About the Model

Probabilist was built around a small number of first principles about how stock options derive their value, and where existing tools in the market fall short.

First principles

The first principle is straightforward: the value of an option at expiry is determined by the relationship between the price of the underlying stock and the strike price of the option itself. As an example, for a call option to expire profitably the stock price at expiry must exceed the strike price plus the premium paid for the option.

The second principle follows naturally from the first. The current price of a stock reflects the market’s collective expectations about that company’s future earnings and prospects. Those expectations are informed, at least in part, by the company’s current financial position — as reported in its income statements, balance sheets, and statements of cash flows.

Taking these principles to their logical conclusion, we believe that whether an option ultimately expires profitably must be influenced (at least partially) by the company’s underlying fundamentals.

However, when surveying most tools available to retail traders today, we found that the overwhelming majority focus almost exclusively on technicals — such as option Greeks, implied volatility, and price patterns — while largely ignoring fundamentals.

This gap between first principles and what is commonly available in the market is what motivated the creation of Probabilist.

How the system works

At a high level, the system ingests several categories of publicly available historical data:

• Company financial statements (income statements, balance sheets, and cash flow statements)
• Historical stock prices, including trailing return and volatility characteristics
• Historical options prices and contract details, across calls, puts, and long/short strangles
• Sector classification and analyst rating trends
• Market-wide conditions, such as volatility (VIX) and Treasury yield trends

These inputs are combined and fed into a machine learning model — trained separately for calls, puts, and long and short strangles, so that each strategy's distinct payoff structure is learned on its own — to estimate the probability that a given option will expire profitably. In this context, “profitable” means that, at expiry, the option's payoff exceeds the premium paid for it.

Each model is validated using walk-forward testing: it is repeatedly retrained on an expanding window of historical data and evaluated only on the subsequent period it has not seen, across many sequential windows spanning multiple years and market conditions. This is meant to approximate how the model would have performed if it had been used in real time, rather than relying on a single historical snapshot.

Rather than producing a binary “yes” or “no” outcome, the model outputs a probability. This reflects both the inherent uncertainty in financial markets and the reality that no single set of inputs can fully explain price movements.

Important limitations and assumptions

Probabilist does not claim that a company’s financial statements are the only factors that influence option outcomes. In practice, many additional forces matter, including (but not limited to):

• Options Greeks and implied volatility
• Macroeconomic conditions
• Market sentiment and news events
• Liquidity and supply–demand dynamics

For this reason, Probabilist presents its output as probabilities rather than absolute predictions. The goal is not to replace trader judgment, but to augment it by surfacing information that is often overlooked in traditional options analysis.

Intended audience

This model assumes that users have a baseline understanding of stock options and how they work. If you are new to options trading or would like a refresher, we recommend the following resource:

Investopedia — Options Basics Tutorial