Trust and Transparency
Evidence & Validation
Direct answer: Instead of a cherry-picked performance track record, Swoopr Investment publishes auditable process evidence: automated formula tests for every calculator, a documented reference-sourcing standard with automated coverage checks, a real correction log, and a versioned methodology history. This page is the index of that evidence, not a marketing summary.
Why evidence instead of performance claims
A single backtested strategy, a cherry-picked example, or a "Swoopr score" applied retroactively to winning trades is easy to construct and easy to make misleading, on purpose or by accident. Swoopr Investment does not publish a proprietary stock index, model portfolio, or track record built to demonstrate that a methodology "works." Instead, this page collects the parts of the site that are genuinely auditable: process, not outcomes. See The Swoopr Research Loop for the research process this evidence backs, and Risk Disclosure for why no backtest, simulation, or score is a promise about future results.
Formula validation
Every interactive calculator and simulator on Swoopr Investment keeps its formula logic in a dedicated, DOM-free JavaScript module (the site's *-core.js convention), separate from the page that displays it, imported directly by that tool's automated unit tests. As of this page's last update, 142 such modules exist across the site, each with its own direct test coverage rather than relying on end-to-end UI tests alone to catch a broken formula. This structure is what makes a formula auditable: the calculation a reader sees in the browser is the exact same code path the automated test suite runs against known inputs and expected outputs, there is no second, undocumented copy of the math.
Data source catalog and freshness
Where a page's claim depends on live or third-party market data, rankings, or AI-generated summaries, Data, Rankings & AI Methodology Disclosure documents where that data originates and what it does and doesn't represent. Where a page's claim depends on a regulatory rule, tax treatment, or protocol specification, our Citation & Sources Policy documents the source hierarchy used and requires a visible References section. Time-sensitive pages carry a review cadence and a recorded next-review date; an automated check flags any page that has gone past its scheduled review date, so a stale rule page doesn't just sit unreviewed indefinitely.
Reference coverage
High-risk educational pages, content involving taxes, brokerage and trading rules, and regulated account types, are required to cite at least one primary or institutional source in a real References section before they're considered compliant with our sourcing standard. That requirement is enforced by an automated check that reads each page's actual References section and its actual cited sources, not a self-reported claim, run on every commit and in CI. A page is never allowed to declare a stronger review tier than what it actually cites.
Correction log
Under our Corrections Policy, a factual error reported and confirmed results in a page update and a changed "Updated" date. This is the real, append-only log of those corrections, generated directly from Swoopr's internal correction record, not hand-edited on this page.
- 2026-08-21 · /legal/terms/: Positioning language described Swoopr Investment as "a market-information and research platform", a pre-rename phrasing already corrected on other pages; aligned to the canonical description used site-wide (internal review)
- 2026-08-21 · /trading-technology/market-data-engineering/trades-quotes-bars-and-depth-feeds/: Order-routing disclosure requirement was misattributed to FINRA; Rule 606 is an SEC rule (Regulation NMS), corrected and verified against the site's own glossary entry (internal review)
Methodology version history
The Swoopr Research Loop is versioned so a material change to the process itself is visible, not silently edited in place:
- v1.0, published August 21, 2026, first public version: Define, Understand, Measure, Compare, Value, Stress-Test, Document, Review.
Content review cadence
Glossary terms and pages covering laws, taxes, regulatory rules, and protocol specifications carry an explicit review frequency (quarterly for rules-based content, annual otherwise) and a recorded next-review date, following the same convention across the glossary, the Answer Hub, and the Citation & Sources Policy. An automated report checks every page against its own recorded next-review date and flags anything overdue, run in continuous integration on every push, not only when someone remembers to look.
Security and data handling summary
Public educational content and tools require no account and store data locally in your browser, not on Swoopr's servers. For users of the separate, invite-only trading dashboard, Composer.trade and OpenAI API keys are encrypted at rest with AES-256-GCM under a master key held in the operating system's own credential manager, never saved in plain text, and used only to make requests on that signed-in user's behalf. See Privacy Policy for the full data-handling policy.
AI-assisted content
AI tools may assist with research organization, outlining, and drafting across this site; Swoopr Investment remains responsible for the final content it publishes. See AI-Assisted Content for the full disclosure, including what AI does not guarantee and how AI-generated examples and calculations are checked.
Frequently Asked Questions
Why does a calculator keep its formula in a separate module?
So the calculation can be tested independently of the page that displays it. A formula written inline in a page script can only be verified by loading the page and trying it; the same formula in a dedicated module without any browser dependency can be called directly by automated tests that check it against worked examples and edge cases. It also prevents the same calculation from being implemented twice and drifting, since one module serves every surface that needs it.
How can a reader check that a calculator matches its stated method?
By working the published formula through by hand on the page's own worked example and comparing the result to what the tool returns. Every calculator states its formula in plain text alongside the inputs it expects, which is what makes that check possible without reading any code. Where a tool's assumptions materially limit the result, those are stated on the same page rather than left to be inferred from the output.
What does the correction log include, and what does it leave out?
It records confirmed factual corrections: what was wrong, on which page, and when it was fixed. It does not record routine editorial work such as rewording, restructuring or adding new material, because those are not corrections and logging them would bury the entries that matter. A page can therefore be substantially revised without generating a log entry, while a small factual fix does generate one.
What is an append-only log, and why does it matter here?
Entries are added and never removed or rewritten, so the record of past errors stays visible after they are fixed. A log that could be edited would let an inconvenient entry quietly disappear, which would make the whole record uninformative. Keeping it append-only means the log grows as a history rather than describing only the present, and it is generated from the internal correction record rather than maintained by hand on the page.
What does a stated review cadence commit the site to?
Revisiting specific content on a schedule rather than only when someone notices a problem. Content covering laws, taxes, regulatory rules and protocol specifications carries an explicit review frequency and a recorded next-review date, with rules-based material on a faster cycle than the rest. The commitment is to the review itself. It does not assert that an external rule cited on a page remains unchanged between reviews, which is why the review date is published alongside the content.
Does an automated reference check confirm that a source supports the claim?
No, and the distinction is important. Automation can verify that a References section exists, that its links resolve, and that a page in a category requiring citation has one. Whether a specific source actually supports the specific sentence next to it is an editorial judgment that no check performs. The automated layer catches structural absence and link rot; it does not substitute for reading the source.
What is the difference between a test passing and a formula being correct?
A test confirms that the code produces the result the test expects. If the expectation itself encodes a misunderstanding of the formula, the test passes and the formula is still wrong. That is why the published formula on the page, the worked example and the test are treated as three separate artifacts that have to agree: a reader checking the worked example by hand is verifying something the test cannot verify about itself.
What does the public site store about a visitor?
The public guides and tools require no account, and the calculators hold their inputs in the browser rather than transmitting them, which is why a tool can be used without anything being sent anywhere. Analytics and consent handling are covered by the Privacy Policy, which is the authoritative description rather than this summary. The separate invite-only dashboard operates under different arrangements, described elsewhere on this page.
Where does the material on this page come from?
From the site's own operational records rather than from a narrative written about them. The correction log is generated from the internal correction record, the module and coverage figures come from the test suite as it stands, and the review cadence reflects the metadata actually recorded against content. That sourcing is the point of the page: it collects checkable facts about how the site is built, which is a different claim from asserting that the content is good.