UFC Fighter Career As Coin Flips Part 2
Overview:
In expanding on the idea of coin flipping, I've tried to bring a touch more realism to the mix. Although I acknowledge that this synthesized data, built on simplistic assumptions, does not perfectly mirror the intricacies of actual UFC data, there's a certain appeal in starting with a straightforward concept and iteratively refining it.
Moreover, it occurred to me that this method could provide a simple solution for generating fake data for sports analytics projects, or even a fantasy sports game.
Let's take a look at the assumptions I incorporated:
Each fighter can have a career spanning up to 13 fights.
If a fighter suffers three losses, their career is effectively over.
Each fighter is assigned a win probability randomly picked between 40% and 80%.
Results
The distribution of win-loss records, given these assumptions, is as follows for a simulation of 10,000 careers:
Next, let's dive into the distribution of win percentages:
To add an extra layer of realism, I've used Python Faker library to generate names, leading to the following results:
If you’d like to delve deeper into the code used for this posting, it's available at: https://github.com/JKolodny/MMA-Simple-Fight-Simulations
Here's to making simulated data more realistic, one iteration at a time!