Analyses / Methodology
How our analysis works
Every Sigma Terminal analysis is generated from latest available provider data through a fixed, structured framework — not freeform opinion. Here's exactly how, and what it can and can't tell you.
The data behind every analysis
We start from provider market data, not the model's memory. For each ticker we pull:
- Latest available quotes — current provider price, day range, 52-week range, volume.
- Fundamentals — income statement, balance sheet, and cash-flow figures; margins, growth, returns on capital, leverage, and valuation multiples.
- Insider activity — recent insider buying and selling.
- Recent news — company headlines from the past week.
Market data is sourced from institutional data providers. Figures may be delayed or revised, and we stamp every analysis with an "as of" date so you always know how fresh it is.
How the analysis is generated
Gather. We assemble the structured data bundle above for the ticker.
Reason. A large language model (Anthropic's Claude family) analyzes only that supplied data against a fixed analyst framework — it is instructed not to invent figures that aren't present.
Structure. The output is a consistent research brief: a bull case, a bear case, a valuation verdict and score, technical levels, financial-health checks, catalysts, risk flags, and — for the full version — watch zone, invalidation level, and multi-horizon scenario ranges.
Refresh. Public analyses are refreshed on scheduled runs when provider data is available, so stale or missing data should be verified before use.
What the scores mean
Valuation score (0–100) weighs the multiples (P/E, P/S, P/FCF, etc.) against the company's growth and quality — lower means richer/more demanding, higher means cheaper relative to fundamentals. Confidence reflects how cleanly the data supports the setup. These are analytical signals, not guarantees.
Honest limitations
- AI can misread data or miss context. Treat every analysis as a starting point for your own research, never the final word.
- Data can be delayed, cached, or revised after the fact.
- Forecasts are probabilistic estimates, not promises. Markets are uncertain.
- Nothing here accounts for your personal financial situation, goals, or risk tolerance.