What Is a Monte Carlo Simulation

Monte Carlo in five lines

A Monte Carlo simulation takes the trades a rule actually made, shuffles their order and replays them, many times over. The reason is that the order the trades came in was one accident among many, and while the total does not care about order, the drawdown depends on it entirely. What it gives is a spread of the drawdowns and the endings the same trades could have produced. What it cannot do is invent trades the rule did not make; it reshuffles, it does not predict. In the example, the plain 10:00 to 18:00 rule's 135 trades on the euro in 2026 were shuffled a thousand times: the drawdown came out anywhere from 178 to 599 pips, where a pip is the fourth decimal place of the price, and the real order gave 299.5.

Tiles drawn from a bag

Picture a bag of tiles, one tile per trade the rule made, green for the winners and red for the losers, 68 green and 67 red. Draw them all out and lay them in a row, and you have one story of the rule, with one worst run in it. Put them back, shake, and draw again: the same tiles, the same sum of 150.3 pips, and a different row, with a different worst run. A thousand draws is a thousand rows, and the drawdown of each is a number the real order could just as well have given, because the real order was one draw.

This paper is locked