1. The screen (all markets)
Across five asset classes — stocks, indices/ETFs, forex, commodities, and crypto — Aethon builds a macro backdrop from official, dated sources: interest-rate and inflation data (FRED), global growth (IMF), world-news risk (GDELT), and live market prices. Each instrument is scored for market context strength (0–100, how well current inputs align) and a direction, with volatility and data-freshness penalties applied. This is a screening layer, not a recommendation.
2. The report (one ticker, on demand)
When you open a report, Aethon runs up to seven evidence providers live for that ticker: SEC filed-company data, verified catalysts, official fund facts, two-sided forex policy settings, crypto protocol/institutional facts, gold-macro drivers, and energy fundamentals. The report states what each source did — and did not — return.
3. The report-grade conclusion
The report's headline confidence is not the raw screen score. It is adjusted, transparently, by:
- Evidence coverage — how much applicable official evidence actually returned. Thin coverage pulls the score toward neutral.
- Outcome calibration — how accurate past conclusions for that setup have been. Until enough outcomes are measured, a conservative prior applies.
- Regime fit — whether the directional lean works with or against the current market regime.
The report shows this adjustment in full ("how this score was set"), plus a strongest-case-against counter-thesis. We lead with what could be wrong.
4. The quality gate
Aethon only sells a report for a ticker whose official-evidence bundle is complete enough to support a paid conclusion. Tickers that fail are marked developing coverage and are not offered for purchase. We would rather decline than overstate.
5. Calibration — measuring whether we were right
Every conclusion records the directional setup at the time. After a fixed horizon (currently 7 days), Aethon measures whether the stated direction matched the later price move and stores the outcome. For the public track record, outcomes are deduplicated to one per ticker per day (so a single batch of correlated calls can't inflate the count), a headline hit-rate is only shown once at least 50 independent directional outcomes exist, and it is always published with a 95% confidence interval so a modest sample is never read as a precise number. This measures whether the direction was right — it is not a return forecast, and past results do not guarantee future ones.
6. What the numbers are — and aren't
- Market context strength = how well current inputs align. Not a probability of profit.
- Conclusion support = how complete the official evidence is. Not a price target.
- Hit rate = measured directional accuracy over the horizon. Not advice to trade.
7. Sources
FRED (macro), IMF (growth), GDELT (world-news risk), SEC EDGAR (filings), and licensed market and fund-data providers. Data may be incomplete, delayed, or unavailable; reports reflect what was retrievable at the moment they were generated.
Questions about the method: info@fumie.org.