I started playing this lovely game called Sledding Game Demo with a few of my friends, and after racing each other on sleds for a few days, one of my friends found a fishing minigame.

There were rare fish and all that and there were traits and fun things. As I am a bit of a data scientist I wanted to know what the fish drop rates were and for that I needed a dataset. I didn’t want to make one entirely by hand so I made a script. The original motivation was actually the sheer mountains of money you get from catching big fish and the joy of gambling it all away at the roulette table in the lodge. This was my first time using OCR and I used Google’s Tesseract. The script uses the amount of red on the screen as a signal for when to reel. Because the bobber is red on an otherwise snowy map you can detect when it disappears under the water’s surface this way.

A full writeup and code will be available in the future.