Skip to content
Menu
Finabase Learn

Trading · Options

How options are priced: the intuition behind the models

Why an option’s price is the cost of copying it with stock and cash, shown with a one-step example, and what Black-Scholes and binomial models assume and get wrong.

On this page
  1. The idea: copy the payoff
  2. How it works
  3. Where it fails
  4. Compared with its neighbours
  5. How Finabase does it
  6. Frequently asked questions
  7. Summary
  8. Check your understanding

An option pricing model answers one question: what would it cost to build something else that pays exactly what the option pays? If a mix of stock and cash can copy the option, the option must cost what the mix costs. Anything else would hand a free profit to whoever bought the cheaper of the two and sold the dearer.

That idea, called replication, is the whole foundation. It explains a fact that surprises most people: the price of an option does not depend on whether the stock is expected to go up. It depends on how far the stock can move, on time, and on the cost of money.

The Option premium and time value: decay, volatility and Black-Scholes lesson lists what moves a premium and introduces the Black-Scholes formula. This page goes underneath it. It prices an option by hand in a world with two outcomes, shows how that small example grows into the models traders use, and is plain about what those models assume.

The idea: copy the payoff

Take the smallest market there can be. A stock is at $100. One period from now it will be either $110 or $90. Money earns no interest, to keep the arithmetic clean. A call with a $100 strike will then be worth $10 if the stock rises and nothing if it falls. What is the call worth today?

Pricing a $100 call by copying it: one period, two outcomes, no interest Today Stock: $100 $100 call: worth ? Stock rises to $110 Call pays $10. Copy: 0.5 share − $45 loan = $55 − $45 = $10 Stock falls to $90 Call pays $0. Copy: 0.5 share − $45 loan = $45 − $45 = $0 The copy costs 0.5 × $100 − $45 = $5 today, so the call must cost $5 Half a share and a $45 loan pay what the call pays either way. No probability was used.
Illustrative one-period example with no interest. Half a share financed with a $45 loan pays the same as the $100 call in both outcomes, so the call is worth what that portfolio costs: $5.

The figure shows the answer before any formula. Buy half a share for $50 and borrow $45 to help pay for it. Your own outlay is $5.

  • If the stock rises to $110, half a share is worth $55. Repay the $45 loan and $10 is left.
  • If the stock falls to $90, half a share is worth $45. Repay the loan and nothing is left.

That is exactly what the call pays. Two things with identical payoffs must have the same price, so the call is worth $5.

Notice what was never used: the chance that the stock rises. It could be 50% or 90%. The copy works in both outcomes whatever the odds, so the odds cannot matter. A trader who is sure the stock will rise still cannot justify paying $7 for this call, because someone else would sell it to them at $7, build the copy for $5, and keep $2 with no risk.

How it works

The recipe

For one step, two numbers define the copy.

Volatility is the input that matters

Widen the two outcomes to $120 and $80, with the stock still at $100 today.

The range doubled and the call's value doubled, from $5 to $10. Nothing was said about the stock being more likely to rise. A wider spread of outcomes makes the option worth more because the option keeps the good outcome and discards the bad one. This is the reason volatility is the main input to every pricing model.

Risk-neutral probabilities

There is a shortcut that gives the same prices. Ask: what probability of an up move would make the stock itself fairly priced if investors demanded no reward for risk? In the first example the stock is $100 and can be $110 or $90 with no interest, so that probability is 50%: 0.5 × 110 + 0.5 × 90 = 100. Apply the same probability to the call: 0.5 × $10 + 0.5 × $0 = $5.

This is called risk-neutral pricing. The 50% is not a forecast. It is a bookkeeping device that reproduces the cost of the copy. Probabilities that models print, such as "chance of finishing in the money", are of this kind, which is one reason they should be read loosely.

From one step to a model

One step with two outcomes is too crude. Split the period into two steps, then ten, then a thousand, with a small up or down move at each. At every node the same recipe applies, working backwards from expiration to today. This is the binomial model of Cox, Ross and Rubinstein (1979). It handles early exercise naturally: at each node, compare the value of holding with the value of exercising, which is why it is widely used for American-style stock options.

Let the steps shrink toward zero and the tree's range of final prices becomes a smooth curve. The option value converges to the Black-Scholes formula (1973), extended to dividends by Merton. The copy is still there: the formula's first term is delta shares of stock, the second is the loan.

Where the model thinks a $100 stock can be in 45 days (30% volatility) 0.00 0.01 0.02 0.03 0.04 $65 $75 $85 $95 $105 $115 $125 $135 $145 Strike $105 Relative likelihood of each price Only this tail pays the $105 call Everything left of the strike pays $0 Stock price at expiration Likelihood (density)
Illustrative: the range of prices the Black-Scholes model assigns to a $100 stock after 45 days at 30% volatility and a 4% rate. A $105 call is paid only by outcomes to the right of the strike, and paid more the further right they are.

The curve shows what the formula is doing. For a $105 call, every outcome left of the strike pays nothing. Outcomes to the right pay the distance beyond $105. The call's value is the average of those payments over the whole curve, discounted to today. With the inputs in the figure the model puts the chance of finishing above $105 at about 32% and values the call at $2.42.

Change only the volatility and the curve changes width. At 15% the same call is worth $0.61; at 60% it is worth $6.52. The chance of finishing above the strike moves much less, from 20% to 38%. Most of the extra value comes from how far beyond the strike the stock can go, which is the same lesson as the $120 and $80 example.

Checking the model against a published value

The figures on this page and across the options lessons come from one implementation of the Black-Scholes formula. A standard textbook example has a stock at $42, a strike of $40, six months to expiration, 20% volatility and a 10% interest rate, and gives a call value of $4.76 and a put value of $0.81 (Hull). The implementation used here returns $4.7594 and $0.8086.

Where it fails

Every result above depends on being able to hold the copy and adjust it. The models assume conditions that markets only approximate.

Volatility is not constant. Black-Scholes takes one volatility for the life of the option. Actual volatility changes daily, and the market prices each strike at its own level. The pattern is described in Volatility skew: why implied volatility differs by strike.

Prices jump. The copy has to be adjusted as the stock moves: more shares as it rises, fewer as it falls. That works if the price moves smoothly. An earnings gap or an overnight shock moves the stock past many adjustment points at once, and the copy fails exactly when it is needed. Sellers charge for that risk, which lifts the price of options that pay in a jump.

Trading is not free or continuous. Each adjustment costs a bid-ask spread and commissions. In practice the copy is adjusted at intervals, and the result differs from the model.

Returns are not bell-shaped. The model's curve underestimates the frequency of very large moves. Deep out-of-the-money options are worth more in the market than a single-volatility model says.

Early exercise and dividends. Black-Scholes values a European option on a stock without dividends. An American put, or a call just before a dividend, can be worth more because it can be exercised early. Tree models allow for that.

A model cannot supply its own input. Feed it a volatility and it returns a price. The volatility is the unknown. In practice the model runs in reverse, from price to implied volatility, and serves as a common language for quoting.

Compared with its neighbours

Approach Exercise style it handles Volatility assumed Used for
One-step copy European Two outcomes Understanding why the price is what it is
Binomial tree European and American Constant (can be varied) US stock options, dividends, early exercise
Black-Scholes European Constant Quoting implied volatility, quick Greeks
Black-Scholes-Merton European Constant, dividend yield Index options, stocks with steady dividends
Simulation Any, with more work Whatever the modeller picks Options whose payoff depends on the price path

All of them rest on the same replication argument. They differ in how they describe the stock's possible paths and in what they can handle.

How Finabase does it

Finabase does not run an option pricing model on its pages. An option chain such as AAPL options shows market quotes: bid, ask, last price, change, volume and open interest for each contract, with the implied volatility reported by the quote provider. There is no theoretical value column, no probability of finishing in the money and no Greeks.

The chain does give you every observable input to a model: the stock price in the header, the strike in the centre column, the expiration you selected, and the implied volatility of the contract. The figures in the options lessons are computed from inputs like these with the Black-Scholes formula, and each lesson states its inputs so you can reproduce them. Quotes may be delayed, and Finabase does not place orders.

Frequently asked questions

Why does the expected return of the stock not affect the option price?

Because the option can be copied with stock and cash, and the copy pays the same in every outcome whatever the odds. The stock's expected return is already in the stock price, which the copy uses.

Is the Black-Scholes price the correct price?

It is the correct price if its assumptions hold and you know the volatility. Neither is true in practice. The market price is the price; the model is a way to translate it into an implied volatility and into sensitivities.

What is the difference between the binomial model and Black-Scholes?

The binomial model moves the stock in small up-and-down steps and works backwards through a tree. Black-Scholes is the result as the steps become infinitely small. For a European option with no dividend they agree; the tree can also handle early exercise.

Does a delta of 0.5 mean a 50% chance of finishing in the money?

Roughly, not exactly. Delta is the number of shares in the copy. The model's chance of finishing in the money is a related but different figure, and both are risk-neutral quantities, not forecasts. The Option delta: share equivalence, hedge ratio and its limits lesson compares the two.

Summary

An option is priced by finding the portfolio of stock and borrowed cash that pays the same amount in every outcome; the option must cost what that portfolio costs. In a two-outcome example the copy is half a share and a $45 loan, and the call is worth $5 whatever the odds of a rise. Repeating the step gives the binomial model, and shrinking the steps gives Black-Scholes. The price depends on the spread of outcomes, time and interest, and not on the stock's expected return. The models assume smooth prices and constant volatility, so in practice they serve to translate market prices into implied volatilities and Greeks.

Check your understanding

4 questions on this lesson. Each answer comes with the reason, and nothing is scored.

  1. A stock at $50 will be $60 or $40 next period, with no interest. What is a $50 call worth, and how is it copied?

    1. $10: one share and no loan
    2. $5: half a share and a $20 loan
    3. It depends on the probability that the stock rises
    Show the answer

    B. $5: half a share and a $20 loan Delta = (10 − 0) ÷ (60 − 40) = 0.5 shares. The loan is 0.5 × 40 = $20. Cost today: 0.5 × 50 − 20 = $5. In the up state the copy is worth 30 − 20 = $10 and in the down state 20 − 20 = $0, the same as the call. The probability never enters.

  2. Two analysts agree on a stock’s volatility. One expects it to rise 20% this year, the other expects it to fall. According to the replication argument, how do their fair values for a one-year call compare?

    1. The same
    2. The optimist’s value is higher
    3. The pessimist’s value is higher
    Show the answer

    A. The same The call is worth the cost of the stock-and-cash portfolio that copies it, and that cost depends on the stock price, volatility, time and interest. The expected return is not an input. The analysts would disagree about whether to own the call, not about what it costs to copy.

  3. In the one-step example the two outcomes widen from $110 and $90 to $120 and $80. Why does the $100 call double in value?

    1. Because the stock has become more likely to rise
    2. Because the loan in the copy becomes larger
    3. Because the call keeps the larger gain and still loses nothing in the bad outcome
    Show the answer

    C. Because the call keeps the larger gain and still loses nothing in the bad outcome The call pays $20 instead of $10 in the up state and $0 in the down state either way. A wider spread of outcomes raises the average payoff of something that cannot go below zero. This is why volatility is the main input to an option’s price.

  4. Which assumption of the Black-Scholes model fails most visibly on the night of an earnings report?

    1. That the interest rate is constant
    2. That the stock price moves smoothly, so the copy can be adjusted along the way
    3. That the option has a fixed strike
    Show the answer

    B. That the stock price moves smoothly, so the copy can be adjusted along the way The copy must be adjusted as the stock moves. A gap takes the price from one level to another with no chance to trade in between, so the copy breaks. Sellers charge for that, which is one reason options before an event carry a higher implied volatility.

Sources and further reading

  1. Black, F. and Scholes, M., The Pricing of Options and Corporate Liabilities, Journal of Political Economy 81(3), 1973
  2. Merton, R. C., Theory of Rational Option Pricing, Bell Journal of Economics and Management Science 4(1), 1973
  3. Cox, J., Ross, S. and Rubinstein, M., Option Pricing: A Simplified Approach, Journal of Financial Economics 7(3), 1979
  4. Hull, J., Options, Futures, and Other Derivatives (Pearson): the worked Black-Scholes-Merton example with a $42 stock and a $40 strike
  5. The Options Clearing Corporation — Options Disclosure Document (Characteristics and Risks of Standardized Options)