When something gets this complicated, it becomes very challenging to directly calculate the odds. However, this is pretty shaky, since there may be tables in the way or he may only be able to take 12 total steps, or maybe the door to the restroom is down a hallway, so you have to be able to account for all of that. This would give you an idea of how much of the room counts as a victory - and thus some first-level approximation of his chance of success. What if you wanted to know the odds that your sauced compatriot will actually make it to the bathroom? The traditional way to do that would be to calculate what percentage of the room is covered by the bathroom, and then take a ratio of bathroom to unbathroom. Let's talk about how this is a great tool. Until we reach some sort of stopping condition.Īt this point, you may be thinking, "Great, so we can essentially map randomness. Then we take 1-step forward based on whatever the dice said. In this case, we can roll a dice to decide what angle our inebriated friend is going to step. Take a look at a possible path in the image below, and then we'll talk about why this counts as a Monte Carlo.Ī Monte Carlo simulation means that we're using a set of dice to decide how our actor behaves. This idea is sometimes called the "drunkard's walk" and we can look at it by using a Monte Carlo approach. like, (hic) come'on man (hic) I'm finnnennene." He has no control over his limbs at this point and in order to walk he just staggers randomly in any direction. He stands up and immediately proclaims that he's "not that drunk. He decides that he needs to use the restroom. Let's imagine there's a very, very drunk guy at a bar. To get started, let's take a look at a simple example. In this first section, we'll start out just by discussing what a Monte Carlo simulation is in the first place.
#What is monte carlo simulation series
To address that, I've decided to put together a series of small projects that demonstrate the power of Monte Carlo methodology in a few different fields. However, I've found that for many folks the concept of using Monte Carlo is obscured by a fundamental misunderstanding of what it is. It's super flexible and extremely powerful, since it can be applied to almost any situation if the problem can be stated probabilistically. One of the most powerful techniques in any data scientist's tool belt is the Monte Carlo Simulation. In part 5, we'll apply these techniques to a business case. Part 4 attacks the problem of trying to do particle physics simulations with Monte Carlo. In part 3 we try to beat the casino in video poker. If you already know about what a Monte Carlo is and are just interested in implementing them, part 2 introduces how we do Monte Carlo in Python. This is part 1 of a several part series dedicated to investigating how Monte Carlo can be a great tool. What is a Monte Carlo Simulation (Part 1) Python