Tekken 8 Data Tools

Basic data gathering and analysis tool for Tekken 8 replays.

---

Getting Started

Requirements

1. Any modern version of Python

   • Exact minimum version TBD - Python 3.14 will NOT work

2. CPU that supports AVX

   • Most CPUs released after 2011 will support AVX

3. The UV package and project manager

Installation

1. Clone the repo

git clone https://github.com/chiefeelface/Tekken-8-Data-Tools.git
cd Tekken-8-Data-Tools

2. Install dependencies and activate the virtual environment

uv sync
.\.venv\Scripts\activate

Usage

Simply run the main file and follow the prompts

py main.py

If your terminal does not have a (.venv) in front of the prompt, you will need to activate the virtual environment before running the file

.\.venv\Scripts\activate

If downloading replays, it will download 700 seconds worth of replays every 1 second as per the Wavu Wank wiki api, so downloading a month of replays will take about an hour.

Saving replays to an SQLite database is only recommended if you are going to write your own queries, as the analysis is much slower (10-15x) then when using a CSV file. If you choose to save the replays to a SQLite database file there are lookup tables you can join on to get the names of characters, stages, etc. to make the data readable.

The replays will be saved intermittently once the total replays downloaded reaches 1,000,000 to the downloaded_replays directory, with numerous fail-safes to prevent any downloaded replays from being lost in the event of network failure or any other errors.

If you choose to save the analysis to an excel file, it will be in the results directory.

Analysis Ideas

• Ping Tekken 8 API to get highest ranked character by player.
• Get most played character win rate and win rate without most played character by player.
• Get average win rate by rank.
• Get average win rate by character by rank.