For a correctly formatted copy, use the PDF button on the study page.
For a correctly formatted copy, use the PDF button on the study page.
AI-written text · figures from the filings or computed in code
| Measure | Share | What it means |
|---|---|---|
| Extraction coverage | 86% | the share of the figures the analysis looks for that were found in your documents |
| Analysis computed | 89% | the share of the indicators computed from your data |
Overall confidence: high (index 0.88).
Study quality: Substantial, the lowest of the three readings below.
| Reading | Measured | Grade |
|---|---|---|
| Extraction coverage | 86% of the figures the analysis looks for | Complete |
| Analyses fully computed | 11 of 22 analyses | Substantial |
| Score quality | highly robust | Complete |
Reporting quality: Strong.
Study quality says how much of the analysis the data allowed; reporting quality describes the filings. Neither is a judgement of the company.
| Observation | Read from |
|---|---|
| The audit opinion is unmodified. | Auditor Opinion |
| Audit fees moved +7% against the prior year. | Auditor Fees |
| The filings state the accounting framework they were prepared under. | Accounting Standards |
| No departure from the expected digit distribution is detectable in the 76 figures tested. | Digit Distribution |
| 4 published figures could not be checked against a figure that was not confirmed; every other published figure passes the statement identities it takes part in. | Statement identities |
| 2% of the published figures were not confirmed by a second reading. | Re-read figures |
482 figures were published from your documents.
| Reading confidence | Figures | Share |
|---|---|---|
| high | 403 | 84% |
| low | 56 | 12% |
| medium | 23 | 5% |
| Line item | Where it is printed | Analyses blocked | Impact |
|---|---|---|---|
| Net Fixed Assets | Balance Sheet - Non-Current Assets | 2 | medium |
| Preferred Equity | Balance Sheet - Equity Section | 2 | medium |
| Capitalized Interest | Notes to Financial Statements | 1 | low |
| Acquisitions and Divestitures | Cash Flow Statement - Investing Section | 1 | low |
| Closely Held Shares | Proxy Statements | 1 | low |
| Investment Property | Balance Sheet - Non-Current Assets | 0 | low |
| Abnormal Items | Income Statement or Notes | 1 | low |
| Sale of Business | Income Statement | 0 | low |
| Name Change Count | Corporate Filings | 1 | low |
| Unrealized Gains/Losses | Income Statement or Notes | 1 | low |
| Analysis | Indicators computed | State | Why not complete |
|---|---|---|---|
| Bankruptcy Risk | 6 of 6 | healthy | — |
| Assets | 4 of 4 | healthy | — |
| Digit Distribution | 4 of 4 | healthy | — |
| Business Model | 0 of 2 | severely degraded | needs 14 consecutive years of figures; Lynx has 4 in a row |
| Cash | 8 of 9 | healthy | depends on another indicator that could not be computed |
| Accrual Quality | 0 of 1 | severely degraded | needs 7 consecutive years of figures; Lynx has 4 in a row |
| Working Capital Accrual Quality | 0 of 1 | severely degraded | needs 9 consecutive years of figures; Lynx has 4 in a row |
| Credit | 12 of 12 | healthy | — |
| Governance | 9 of 11 | healthy | missing Closely Held Shares, Name Change Count |
| Growth | 5 of 7 | degraded | needs 5 consecutive years of figures; Lynx has 4 in a row |
| Real Activity Cost Management | 0 of 0 | severely degraded | it is estimated from the activity of listed companies in the same industry and year, and that comparison is not available yet |
| Income | 11 of 15 | degraded | missing Abnormal Items, Capitalized Interest, Unrealized Gains/Losses |
| Investing | 2 of 3 | degraded | missing Acquisitions and Divestitures |
| Discretionary Accruals | 0 of 3 | severely degraded | needs 7 consecutive years of figures; Lynx has 4 in a row |
| Margins | 6 of 8 | degraded | missing Net Fixed Assets; depends on another indicator that could not be computed |
| Miscellaneous | 8 of 8 | healthy | the figures were read, and a comparison with other companies on the same measure is not available yet; the figure it needs is not one the analysis reads from the documents |
| Manipulation Detection | 6 of 9 | degraded | missing Net Fixed Assets; depends on another indicator that could not be computed; needs the same figures for an earlier year, which Lynx did not find in the filings uploaded |
| Financial Strength | 11 of 11 | healthy | — |
| Real Activity Manipulation | 0 of 0 | severely degraded | it is estimated from the activity of listed companies in the same industry and year, and that comparison is not available yet |
| Smoothing | 7 of 8 | healthy | needs 5 consecutive years of figures; Lynx has 4 in a row |
| Valuation | 4 of 5 | healthy | missing Preferred Equity |
| Working Capital | 11 of 11 | healthy | — |
| Year | Rating | Score quality | Analysis computed | Carried by |
|---|---|---|---|---|
| 2025 | A | highly robust | 89% | — |
| 2024 | A | highly robust | 85% | — |
| 2023 | A | highly robust | 85% | — |
| 2022 | A (reduced reliability) | fragile | 58% | — |
The study quality grade is Substantial, and what decides it is not how the documents were read but how many analyses they could carry. The extraction coverage reading is graded Complete and the score quality reading is Complete as well, with the indicators that were computed described as highly robust, so it is the analyses-fully-computed reading, graded Substantial, that sets the grade: enough analyses ran to carry a real study, but not all of them. The figures that were read were read confidently. Nearly all figures were read with better than low confidence, every one matched its source in the documents exactly, and the reconciliations behind the figures were all checked; of the figures that were re-read, most agreed, a small number were corrected, and the rest were not re-read. The analyses that could not be computed fall into groups. Several need longer unbroken runs of years than the filings supply, which holds back the Business Model, Accrual Quality, Working Capital Accrual Quality, Discretionary Accruals and, in part, the Growth and Smoothing analyses. The rest are short of particular line items: Net Fixed Assets is the most consequential, holding back the Margins and Manipulation Detection categories, and Preferred Equity holds back part of Valuation; lower-impact gaps include capitalized interest, acquisitions and divestitures, closely held shares, abnormal items and unrealized gains and losses. The confidence band is high, and nothing forced it lower.
No upload is recommended, because the assessment lists none: the analyses still missing are held back by the length of the filing history and by a comparison with other companies that is not yet available, rather than by figures a further document would supply.
This study grades Substantial for study quality and Strong for reporting quality, and together they describe a study you can rely on within clear limits. Study quality says how much of the planned analysis your filings allowed; reporting quality describes the condition of the filings themselves, not a judgement of how the business is run. The strongest reading behind the study grade is extraction coverage, which measures the share of the figures the analysis looks for that were actually found in the documents you uploaded. Here that reading is graded Complete, meaning nearly all the figures were found, and what was found was read with confidence: nearly all figures were read with better than low confidence, each matched its source in the documents exactly, and the figures were checked against the totals they belong to. The reporting grade rests on facts such as an unmodified audit opinion, accounting standards being stated, audit fees that moved modestly against the prior year, and figures that pass the arithmetic checks connecting them. The rating is for 2025, and every year in the study carries one, back to 2022; the earliest year's rating carries a caution that its reliability is reduced, because fewer figures could be read for that year. What holds the study grade down is not how the documents were read but how many analyses they could carry. The assessment lists no upload, so there is no single document to send that would lift the grade; the remaining gaps are ones a new upload would not close.