Every card below shows the score and the exact filing data it was computed from — no unexplained numbers. Some signals (P-02, P-03, P-08, P-09, P-10, P-11) pull live documents from SEC EDGAR the first time you view this company; a few of those (P-03, P-10, P-11) use AI to read the filing text, but every AI claim must quote the actual document verbatim or it's discarded — never a guess.
Ask follow-up questions about the latest 10-K, with citations verified against the actual text. No account needed.
Run the bounded investigation agent — reads this company's signals and writes a short, cited summary.
{
"etr_by_year": {
"2014-12-31": -13.738,
"2013-12-31": 0.0186,
"2012-12-31": 0
},
"statutory_rate": 0.21,
"gap_by_year": {
"2014-12-31": 13.948,
"2013-12-31": 0.1914,
"2012-12-31": 0.21
},
"note": "Low ETR can be fully legitimate; this flags persistence + magnitude for review.",
"stale": false,
"latest_filing_period": "2014-12-31",
"source": {
"url": "https://www.sec.gov/Archives/edgar/data/1070523/000114420415031754/0001144204-15-031754-index.htm",
"form": "10-K",
"filed": "2015-05-15"
}
}