AugurSignals makes its analytical process inspectable. Every output traces to its inputs. Every limitation is visible alongside the result.
This page describes what that looks like in practice. For the calculation methodology itself, see Our Methodology. For why the platform does not make recommendations, see Why No Recommendations.
From Career Record to Source Filing
A user examining an insider's profile sees a career-level performance record built from the P-Factor methodology. That record is not a terminal output. It is the top of a navigable structure.
Each transaction that contributed to the career record is individually accessible. For every transaction, the platform constructs a structured analysis: whether the insider paid cash, how their share-level exposure changed, the approximate economic value of that change, and a confidence rating reflecting how interpretable the filing data is for that specific record. These four dimensions sit alongside each other rather than collapsing into a single score.
The transaction links to related activity — other transactions in the same Form 4 filing, and nearby transactions by the same insider at the same company within a seven-day window. The accession number is visible. The original SEC filing on EDGAR is linkable.
Each layer of analysis adds structure. None of them replace the layer beneath. The depth of this chain varies by data availability — not every insider's record carries the same richness.
What the Platform Shows When It Does Not Know
Every owner profile carries a data quality score — a visible measure of how much of the analytical framework the available data can support. When fewer than five qualifying transactions exist, the platform states this directly and identifies what is missing. Individual transaction analyses carry their own confidence rating: a filing with ambiguous instrument classification or missing price data receives a lower confidence grade than one with complete disclosure. Both grades are visible.
The interface adapts to data availability. Analytical cards appear only when sufficient data exists to populate them. When a section is absent, that absence is informational — it means the data could not support the analysis, not that the analysis was skipped. Where data exists but is incomplete, the platform states what percentage of records contributed and what was excluded.
The distinction between "no data displayed" and "data insufficient to calculate" is structural. Both states are visible to the user.
Historical Inspectability
Analytics can be queried at historical points. A user can ask what an insider's profile would have shown a year ago, and the platform constructs that view using only information available before the specified date.
No future data enters historical assessments. Calculations are deterministic — same inputs, same date, same outputs. This is the mechanism that enables honest backtesting: evaluating whether an insider's track record held up in real time, not only in retrospect.
Temporal correctness depends on the completeness of historical source data, which varies by insider and time period. The platform does not interpolate where records are absent.
Constraints Alongside Results
The platform documents its own analytical boundaries: single-vendor price data, fixed measurement horizons, equal transaction weighting, exclusion of sales from primary scoring. These are not separated into fine print or a different page.
Every owner profile includes a data snapshot that states what was calculated, which sources contributed, when the calculation was last run, and what the known limitations are. The methodology choices and the results they produce are visible in the same place.
Documenting constraints does not eliminate them. It makes the boundaries of the analysis visible rather than hidden.