Black-Scholes Option Calculator

An option is a promise with a choice built into it. The investor education site run by the U.S. Securities and Exchange Commission describes options as contracts that give the buyer the right, but not the obligation, to buy or sell a security at a fixed price within a set period. A call is the right to buy. A put is the right to sell. The fixed price is the strike.

That small word, right, is what makes options hard to value. A share is worth what it trades for. An option is worth something that depends on where the share might go, how fast it moves, how long the choice stays open and what money earns in the meantime. The same agency classes options among derivatives: instruments whose performance comes, at least in part, from the performance of something else.

A formula built on a frictionless world

In 1973 Fischer Black and Myron Scholes published “The Pricing of Options and Corporate Liabilities” in the Journal of Political Economy, and in the same year Robert Merton published “Theory of Rational Option Pricing” in the Bell Journal of Economics and Management Science. Merton and Scholes received the 1997 Nobel Memorial Prize in Economic Sciences for the work. Black had died in 1995 and was not eligible, though the prize committee mentioned his contribution.

Their central move was to hedge. Hold the share and sell the option in just the right ratio, keep adjusting that ratio as the price moves, and the combined position stops depending on where the share goes. A position with no risk must earn the risk-free rate, or someone could make money from nothing. The ratio must be revised continuously, which is why the method is called continuously revised delta hedging. That single argument fixes the price of a European option, one that can be exercised only on its expiry date. Notably, it does so without asking what return anyone expects from the share.

The argument needs a tidy world. The model assumes a constant risk-free rate; a share price that follows geometric Brownian motion with constant drift and constant volatility; no arbitrage; the freedom to borrow, lend, buy and short any amount, even fractions; and no fees or other costs. The original version also assumed the share paid no dividend. Later work relaxed that: for an index, a dividend paid continuously as a fixed fraction q of the price adds one term, and the formula keeps its shape.

Under those assumptions the value of a call is the spot discounted at the dividend yield times N(d1), minus the strike discounted at the risk-free rate times N(d2). The put mirrors it. N is the cumulative standard normal distribution. The quantities d1 and d2 combine the log of spot over strike, the rates, the volatility and the time left, and they differ by volatility times the square root of time.

Here is a worked example you can type into the calculator above. Take a spot of 100, a strike of 100, one year to expiry, a risk-free rate of 5 percent, no dividend and a volatility of 20 percent. Then d1 comes out at exactly 0.35 and d2 at 0.15. The call is worth 10.4506 and the put 5.5735, both in the same unit as the spot. At 25 percent volatility the call rises to 12.3360. The older page at this address offered exactly these inputs; it failed on load because it called a JavaScript error function that does not exist.

That last detail is not trivia. N looks innocent, but computing it well takes care. The normal distribution can be written through the error function, and the complementary error function, erfc, is one minus it. Python’s documentation notes that erfc exists for large arguments, where subtracting from one would lose significance. Deep in the tails, where an option is far from its strike, that loss is the whole answer.

So the calculator builds erfc itself from the NIST Digital Library of Mathematical Functions: a power series for arguments below one half and a continued fraction above, with the symmetry rule for negative arguments. Measured against Python’s own erfc on a dense grid reaching far into the tails, the largest relative difference was about one part in a thousand trillion.

What the Greeks measure

The Greeks are partial derivatives of the option value: each says how much the value moves when one input moves and the rest stay put. They are the language in which risk is discussed and limited, because a desk can cap its exposure to each one separately. They are also the place where small formula errors hide most easily, which is why every Greek here is checked against a finite difference of the value itself.

Delta is the change in value per unit change in the spot. In the worked example the call delta is 0.6368 and the put delta minus 0.3632; the two always differ by the discount factor for the dividend, which here is one. Delta is also the hedge ratio in the original argument, the number of shares that offsets one option for a small move. A trader who wants no exposure to direction keeps it near zero.

Gamma is the change in delta per unit change in spot, 0.0188 in the example, and it is the same for a call and a put with the same strike. It measures how quickly a delta hedge goes stale. Vega is the change in value per unit change in volatility, again the same for both. Markets usually quote it per volatility point, so the example shows 0.3752. Vega is not a Greek letter at all; the name seems to come from reading the letter nu as a V.

Theta is the change as calendar time passes with everything else fixed. For the example call it is minus 6.4140 a year, or about minus 0.0176 a day on a 365-day count; some desks divide by 252 trading days instead, and the calculator lets you pick. Rho is the change per unit of the risk-free rate, and the dividend rho does the same for the dividend yield. Both are shown per percentage point.

Two cases deserve plain statements. With no time left, the formula collapses to the payoff: the larger of spot minus strike and zero for a call. With zero volatility, it collapses to the discounted payoff of the forward. At the strike itself some Greeks have no finite value, and the calculator says unbounded rather than printing a number that only looks precise.

Implied volatility turns the formula around

Of the inputs, only volatility cannot be read off a screen. The rate is published, the strike and expiry are in the contract, the spot is the last trade. Volatility is the average future variability of the share, and nobody knows it in advance. That gap invites the model to be run backwards.

Because the option value rises steadily with volatility, at most one volatility can reproduce a given market price. That number is the implied volatility. There is no general closed form for it, so it is found by root finding. The Wikipedia article on the subject gives a call with strike 50, 32 days to run, a 5 percent rate and a spot of 51.25 trading at 2.00, and reports an implied volatility of 18.7 percent; the solver here agrees to the first decimal.

In practice implied volatility has become a quoting convention. Two options with different strikes and expiries are hard to compare by price, but easy to compare by the volatility each price implies. In the worked example, a call price of 12 implies 24.1117 percent. The solver takes Newton steps, which DLMF notes roughly double the correct digits each time near the root, and falls back to bisection whenever a step would leave the bracket known to hold the answer.

Some prices imply nothing. A call can never be worth less than the discounted payoff of the forward, nor more than the spot discounted at the dividend yield. In the worked example the lower bound is 4.8771, so a call price of 4 has no implied volatility. The calculator then names the bound that was crossed rather than returning a number.

That same 4.8771 appears in put-call parity. A long call and a short put on the same strike and expiry together behave like a forward contract, so the call minus the put must equal the discounted forward minus the discounted strike. Parity rests only on static replication, so it needs far weaker assumptions than the full model. The calculator shows both sides, and in the worked example both are 4.8771.

Where the model breaks down

The assumptions are not facts about markets. Real prices jump, so extreme moves come far more often than a lognormal model allows; trading costs money and sometimes cannot happen at all; volatility drifts and clusters; and interest rates move. The model also tends to misprice options far from the strike, and it assumes a positive price for the underlying.

Nor does it value American options, which can be exercised early. For an American put there is no general closed form, and this calculator does not attempt one. Its numbers are model values for European exercise and nothing more.

None of this makes the formula useless. It makes it a ruler: exact under stated rules, and honest about which rules those are. This page computes that ruler carefully and shows its working. It is not trading advice, it recommends no trade, and a model value is not a forecast of any market price.