039770 | 5 (CAAVE,CLINK,CMKR,CSUSHI,CYFI) | 223 | 0 | 0 | yes
66 | 13583937 | 6 (CAAVE,CLINK,CMKR,CSUSHI,CUSDC,CYFI) | 5334 | 0 | 0 | yes
69 | 13635170 | 4 (CLINK,CMKR,CSUSHI,CYFI) | 484 | 0 | 0 | yes
71 | 13699346 | 4 (CBAT,CDAI,CETH,CZRX) | 8584 | 9 | 9 | yes
72 | 13765755 | 3 (CCOMP,CUNI,cWBTC) | 3157 | 0 | 0 | yes
74 | 13841132 | 3 (CBAT,CLINK,CWBTC) | 3334 | 0 | 0 | yes
81 | 14030172 | 1 (CETH) | 6937 | 0 | 0 | yes
85 | 14292829 | 6 (CAAVE,CLINK,CMKR,CSUSHI,CUNI,CYFI) | 1635 | 0 | 0 | yes
90 | 14430843 | 3 (CMKR,CSUSHI,CZRX) | 1675 | 0 | 0 | yes
98 | 14563136 | 2 (CSUSHI,CYFI) | 170 | 0 | 0 | yes
103 | 14700684 | 2 (CAAVE,CMKR) | 205 | 0 | 0 | yes
107 | 14879331 | 5 (CDAI,CLINK,CMKR,CUSDC,CYFI) | 7332 | 0 | 0 | yes
108 | 14959593 | 5 (CAAVE,CLINK,CSUSHI,CUSDC,CYFI) | 5112 | 0 | 0 | yes
111 | 15077600 | 2 (CDAI,CUSDC) | 7055 | 0 | 0 | yes
133 | 15911677 | 1 (CSUSHI) | 168 | 0 | 0 | yes
186 | 18385597 | 7 (CAAVE,CCOMP,CLINK,CMKR,CSUSHI,CYFI,CZRX) | 2674 | 21 | 21 | yes
190 | 18519451 | 7 (CAAVE,CCOMP,CLINK,CMKR,CSUSHI,CYFI,CZRX) | 2666 | 19 | 19 | yes
193 | 18612241 | 7 (CAAVE,CCOMP,CLINK,CMKR,CSUSHI,CYFI,CZRX) | 2660 | 21 | 21 | yes
200 | 18762889 | 7 (CAAVE,CCOMP,CLINK,CMKR,CSUSHI,CYFI,CZRX) | 2641 | 22 | 22 | yes
213 | 19062517 | 8 (CAAVE,CBAT,CCOMP,CLINK,CMKR,CSUSHI,CYFI,CZRX) | 3320 | 45 | 45 | yes
214 | 19161536 | 8 (CAAVE,CBAT,CCOMP,CLINK,CMKR,CSUSHI,CYFI,CZRX) | 3317 | 47 | 47 | yes
217 | 19262734 | 8 (CAAVE,CBAT,CCOMP,CLINK,CMKR,CSUSHI,CYFI,CZRX) | 3312 | 33 | 33 | yes
243 | 19760928 | 9 (CAAVE,CBAT,CCOMP,CLINK,CMKR,CSUSHI,CUNI,CYFI,CZRX) | 3596 | 38 | 38 | yes
252 | 19960052 | 8 (CAAVE,CCOMP,CLINK,CMKR,CSUSHI,CUNI,CYFI,CZRX) | 2926 | 35 | 35 | yes
295 | 20470106 | 1 (CCOMP) | 1023 | 53 | 53 | yes
301 | 20519734 | 9 (CAAVE,CBAT,CLINK,CMKR,CSUSHI,CUNI,CUSDC,CYFI,CZRX) | 5468 | 65 | 65 | yes
352 | 21063478 | 9 (CAAVE,CBAT,CLINK,CMKR,CSUSHI,CUNI,CWBTC,CYFI,CZRX) | 3637 | 60 | 60 | yes

## Method

- Population: accounts that entered a changed market AND carried debt.
- Collateral buffer computed at block b-1, fixed-point integer arithmetic (1e18).
- No sampling, no Monte Carlo, no randomness. Same input, same output.
- Falsification criterion: exact set equality between predicted and observed
  transitions. One mismatch demotes the result; no averaging, no tolerance.
- Route A / Route B cross-check: Route A truncates in integers, Route B does not.
  Every disagreement is a truncation artefact; the flip sets agree exactly.

## Pages

- [Dashboard](https://cr141.com/): five surfaces — proposal explorer, ranking, flip explorer,
  what-if scenarios, validation gates. Static, no network at read time.
- [About](https://cr141.com/about): definition, method, scope, limits, FAQ.
- [Dataset](https://cr141.com/dataset): the historical replay dataset and what it contains.
- [API](https://cr141.com/api): the four endpoints in human language.
- [For AI agents](https://cr141.com/for-ai-agents): this content as plain text.
- [OpenAPI 3.1 document](https://cr141.com/openapi.json): the machine-readable API contract.

## API

Status: the OpenAPI 3.1 document is public; the API host is not. `openapi.json` is complete
enough to generate a client without calling anything. Run `docker compose up -d` to get the
same endpoints on http://localhost:8000 with Swagger UI at /docs. No key, no token, no account.

OpenAPI document: https://cr141.com/openapi.json (4 paths, 19 schemas, captured examples)

- GET /evidence/status
  Parameters: none.
  Returns EvidenceStatus: audit, audit_closed, events_measured, gates, exactness,
  reproducible_package, cache, generated_utc.

- GET /proposal/{pid}
  Parameters: pid — string, path, required; must be an integer in decimal, e.g. 301.
  Returns Proposal: pid, block, block_evaluated, markets, population, targets,
  predicted_flips, observed_flips, exact_set_equality, density_pct, route_a_b,
  rpc_failures, completion_rate.

- GET /proposal/{pid}/flips
  Parameters: pid — string, path, required, must be an integer; attribution — boolean, query, optional;
  limit — integer, query, optional; cursor — string, query, optional.
  Returns FlipPage: pid, total, next_cursor, attribution, flips[] (account, buffer,
  predicted, observed, driver_market, contribution).

- POST /simulate
  Body: {"changes": [{"pid": integer, "market": string, "cf_after": number}]}
  Returns SimulateResult: pid, block, block_evaluated, changes[], affected_accounts,
  potential_flips, newly_flipped, healed, collateral_impact_wei, borrow_impact_wei,
  flips[], flips_truncated, invalid, model.

Errors: 404 when a pid is not in the audited corpus; 422 for a validation failure. Every
success response is a declared Pydantic model, so the schema is the contract.

## Terminology

Governance Impact Analysis; Governance Impact Verification; Collateral Factor
Changes; Proposal Impact Analysis; Counterfactual Governance Analysis; DeFi Risk
Modeling; Account State Transition. These terms are used without variation.

## Limits

Compound only. Executed proposals only. Fixed block state. No price path, no
liquidation sequence, no forecast. What-if scenarios on the dashboard are frozen,
pre-computed counterfactuals — not live simulations.

## Canonical definition

CR141 verifies the account-level consequences of DeFi governance parameter changes
before execution and has been validated on 28 historical Compound governance
proposals with 474 predicted and 474 observed account state transitions.