Methodology & Data Sources
Transparency is the foundation of trustworthy analysis. This page documents exactly where Stock Smith's data comes from, how the models work, what the AI does and does not decide, and the limits you should keep in mind when using our numbers.
Data Sources
Every analysis starts with real market data. Price history, quotes, and option chains are sourced from public market data feeds (via Yahoo Finance). For each ticker we pull approximately two years of daily open-high-low-close-volume history, plus a real-time quote for the current price and recent change.
We do not use proprietary or paid alternative-data sets, and we do not fabricate prices. When live data is unavailable for a symbol, the analysis says so explicitly rather than silently substituting a stale or made-up value.
The Simulation Model
Price projections use a Monte Carlo simulation built on Geometric Brownian Motion (GBM) — the same model that underpins Black-Scholes option pricing. From the two years of daily returns we estimate two parameters: drift (average return) and volatility (the size of the swings).
We then simulate 1,000 independent one-year price paths, each taking a different random walk seeded by those parameters. Summarizing the 1,000 endings gives the expected price, the 80% and 95% confidence bands, the probability of finishing above today's price, and the modeled best- and worst-case levels.
Where options income is modeled, premiums come from the live option chain when a bid is quoted, with a Black-Scholes estimate used as a transparent fallback when no live bid is available — and the source is labeled either way.
What the AI Does — and Doesn't
The AI layer is given the simulation's quantitative output — expected price, probability of gain, volatility — alongside current market context. From that it produces written sentiment, bullish and bearish catalysts, key risks, and example trade strategies with entry, target, and stop-loss levels.
The AI does not invent the numbers. The math comes from the simulation; the AI explains and contextualizes it. If the AI service is unavailable, the platform falls back to deterministic, rules-based strategies so you always receive a grounded result rather than an error.
Assumptions & Limitations
Our models assume the future will broadly resemble the recent past. That assumption breaks down around earnings surprises, regime changes, and rare shocks. Geometric Brownian Motion also assumes normally distributed returns, which understates the frequency of extreme moves — real markets have fatter tails than the model.
Confidence bands are probabilities, not fences: a 95% band still leaves real scenarios outside it. Treat every projection as a disciplined map of uncertainty to inform position sizing and risk management — never as a guarantee or a recommendation.
Editorial Integrity
Stock Smith is a research and education company. We do not accept payment to feature, rate, or recommend any security. Our educational guides and methodology are written and reviewed by the Stock Smith research team. Nothing on this site is personalized investment advice, and we are not a registered broker-dealer or investment adviser.
Stock Smith is a research and educational tool. Its simulations, AI analysis, and trade strategies are probabilistic models based on historical data and do not predict the future. They are not investment, financial, legal, or tax advice, and nothing here is a recommendation to buy or sell any security. Markets carry risk, including loss of principal, and past performance does not guarantee future results.