LessonThis is a lesson restated in new words. Any figure in it is an example from the source, not this desk's record.
Black-Scholes does not tell you where a stock will be tomorrow. It prices a European call as the discounted average of the contract's payoff over the paths allowed by its assumptions. Hand it spot, strike, volatility, the financing rate, and tenor, and it returns that average. In the lecture's example those inputs are 100, 100, 30 percent, 5 percent, and one year, and the call price is 14.23. A quote of 14.10 is below that model value. Across 100,000 simulated paths the average gain and loss of buying at 14.10 was positive. One path still lost. The average is not a claim that a trade will make money.
Why it matters
The model is the solution of the Black-Scholes partial differential equation. That equation comes from a hedging argument: hold the option and a changing amount of stock so the portfolio's random piece cancels, then the hedged portfolio should earn the riskless rate if there is no arbitrage. The stock in that argument follows geometric Brownian motion, a continuous random path with constant volatility. Those assumptions are the reason a formula exists. They are also the reason the formula can be wrong in the world. If the true dynamics were known, no model would be required.
Each contract has one payoff at expiration. You do not get the average. You get one path. The model is useful only as a statement about many repetitions under its own rules.
The idea in plain words
The lecture draws the strike as one line and a simulated price path as another. Finish above the strike by more than the premium and the long call gains. Finish lower and the loss stops at the premium. Upside is unbounded. Repeating the draw produces a different path every time. One illustrated path made 46 after a premium of 14.23. Another finished above the strike and still lost, because the gain did not cover the premium. Averaging the discounted results of all such paths is what the formula compresses into 14.23. Now put that number next to a market. The lecture's market maker quotes 13.43 at 14.10, so the call can be bought at 14.10 against a model value of 14.23. Quotes move for many reasons. The lecture sets those reasons aside and keeps the gap as a teaching case. Buying at 14.10 can lose on the next path. The lecture says so while the simulation is still on the screen.
How to use it
Use the formula as a benchmark under stated inputs, not as a forecast. If you believe the assumptions for the moment, a purchase below the model value has a positive expected payoff in the model, and a purchase above it has a negative one. The lecture's 14.10 versus 14.23 is that comparison. Surviving the short run is part of the comparison. The casino analogy in the lecture is only about repetition: a small edge needs many plays, and a player who is ruined early does not receive the long-run average. Then change an input on purpose. The lecture's warning example pushes the drift assumption to a negative 50 percent case and says the whole path of results changes. Volatility is the input that most often fails in silence. It is not constant, it clusters, and prices jump. A Heston-style model lets volatility be its own random process. Jump models add discontinuities.
What usually goes wrong
Treating one green path as confirmation. The path that made 46 and the path that missed the premium are both allowed by the same 14.23 price. A second mistake is levering a single expiration because the model value sits above the quote. The lecture's own picture is the opposite: the positive average showed up across tens and hundreds of thousands of simulated contracts, not in one bet. A third is ignoring the shortfall between model and market that comes from bad inputs. If the 30 percent volatility is the wrong constant, 14.23 is the right output of the wrong problem, and buying 14.10 is not a discount to a meaningful value.
The lecture is explicit that geometric Brownian motion omits clustering and jumps. Using the formula without naming which assumption you think the market has wrong turns the 0.13 gap into a slogan.
Questions people ask
Which five numbers produce 14.23?
Spot 100, strike 100, volatility 30 percent, interest rate 5 percent, and one year to expiration. Change any one of them and the model price moves. The market quote in the example is 13.43 bid, 14.10 ask.
Does a positive average across 100,000 paths mean the next trade wins?
No. Each path is one payoff. The average was positive for buying at 14.10 under the simulated dynamics. The lecture also shows paths that lose the premium.
Why not just predict the terminal price?
Under the model's own motion, the terminal price is a draw from a distribution. Predicting one draw is a different game from pricing the average of the payoff. The lecture calls the first game a gamble and spends the article on the second.