Reconstructing a historical fundamental screen requires more than selecting an earlier date and retrieving financial statements. A database queried today may contain filings that were accepted after the historical decision date. Using those filings can introduce information that falls outside the intended screening cutoff.
A fiscal year's end date does not establish when its annual filing became eligible for a historical screen. The workflow must connect each financial value to its supporting filing and compare that filing's acceptance timestamp with the decision cutoff.
This article uses Claude with Financial Modeling Prep's MCP server to reconstruct an annual revenue-growth screen for Apple (AAPL) and Microsoft (MSFT) as of October 15, 2024. It retrieves annual as-reported financial data, verifies filing provenance, resolves consolidated revenue values, and applies the screening rule only after determining which filings qualify.
The result is a filing-based historical reconstruction using records retrieved at a later observation date. It does not reproduce FMP's historical database or establish the first moment every financial figure became public. Its purpose is to control which filing-supported values enter the calculation, with unresolved evidence marked Review Required.
FMP Data and Claude MCP Setup
The workflow uses the FMP MCP server to retrieve financial statements, SEC filing records, and historical prices within Claude.
To reproduce the analysis, follow the MCP setup documentation, connect and authorize the server, and confirm access to the required datasets before running the analytical prompt.
|
FMP dataset |
MCP operation recorded in this analysis |
Role in the workflow |
|
as-reported-income-statements |
Retrieve annual financial values and identify the consolidated revenue or net sales figure. |
|
|
search-by-symbol |
Identify each annual filing, accession number, filing date, acceptance timestamp, and source URL. |
|
|
historical-price-eod-light |
Retrieve the decision-date closing price for context. |
|
|
financial-reports-form-10-k-json |
Resolve ambiguous revenue values against the corresponding fiscal year's consolidated financial statement. |
The operation names above are the MCP labels recorded in this analysis. They should not be treated as literal REST URL paths; the linked documentation identifies the corresponding API endpoints.
The 10-K JSON report is used when the flattened as-reported response does not establish the consolidated top line reliably. An XBRL revenue concept can appear in several contexts, including consolidated results, products, services, segments, and other dimensional disclosures. Selecting a familiar field name alone is therefore insufficient.
For example, a product-sales subtotal cannot substitute for consolidated net sales, and Microsoft Cloud revenue cannot substitute for Microsoft's total revenue. The workflow checks the statement structure and retains the company-level figure from the same fiscal-year filing used for the availability test.
This is the same source-selection issue explored in detecting financial statement mapping risk with Claude MCP. Here, that check is combined with filing acceptance timestamps so the selected value has both financial and chronological support.
Point-in-Time Screening Methodology
The historical decision cutoff is October 15, 2024, at 4:00 p.m. Eastern time, equivalent to 16:00 EDT or 20:00 UTC.
The analysis uses annual data for FY2022 through FY2024. For each company, it connects the consolidated annual revenue value to the corresponding 10-K and checks whether that filing was accepted by the cutoff.
Three dates must remain distinct:
- Fiscal-period end: the end of the period covered by the financial statements.
- Filing date: the official filing date assigned to the submission.
- Acceptance timestamp: the date and time EDGAR accepted the submission.
Keeping these dates separate is also central to aligning earnings dates, fiscal quarters, estimates, and reported results.
What “Available” Means in This Screen
“Available” is an operational classification based on filing acceptance and verified financial provenance.
|
Classification |
Requirement |
Treatment |
|
Available |
The filing was accepted by the cutoff, the consolidated value is established reliably, and the value and acceptance timestamp refer to the same filing. |
Eligible for selection into the screening pair. |
|
Later Filed |
The supporting filing was accepted after the cutoff. |
Excluded from the calculation; the evidence row is marked Review Required. |
|
Unavailable / Review Required |
The filing, timestamp, financial value, or relationship between them cannot be established reliably. |
No estimation or substitution; the unresolved evidence is identified. |
SEC acceptance is not an exact public-dissemination timestamp. The SEC's explanation of EDGAR timestamps and publication timing states that filings often appear on sec.gov within one to three minutes of acceptance, but that delay is not guaranteed and can increase. This workflow therefore establishes acceptance-based eligibility rather than precise intraday access.
None of the six filings in this example was accepted close to the October 15, 2024 cutoff. The timing clarification does not change the selected periods or results.
Selecting the Annual Comparison
For each company, the screen selects the latest annual period classified as Available and its immediately preceding eligible annual period.
A later-filed period does not automatically remove the company from the screen. The workflow excludes that period and steps back to an earlier valid comparison.
Review Required can therefore apply to an individual evidence row without applying to the final company result. If two eligible annual periods remain fully supported, the screen can still return PASS or FAIL.
The calculation is:
Revenue Growth (%) = [(Latest Eligible Annual Revenue / Previous Eligible Annual Revenue) − 1] × 100
The screening rule is:
- PASS: Revenue Growth > 0%
- FAIL: Revenue Growth ≤ 0%
- REVIEW REQUIRED: A reliable two-period calculation cannot be completed.
Calculations use the exact displayed consolidated values, with only the final percentage rounded to two decimal places.
Historical closing prices for October 15, 2024 are reported separately as context. They do not affect the screening result, and the analysis does not calculate subsequent returns.
Claude Prompt for Filing-Date-Aware Analysis
The following prompt defines the original workflow, including filing eligibility, same-year financial reconciliation, and the screening rule. In this example, its reference to the U.S. regular-market close means the cutoff specified above.
mensional revenue value instead of the consolidated statement-level figure.
|
Use the connected Financial Modeling Prep (FMP) MCP server to reconstruct a filing-date-aware fundamental screen for:
Historical decision date:
This is a filing-based historical reconstruction using FMP records retrieved in the current Claude session. It is not a reconstruction of the historical state of FMP's database on October 15, 2024. Use FMP MCP only. Do not use web search, external sources, analyst estimates, quarterly data, TTM data, or standardized financial statements to substitute for missing evidence. Keep the workflow compact and limited to the evidence required below. 1. Record the Retrieval Observation Time At the beginning of the analysis, record:
This timestamp describes when the FMP responses were retrieved in the current analysis. Do not imply that FMP's database looked the same on October 15, 2024. 2. Retrieve Annual Financial Evidence Retrieve annual as-reported income-statement data for FY2022 through FY2024 for AAPL and MSFT. For each fiscal year identify:
The required value is the consolidated company-level top-line figure from the annual income statement. Do not use:
If the flattened as-reported response clearly provides the consolidated statement-level value, retain it directly. If the flattened response is ambiguous or returns a dimensional subtotal, use the FMP MCP 10-K JSON financial-report tool for that same fiscal year to identify the consolidated statement-level value. When 10-K JSON reconciliation is required:
Where useful, reconcile the selected value against related annual statement lines such as cost of sales / cost of revenue and gross profit / gross margin. If the MCP output does not provide enough context to establish confidently that the selected value is the consolidated statement-level figure from that filing, mark that fiscal period: Review Required Do not estimate or substitute the value. 3. Retrieve Filing Provenance Retrieve the SEC annual filing records required to support FY2022 through FY2024 for both companies. Focus on:
For every fiscal-year evidence row, retain:
The accession number and filing URL must identify the same filing whose acceptance timestamp is used for the historical availability test. For every financial value, confirm explicitly: Financial value and acceptance timestamp refer to the same filing: Yes / No / Review Required If the value was reconciled using 10-K JSON, confirm that the JSON report corresponds to the same fiscal-year filing whose accession number, URL, and acceptance timestamp are shown. If that relationship cannot be established reliably, mark the period Review Required. Do not pair a value sourced from one filing with an acceptance timestamp from another filing. Retrieve the relevant 10-K / 10-K/A records needed for FY2022 through FY2024. Do not perform an exhaustive forensic filing audit. However, if the MCP retrieval does not provide enough coverage to determine whether the relevant annual filing set has been captured, state: SEC filing coverage: Review Required rather than claiming that no additional 10-K or 10-K/A exists. Retain the actual MCP tool name used for the SEC filing retrieval so that the retrieval sequence can be audited separately in a follow-up. 4. Determine Historical Availability Use the SEC filing acceptance timestamp, not the fiscal-period end date and not merely the filing date, to determine historical availability. Classify each fiscal-year evidence row as: Available Use only when:
Later Filed Use when the filing that supports the financial evidence was accepted after the historical cutoff. For Later Filed evidence:
Unavailable / Review Required Use when any required element cannot be established reliably, including:
A Review Required evidence row does not automatically force the whole company result to Review Required if an earlier valid two-period comparison can still be established. 5. Retrieve Historical Price Retrieve the historical regular closing price for:
The historical price is contextual only. Do not use it in the fundamental screen and do not calculate subsequent returns. 6. Construct the Fundamental Screen After completing the availability assessment:
Calculate: Revenue Growth (%) = Calculate from the exact displayed values. Round only the final percentage to two decimal places. Apply:
A later-filed fiscal period does not automatically remove the company from the screen. Exclude that period and step back to the latest annual period that was genuinely available. 7. Output First state: Retrieval Observation Time: [date, time, timezone if available] Then state: Reconstruction Basis: Filing-based point-in-time reconstruction performed using FMP records retrieved at the observation time above. This analysis does not claim to reproduce the historical state of FMP's database on October 15, 2024. Table 1 — Filing Availability and Source Provenance | Company | Fiscal Period | Form | Accession Number | Filing Acceptance Timestamp | Filing URL | Financial Metric / Exact Reported Label | Consolidated Value | FMP Financial Source Used | Source Filing Fiscal Year | Value + Acceptance From Same Filing? | Availability Status | Included / Excluded | Review Required | Evidence Note | Requirements:
Use Included / Excluded labels only after determining the final eligible screening pair:
Table 2 — As-of-Date Fundamental Screen | Company | Latest Eligible Fiscal Period | Prior Eligible Fiscal Period | Latest Eligible Revenue / Net Sales | Prior Eligible Revenue / Net Sales | YoY Revenue Growth | Historical Price on Decision Date | Screen Result | Review Required | Analyst Follow-Up Action | Set Analyst Follow-Up Action = None when the historical screen is fully resolved. Only provide a follow-up action when an unresolved evidence problem prevents a reliable screen or leaves material provenance incomplete. 8. Short Interpretation After the tables, provide no more than four concise bullets explaining:
9. Retrieval Note for Follow-Up Audit End with a short note listing only:
Do not provide a detailed page-by-page or parameter-by-parameter tool trace in this run. That retrieval trace will be audited separately if needed. Final Checks Before responding, verify:
Do not add companies, metrics, forward returns, investment recommendations, or post-decision performance analysis. |
Targeted MCP Retrieval and SEC Coverage Audit
The main prompt produces the financial screen and records the source filings. A separate follow-up checks the retrieval process supporting the filing-coverage statements.
Run the following prompt in the same Claude conversation so it can reference the preceding tool-call history. It focuses on the actual calls, search windows, result counts, limits, and annual-filing inventory. It permits additional SEC searches when necessary to check the defined coverage period, while leaving the financial analysis unchanged.
|
Using the immediately preceding as-of-date fundamental-screen analysis, perform only a targeted retrieval-methodology audit. Do not rerun the financial analysis, revenue reconciliation, historical-price retrieval, growth calculations, or PASS/FAIL screen. The purpose of this follow-up is to document the actual FMP MCP call trace used in the preceding run and determine whether the SEC filing retrieval was sufficient to support the filing-coverage statements. 1. Report the Actual MCP Call Trace Use the tool-call history from the immediately preceding run. For every FMP MCP call actually used, report: ● call number; ● exact MCP tool name; ● endpoint or operation selected; ● company / symbol; ● exact parameters supplied; ● start date and end date, where applicable; ● fiscal year / period, where applicable; ● page value, if supplied; ● limit value, if supplied; ● number of records returned, if available; ● purpose of the call. Do not reconstruct or guess parameters from memory. If a parameter, page value, limit, or result count is not visible in the retained tool-call history, state: Not available from retained tool trace Do not invent it. Return: Table A — Actual FMP MCP Call Trace | Call # | MCP Tool | Endpoint / Operation | Symbol | Parameters | Date Range | Page | Limit | Result Count | Purpose | Include calls used for: ● as-reported-income-statements; ● financial-reports-form-10-k-json; ● search-by-symbol; ● historical-price-eod-light. 2. Document the SEC Filing Retrieval Sequence Using only the actual secFilings / search-by-symbol calls from the preceding run, list the SEC filing searches in execution order. For each search show: ● company; ● start date; ● end date; ● page; ● limit; ● result count; ● relevant 10-K or 10-K/A records returned; ● whether the date range overlapped another search window. If multiple windows were used, explain briefly why. If duplicate filing records were returned across overlapping windows, state: ● which accession numbers were duplicated; ● how duplicates were removed. If there were no duplicates, state that explicitly only if the retrieved results establish it. 3. Confirm 10-K / 10-K/A Coverage Determine whether those searches are sufficient to establish that no relevant 10-K or 10-K/A supporting FY2022-FY2024 was omitted. Do not assume that targeted windows prove complete amendment coverage. If the retained SEC searches already establish complete coverage, explain exactly why. If they do not, perform only the minimum additional secFilings retrieval necessary to check annual 10-K / 10-K/A coverage for: ● AAPL ● MSFT across January 1, 2022 through December 31, 2024. Do not retrieve financial statements again. For any additional SEC search, record the exact: ● MCP tool; ● parameters; ● date range; ● page; ● limit; ● result count. If one call cannot provide reliable coverage because of result limits or unrelated filings crowding the response, use the minimum number of narrower date windows needed. Do not perform a broad forensic audit beyond this date range. 4. Filing Inventory Return: Table B — Annual Filing Coverage | Company | Fiscal Year | Form | Accession Number | Filing Acceptance Timestamp | Filing URL | Retrieved In Call # | Coverage Status | Use descriptive hyperlinks for Filing URL. Include every relevant FY2022-FY2024 10-K and any 10-K/A identified by the audited retrieval sequence. Use one of: ● Confirmed in retrieval ● Amendment identified ● Coverage incomplete / Review Required Do not state that no 10-K/A exists unless the audited retrieval sequence supports that conclusion. 5. Coverage Conclusion For each company state only one of: SEC filing coverage confirmed for the defined Jan. 1, 2022-Dec. 31, 2024 scope or SEC filing coverage remains Review Required Then give one sentence explaining why. 6. Retrieval-Time Clarification Retain the retrieval observation time from the preceding analysis. Confirm explicitly: This workflow reconstructs filing-based historical availability using FMP records retrieved on the observation date. It does not establish what FMP's own database would have returned on October 15, 2024. Output Discipline Return only: 1. Table A — Actual FMP MCP Call Trace 2. Table B — Annual Filing Coverage 3. the two company-level coverage conclusions 4. the retrieval-time clarification Do not revise the financial-result tables from the main analysis. Do not recalculate financial values, prices, growth rates, or screen results. Do not introduce external sources or web search. Do not invent missing tool-call metadata. |
Results: Filing Availability and the Historical Screen
Recorded retrieval observation time: September 27, 2026, at 04:20:32 UTC.
Reconstruction basis: The documented run used FMP records retrieved at that observation time to reconstruct filing-based eligibility for the October 15, 2024 cutoff. It does not establish what FMP's database would have returned on the historical decision date.
The filing assessment produced different eligible comparisons for the two companies. Apple's FY2024 had ended before the cutoff, but its 10-K was accepted on November 1, 2024. Microsoft's FY2024 10-K had already been accepted on July 30, 2024.
The evidence is presented below in separate filing, financial, and eligibility tables for readability.
Table 1A: Annual Filing Identity and Acceptance
All six source documents are Forms 10-K. Acceptance timestamps below use Eastern time as displayed in the SEC filing records, not UTC. Each acceptance timestamp links to the corresponding filing index.
|
Source filing |
Fiscal-period end |
Official filing date |
SEC acceptance timestamp, Eastern time |
Accession number |
|
September 24, 2022 |
October 28, 2022 |
0000320193-22-000108 |
||
|
September 30, 2023 |
November 3, 2023 |
0000320193-23-000106 |
||
|
September 28, 2024 |
November 1, 2024 |
0000320193-24-000123 |
||
|
June 30, 2022 |
July 28, 2022 |
0001564590-22-026876 |
||
|
June 30, 2023 |
July 27, 2023 |
0000950170-23-035122 |
||
|
June 30, 2024 |
July 30, 2024 |
0000950170-24-087843 |
Apple's FY2022 and FY2023 records also illustrate why filing date and acceptance date should be retained separately: the official filing dates differ from the acceptance dates.
Table 1B: Consolidated Financial Evidence
All values are in USD millions. Apple's selected statement line is Total net sales in its Consolidated Statements of Operations. Microsoft's selected line is Total revenue in its Income Statements.
“As-reported” below refers to as-reported-income-statements. “Same-year 10-K JSON” refers to financial-reports-form-10-k-json.
|
Company and period |
Consolidated statement label |
Value, USD millions |
Financial source used |
Value and acceptance tied to the same filing? |
|
AAPL FY2022 |
Total net sales |
394,328 |
Same-year 10-K JSON |
Yes |
|
AAPL FY2023 |
Total net sales |
383,285 |
As-reported |
Yes |
|
AAPL FY2024 |
Total net sales |
391,035 |
As-reported; excluded from screen |
Yes |
|
MSFT FY2022 |
Total revenue |
198,270 |
Same-year 10-K JSON |
Yes |
|
MSFT FY2023 |
Total revenue |
211,915 |
Same-year 10-K JSON |
Yes |
|
MSFT FY2024 |
Total revenue |
245,122 |
Same-year 10-K JSON |
Yes |
For every row, the source filing's fiscal year matches the financial period shown. The documented analysis matched the value, accession number, filing URL, and acceptance timestamp to that same filing.
The financial reconciliation notes explain how the values were selected:
- Apple FY2022: The documented flattened response surfaced the Products subtotal of $316,199 million. Reconciliation against the FY2022 10-K JSON identified consolidated Total net sales of $394,328 million.
- Apple FY2023 and FY2024: The documented run retained values from the as-reported key revenuefromcontractwithcustomerexcludingassessedtax. This is the returned data key, not the financial statement's displayed label. The corresponding consolidated statement line is Total net sales. FY2023's $383,285 million value reconciled to the reported gross-profit relationship. FY2024's $391,035 million value was retained only to document the excluded evidence.
- Microsoft FY2022 through FY2024: The documented flattened responses surfaced Microsoft Cloud subtotals of $91,200 million, $111,600 million, and $137,400 million, respectively. Same-year 10-K JSON reconciliation identified consolidated Total revenue of $198,270 million, $211,915 million, and $245,122 million.
These retrieval observations describe this run. They do not imply that every future response will expose the same subtotal or require the same reconciliation.
Table 1C: Eligibility and Selection
|
Company and period |
Availability status |
Treatment in final screen |
Evidence-row Review Required |
|
AAPL FY2022 |
Available |
Included (prior) |
No |
|
AAPL FY2023 |
Available |
Included (latest) |
No |
|
AAPL FY2024 |
Later Filed |
Excluded |
Yes |
|
MSFT FY2022 |
Available |
Available, not used in final pair |
No |
|
MSFT FY2023 |
Available |
Included (prior) |
No |
|
MSFT FY2024 |
Available |
Included (latest) |
No |
Apple FY2024 carries the prompt's required evidence-row flag because its supporting filing was accepted after the cutoff. Its exclusion is resolved, and an earlier valid pair remains available, so Apple's final company-level result does not require further review.
Microsoft FY2022 is eligible evidence but is not used because FY2024 and FY2023 form the latest eligible annual comparison.
As-of-Date Screen Results
After establishing the eligible pairs, the workflow applies the predefined revenue-growth rule.
Table 2A: Fundamental Screen
Revenue and net sales values are in USD millions.
|
Company |
Latest eligible period |
Prior eligible period |
Latest revenue / net sales |
Prior revenue / net sales |
YoY growth |
Result |
|
AAPL |
FY2023 |
FY2022 |
383,285 |
394,328 |
−2.80% |
FAIL |
|
MSFT |
FY2024 |
FY2023 |
245,122 |
211,915 |
+15.67% |
PASS |
The calculations are reproducible from those inputs:
AAPL:
[(383,285 / 394,328) − 1] × 100 = −2.80%
MSFT:
[(245,122 / 211,915) − 1] × 100 = +15.67%
Table 2B: Historical Price and Review Status
The closing prices below are the values reported in the documented run for October 15, 2024.
|
Company |
Historical closing price, USD |
Company-level Review Required |
Analyst follow-up action |
|
AAPL |
$233.85 |
No |
None |
|
MSFT |
$418.74 |
No |
None |
Apple's eligible comparison is FY2023 versus FY2022. FY2024 is excluded because its 10-K was accepted after the cutoff.- Microsoft's eligible comparison is FY2024 versus FY2023. Its FY2024 filing was accepted before the cutoff.
- Apple fails the positive-revenue-growth rule at −2.80%; Microsoft passes at +15.67%.
- Both company-level calculations are resolved. The prices provide context only, and PASS or FAIL describes this single screening rule rather than an investment recommendation.
Data Retrieval and Filing Coverage
The retrieval summary below records the tools, parameters, and SEC result counts documented for the September 27, 2026 run.
Recorded MCP Tools and Operations
|
Purpose |
MCP tool |
Operation |
Recorded parameters or scope |
|
Annual as-reported financials |
mcp__FMP__statements |
as-reported-income-statements |
period=annual, limit=6, separately for AAPL and MSFT |
|
Filing-level financial reconciliation |
mcp__FMP__statements |
financial-reports-form-10-k-json |
AAPL FY2022 and MSFT FY2022 through FY2024; period=FY |
|
SEC filing provenance and coverage |
mcp__FMP__secFilings |
search-by-symbol |
Company-specific filing windows, followed by checks covering January 1, 2022 through December 31, 2024 |
|
Historical price |
mcp__FMP__chart |
historical-price-eod-light |
October 14 through October 16, 2024, separately for each company |
The financial reconciliation used each period's own annual filing. No financial value used in the screen was sourced from a later comparative filing.
Recorded SEC Search Windows
Result counts refer to all records returned by each search, not only Forms 10-K.
|
Company |
Search window |
Page |
Limit |
Result count |
Purpose |
|
AAPL |
September 1 through December 31, 2022 |
Not supplied |
Not supplied |
18 |
Locate FY2022 10-K |
|
AAPL |
October 15 through November 15, 2023 |
Not supplied |
Not supplied |
4 |
Locate FY2023 10-K |
|
AAPL |
October 25 through November 15, 2024 |
Not supplied |
Not supplied |
5 |
Locate FY2024 10-K |
|
MSFT |
July 15 through August 15, 2022 |
Not supplied |
Not supplied |
3 |
Locate FY2022 10-K |
|
MSFT |
July 15 through August 15, 2023 |
Not supplied |
Not supplied |
6 |
Locate FY2023 10-K |
|
MSFT |
July 15 through August 15, 2024 |
Not supplied |
Not supplied |
5 |
Locate FY2024 10-K |
|
AAPL |
January 1, 2022 through December 31, 2024 |
0 |
1000 |
255 |
Check annual-filing coverage |
|
MSFT |
January 1, 2022 through December 31, 2024 |
0 |
1000 |
472 |
Check annual-filing coverage |
The targeted windows located the six annual filings used in the analysis. Within each company, those initial windows did not overlap, and no duplicate-removal step was required for those searches.
The narrow windows were insufficient to establish amendment coverage across the full period. The follow-up therefore checked the January 1, 2022 through December 31, 2024 filing inventory using page=0 and limit=1000.
The recorded full-window searches returned 255 records for Apple and 472 for Microsoft, both below the requested limit. The audited results contained three relevant 10-Ks for each company and no 10-K/A filings within that defined scope.
An attempted formType=10-K/A filter did not narrow the Apple response as expected. Amendment coverage was therefore evaluated from the complete full-window filing results.
|
Company |
Recorded coverage conclusion |
Supporting inventory |
|
AAPL |
SEC filing coverage confirmed for January 1, 2022 through December 31, 2024 |
Three relevant 10-Ks; no 10-K/A identified in the audited results |
|
MSFT |
SEC filing coverage confirmed for January 1, 2022 through December 31, 2024 |
Three relevant 10-Ks; no 10-K/A identified in the audited results |
These conclusions apply to the defined retrieval scope. They are not claims about amendments outside that date range or a guarantee that another retrieval will return identical counts.
The financial retrievals completed without retries. The additional SEC searches checked filing coverage and did not change the financial values, historical prices, growth calculations, or PASS/FAIL results.
Limitations and Analyst Review Requirements
The workflow addresses filing eligibility and source provenance. Several boundaries remain relevant when interpreting or adapting it.
Current retrievals do not establish historical database contents.
The observation timestamp records when this analysis retrieved the FMP responses. It does not show when FMP first ingested each filing or what its database would have returned on October 15, 2024. Maintaining a post-earnings data refresh checklist can help document changes to current model inputs, while historical reconstruction still requires explicit source and timing checks.
Acceptance does not identify the first public disclosure of every figure.
A company may publish selected results in an earnings release or another filing before its annual 10-K is accepted. This screen uses the specified annual-filing evidence. It does not inventory every earlier disclosure channel.
Filings near the cutoff need additional timing review.
A same-day filing requires comparison of the full timestamp and time zone with the decision cutoff. Acceptance alone does not establish the exact time a reader could access the document. An unresolved timing issue should remain visible rather than being converted into a claim of precise intraday availability.
A revenue key does not establish consolidated context.
The same financial concept can appear in consolidated statements and dimensional disclosures. If the corresponding statement-level value cannot be established reliably, the period remains Review Required.
Later amendments must not silently replace earlier evidence.
A 10-K/A may change information from an original filing. If the effect of an amendment on the selected metric cannot be established, analyst review is required. The coverage check in this example is limited to the specified 2022 through 2024 filing window.
Missing evidence must remain explicit.
A failed lookup, missing timestamp, unresolved filing-to-period match, or ambiguous financial value should not be replaced with standardized statements, quarterly figures, TTM data, or an inferred number. The workflow identifies the missing element and uses Review Required when it prevents a reliable calculation.
Historical price requires a matching trading-date record.
If the selected date has no end-of-day record, it should be reviewed rather than silently replaced with a nearby session. Price remains contextual in this example.
From Filing Evidence to a Reproducible Historical Screen
The Apple and Microsoft example shows how filing timing changes the annual comparison used in a historical screen. Apple's FY2024 had ended by October 15, 2024, but its supporting 10-K was accepted later, so the workflow used FY2023 versus FY2022. Microsoft's FY2024 filing qualified, allowing FY2024 versus FY2023.
The practical control is to establish the source filing, acceptance timestamp, and consolidated financial value before applying the screening rule. With those elements visible, readers can trace the selected periods and reproduce the calculation.
To apply the workflow to another decision date, define the cutoff, retain the retrieval observation time, and inspect both the financial evidence and filing coverage. Where the record cannot support a reliable comparison, the result should remain Review Required.
