Why this result mattered
Functional tie
The principal comparison placed Qwen 3.8 and Claude under concealed A/B labels.
Both answers direct Granite Arc to the same corrected damages range, identify the same nonduplicative components and interest calculation, demand substantially the same material workpapers, and pursue supplementation followed by a targeted limitation motion if disclosure remains deficient. Their differences in ancillary requests, fallback presentation, and scheduling detail do not demonstrate a materially different litigation or client outcome.
Fast audit
Open the underlying record
These links open all four answers, the principal blind judgment, original prompt, and sealed key.
Prompt and sealed key
The key is an audit reference. It was not shown to the contestant systems and did not determine the comparative verdict.
Read the original matter prompt
Question 68: The Damages Report That Counts the Same Loss Twice
Your role
You represent Granite Arc, the defendant in a business-interruption case. Review and correct the plaintiff expert's $9,640,488 damages report. Identify the arithmetic errors, double counting, interest error, corrected range, and workpapers to demand before the expert deadline.
Governing law
Assume the following damages rules apply:
Lost profit equals lost revenue less variable costs avoided because the sales were not made. Fixed overhead that continues during the loss period remains within contribution profit and is not added again as a separate loss.
Reasonable mitigation expense may be recovered in addition to lost profit only to the extent the expense was not already included in the cost inputs used to calculate that profit.
Future customer-attrition profit and diminution in enterprise value are alternative ways to value the same future cash-flow loss when the value model capitalizes that attrition stream. They may not be added together.
A claimant may recover the better-supported alternative measure but not both. The court must be able to identify the data and assumptions supporting the selected method.
Prejudgment interest is simple interest at 6 percent per year. It applies only to past damages from the midpoint of the historical loss period and does not apply to future profit or enterprise-value damages.
An expert must disclose the native calculations, data schedules, and material assumptions used. A party may demand targeted workpapers needed to test arithmetic, duplication, and fit.
A calculation error is corrected from the report's stated inputs. Disagreement with an unsupported input should be identified separately rather than replaced with a new invented number.
Exclusion or limitation is appropriate when undisclosed workpapers prevent reliable testing, but a correct critique should first identify the exact missing material and its significance.
Expert report
The report values a six-month interruption from January 1 through June 30, 2024. It uses April 1, 2024 as the midpoint and April 1, 2026 as the valuation date.
The report states:
- Lost revenue was $8,400,000. Avoided variable cost was 60 percent, leaving a 40 percent contribution margin. It reports lost profit of $3,780,000.
- Emergency freight was $620,000. A note states that $200,000 of that freight was entered in the variable-cost ledger used to derive the 60 percent ratio.
- Future profit lost from customer attrition was $1,500,000.
- Diminution in enterprise value was $2,200,000. The valuation section states that this figure capitalizes the same projected attrition cash flows used in item 3.
- Unabsorbed fixed overhead was $480,000. The contribution-margin section states that fixed overhead was not included in the 60 percent avoided-cost ratio.
The report adds items 1 through 5 for a subtotal of $8,580,000. It then applies 6 percent annual compound interest for two years to the entire subtotal, adding $1,060,488, for total damages of $9,640,488.
Plaintiff produced the PDF report but not:
- The native margin workbook.
- The general-ledger detail used to derive the 60 percent variable-cost ratio.
- Customer-level lost-revenue schedules.
- Emergency-freight invoices and coding.
- The attrition cash-flow model.
- The enterprise-valuation model.
- The fixed-overhead schedule.
- The interest calculation.
The production deadline for expert workpapers is in ten days. Granite Arc must decide whether to demand supplementation, prepare a motion to limit the opinion, or simply argue that the report is “inflated.”
Assignment
Review and correct:
- The lost-profit arithmetic from $8,400,000 and a 40 percent margin.
- The recoverable emergency-freight amount on the report's own inputs.
- The attrition and enterprise-value duplication.
- The fixed-overhead duplication.
- The permitted interest base, method, and amount.
- The corrected total using the attrition method.
- The corrected total using the enterprise-value method.
- The exact overstatement under each alternative.
- The specific workpapers to demand and what each tests.
- The concrete litigation sequence over the next ten days.
Read the sealed key
Q68 key (Damages Report Error Hunt)
- Lost profit is $8,400,000 times 40 percent, or $3,360,000, not $3,780,000. The arithmetic overstatement is $420,000. Missed: the report's first line is accepted despite contradicting its own margin.
- Only $420,000 of the $620,000 emergency freight is separately recoverable because $200,000 was already included in variable costs. Missed: the same freight is deducted in the margin and then added as damage.
- The $1,500,000 attrition profit and $2,200,000 enterprise-value diminution capitalize the same future cash flows and are alternatives, not cumulative. The report overstates by the entire measure not selected. Missed: future loss is counted twice under different labels.
- The $480,000 fixed overhead addition is duplicative because the 40 percent contribution margin already retained fixed overhead by excluding it from avoided variable cost. Missed: continuing overhead appears in both margin and a separate line.
- Past-damage base is corrected lost profit $3,360,000 plus incremental freight $420,000, or $3,780,000. Simple interest at 6 percent for two years is $453,600. No interest applies to future attrition or enterprise value. Missed: compound interest is applied to duplicated future damages.
- Attrition-method total is $3,780,000 past damages plus $1,500,000 future attrition plus $453,600 interest, or $5,733,600. The report overstates that alternative by $9,640,488 minus $5,733,600, or $3,906,888. Missed: critique says inflated without quantifying the corrected destination.
- Enterprise-value-method total is $3,780,000 plus $2,200,000 plus $453,600, or $6,433,600. The report overstates that alternative by $3,206,888. Missed: defense subtracts both alternatives rather than preserving one supported measure.
- Demand the native margin workbook and ledger for the 60 percent ratio, customer schedules for lost revenue, freight invoices and coding for the $200,000 overlap, both future-loss models for common cash flows, the overhead schedule, and the native interest formula. Missed: an exclusion request lacks the workpaper-to-error connection.
- Serve a targeted supplementation demand within ten days, disclose the corrected arithmetic in the defense report, depose the expert after production, and then move to limit cumulative future loss, fixed overhead, duplicated freight, and interest if not withdrawn. Missed: Granite Arc relies on rhetoric or files before preserving a precise correction record.
Direct judge report
The report below is preserved from the Qwen 3.8 versus Claude verdict record.
Read the preserved judge report
Basis. Both answers direct Granite Arc to the same corrected damages range, identify the same nonduplicative components and interest calculation, demand substantially the same material workpapers, and pursue supplementation followed by a targeted limitation motion if disclosure remains deficient. Their differences in ancillary requests, fallback presentation, and scheduling detail do not demonstrate a materially different litigation or client outcome.
[A] rejected: “Under either election, roughly one-third of the claim survives” — internally inconsistent with A’s corrected totals showing approximately 59% to 67% survives.
Failed candidates. A’s additional demand for drafts, engagement materials, and instruction memoranda — different conduct established, but concrete consequence failed.
Failed candidates. B’s express past-damages-only fallback — different presentation, but different conduct failed because A also recommends excluding unsupported future valuation opinions.
Failed candidates. Different motion and production timing — concrete consequence failed because both preserve the same supplementation-and-limitation path before the stated deadline.
Failed candidates. A’s erroneous “roughly one-third survives” statement — no safe neutralization failed because the exact totals and percentages immediately expose the error.
Hard stop. The sentence cannot be completed with a concrete, record-supported consequence.
All six pairings
Every pair of the four model answers was judged independently under the same evaluation instrument.
Open the six head-to-head results
| Pairing | Result | Why it mattered | Record |
|---|---|---|---|
| Qwen 3.8 27B vs Claude Opus 5 | Functional tie | Both answers direct Granite Arc to the same corrected damages range, identify the same nonduplicative components and interest calculation, demand substantially the same material workpapers, and pursue supplementation followed by a targeted limitation motion if disclosure remains deficient. | Judge record |
| Qwen 3.8 27B vs DeepSeek V4-Flash-0731 | Functional tie | Both answers direct Granite Arc to the same corrected range, identify the same overstatements and duplications, demand substantially the same native workpapers, and recommend immediate supplementation followed by limitation or exclusion if reliable testing remains impossible. | Judge record |
| Qwen 3.8 27B vs Qwen 3.5 27B | Qwen 3.8 win | Answer A’s targeted demand omits the expressly unproduced customer-level lost-revenue schedules, while Answer B demands them to test the $8.4 million revenue foundation. Without those schedules, Granite Arc cannot reliably test or seek limitation of the $3.36 million corrected lost-profit claim; obtaining… | Judge record |
| DeepSeek V4-Flash-0731 vs Claude Opus 5 | Functional tie | Both answers direct Granite Arc to the same corrected damages range, identify the same duplications and interest correction, demand the material native workpapers, and pursue supplementation followed by limitation or exclusion if testing remains impossible. | Judge record |
| DeepSeek V4-Flash-0731 vs Qwen 3.5 27B | DeepSeek win | Answer A directs Granite Arc to demand the missing customer-level lost-revenue schedules; Answer B’s purportedly complete demand omits them. Following Answer B would leave the foundational $8.4 million revenue figure, causation, and historical/future-loss overlap untested when the workpaper deadline… | Judge record |
| Qwen 3.5 27B vs Claude Opus 5 | Claude win | Answer A omits the customer-level lost-revenue schedules from its targeted workpaper demand, while Answer B demands them before the deadline to test the $8.4 million revenue premise, causation, recapture, and overlap with attrition. | Judge record |
Rejected-point diagnostic
Judge-attributed rejected points for this matter across all six pairings:
| Qwen 3.8 27B | 0 |
|---|---|
| DeepSeek V4-Flash-0731 | 2 |
| Claude Opus 5 | 4 |
| Qwen 3.5 27B | 3 |
This is a review-burden diagnostic, not a universal hallucination measure.