Glossary term
Martingale
A martingale is a stochastic process whose expected future value, given current information, equals its current value.
Updated
Read time
What Is a Martingale?
A martingale is a stochastic process whose expected future value, given current information, equals its current value. In plain English, it is a model of a fair game: based on what is known now, the next expected value is neither higher nor lower.
In finance, martingales appear in asset pricing, derivatives models, risk-neutral valuation, and discussions of efficient markets. The concept is mathematical, but the intuition is useful: if prices fully reflect current information under the model, the next change should not be predictably profitable.
Key Takeaways
- A martingale has no expected drift after conditioning on current information.
- It is often described as a fair-game process.
- Martingales are important in no-arbitrage pricing and risk-neutral valuation.
- A martingale model does not mean prices cannot move sharply.
- The concept depends on the chosen probability measure and information set.
Basic Martingale Expression
A simplified one-period martingale condition can be written as:
In this expression, Xt is the current value of the process, Xt+1 is the next value, and \mathcal{F}t represents the information available at time t. The equation says that the conditional expected future value equals the current value.
What a Martingale Is and Is Not
Idea | Meaning | Common misunderstanding |
|---|---|---|
Fair game | No expected gain from current information alone. | It does not mean no volatility. |
Conditional expectation | Expectation is based on the current information set. | It is not an unconditional forecast. |
Risk-neutral pricing | Discounted asset prices may be martingales under a pricing measure. | It does not mean real-world returns are zero. |
Finance Context
Martingales are central to modern asset pricing because they help formalize no-arbitrage logic. Under a risk-neutral measure, a discounted asset price or payoff process may behave like a martingale. That allows future payoffs to be valued using conditional expectations under the pricing framework.
This does not mean actual investors expect no compensation for risk in the real world. The martingale property depends on the measure being used. Real-world probabilities and risk-neutral probabilities can describe different things.
The information set also matters. A process may look like a martingale under one model but not under another if additional information, transaction costs, funding constraints, or hidden risks change the expected payoff.
Trading Interpretation
The martingale idea is sometimes simplified into the claim that price changes are unpredictable. That can be directionally useful, but it is incomplete. A process can be a martingale and still have changing volatility, jumps, liquidity shocks, or tail risk.
For traders and investors, the key point is humility. If a market price already reflects available information, a predictable edge should not be assumed without a clear reason, cost estimate, and risk control.
The Bottom Line
A martingale is a fair-game process whose expected future value equals its current value given current information. In finance, it is foundational for no-arbitrage reasoning, risk-neutral pricing, and modern derivatives models.