Formula and interface behavior
The library supports arithmetic reconciliation, variance sensitivity, filtering, demonstrations, and product QA across varied cost stacks.
Auditable evidence •
Don’t take our word for it. Inspect the 500-row evidence ledger.
Every headline number opens into the records behind it. Review 500 evidence rows, 2,000 audited synthetic scenario packets, 48,000 automated assertions, and 30 queued validation candidates in the separate dealer-evidence lane.
Redacted public manifestFormula checkedHuman decision required
The proof, without the spin
500Inspectable evidence rows
150Full six-output formula audits
350Additional screening traces
Not claimedAcquisition-ready records
All 500 rows passed the arithmetic check appropriate to their stated audit depth. That validates calculation behavior, not title, condition, repair outcome, or profit. The June 29 source snapshot contains 443 unique normalized source addresses and 57 additional rows sharing an address; availability has not been reverified.
A third, clearly separated layer
2,000 outcome simulations stress-test the estimate-to-closeout math.
Each synthetic packet includes identity, simulated date, repair scope, parts, labor, AvLu estimate, simulated closeout, variance, and formula impact. Simulation is not validation.
2,000Complete synthetic outcome packets
6Designed cost-movement patterns
9 of 9Simulation fields populated
UnderwayDealer validation status
Real-world predictive accuracy
No synthetic record is a dealer quote, mechanic invoice, repair order, publication-ready case, or proof of a successful purchase.
Deterministic and inspectable
Every record identifies AvLu’s scenario generator as its source and uses a synthetic vehicle ID instead of a plausible VIN.
Replace assumptions with permissioned proof
The dealer campaign remains the path to attributable estimates, final repair documents, real variance, and formula impact.
Automated regression QA v4.0.0
48,000 automated assertions across 2,000 scenario packets.
Each synthetic packet runs through 24 named checks covering identity boundaries, source classification, cost integrity, estimate math, closeout math, variance, formula impact, protected maximum, and dealer-outcome guardrails.
2,000Scenario packets audited
48,000Automated assertions passed
2424 checks per packet
100%Regression pass rate
Five reconciliation lanes
Every packet checks estimate, closeout, variance, formula impact, and revised protected maximum.
Dealer flags remain closed
Every packet must remain synthetic, non-VIN, not dealer verified, and ineligible as a dealer outcome.
One result per packet
The public report contains the assertion catalog, source-dataset fingerprint, six-pattern coverage, and 2,000 record-level results.
Automated regression is not dealer validation
The pass rate validates deterministic software and formula behavior, not predictive accuracy, repair outcomes, or realized profit.
Two evidence layers
Each number has a boundary and a next proof requirement.
The ledger keeps broad screening evidence separate from dealer-backed outcome evidence.
Arithmetic audited to stated scope500-row cohort
Formula and screening behavior
- 150 rows received the full six-output formula audit.
- 350 additional rows received recon and ask-versus-ceiling screening traces.
- 168 rows contain valid-format VINs; this is not independent identity verification.
- 443 normalized source addresses are unique; 57 additional rows share an address.
- No rows are labeled acquisition-ready because hard proof gates remain incomplete.
Does not prove: dealer repair cost, title status, physical condition, purchase result, or realized profit.
Queued validation sprint30 candidates
Dealer-backed repair evidence
- 10 original priority vehicles are preserved.
- 20 additional valid-format-VIN dealer/public-web candidates expand the queue.
- 30 external evidence packages are still missing.
- Pending describes written-quote capture in the tracked queue.
- In progress describes evidence-complete cases; publication readiness remains awaiting proof.
Completion requires: dated repair evidence, scope, parts and labor, repair-source identity, AvLu estimate, actual estimate, variance, and formula impact.
Redacted row-level evidence
The headline count is only useful if you can inspect the rows.
Filter the public manifest by audit depth, VIN-format status, or risk. Raw source addresses, VINs, dealer identities, and contact details are intentionally omitted. The remaining vehicle details and source fingerprints are pseudonymous, not guaranteed anonymous.
Loading the redacted evidence manifest…
| Evidence ID | Vehicle | Planning math | Audit depth | Source type | VIN format | Risk | Formula | Next proof action |
|---|
Quote-to-case bridge
Thirty candidates are queued. None is a completed dealer case.
Every candidate currently has the identity baseline and original AvLu estimate. The remaining document, source, variance, and consent gates are still open.
30Queued candidates
PendingWritten quotes captured
In progressEvidence-complete cases
Awaiting proofPublication-ready cases
Loading the redacted validation queue…
| Campaign | Vehicle | Selection group | Baseline math | Completed gates | Written quote | Consent | Evidence status |
|---|
Case-study gate
A lead becomes a case study only after nine checks.
Consent is a separate publication gate. Silence is never treated as permission.
- 1VIN and vehicle identity
- 2Dated repair evidence
- 3Repair scope
- 4Parts and labor
- 5Repair-source identity
- 6AvLu estimate
- 7Actual estimate
- 8Variance
- 9Formula impact on the protected maximum
What the next evidence will show
The useful test is whether a written quote changes the walk-away number.
Each completed case will compare the original planning allowance against attributable repair evidence and show the exact effect on the protected maximum offer and ask gap.
Confirm the vehicle
Recheck availability, identity, VIN, title claim, mileage, and the current listing source.
Collect written scope
Request a dated inspection, estimate, repair order, invoice, or identifiable written dealer summary.
Recalculate the deal
Replace the planning allowance with the documented amount and calculate variance and formula impact.
Publish only with consent
Use named, anonymized, or internal-only treatment according to the dealer’s documented choice.
Help turn one open lead into a documented case.
A redacted inspection, estimate, repair order, invoice, or attributable written repair summary is enough to start the comparison.