What do you think of a maze 'solvable' check

[Kyler]

Hey, all.

 

A long time player here, but first time posting.
 

I'm finally coming out of the shadows to study the code base behind OpenRCT2 and contribute myself if possible.

 

A potential feature I have in mind to implement while studying the code is to add an option where mazes can't be opened unless there's a way from the entrance to the exit.

I know the algorithm for this feature is pretty straightforward, but I'm not sure if the current codebase can handle it.

I was wondering if you could give me your thoughts on the following:

 

• Whether the suggested feature is theoretically feasible with the current implementation of the data structure for mazes in OpenRCT2.
• Whether it is possible to implement this feature as an option so that any current non-solvable mazes people have purposely made do not become "invalid" overnight.
• Whether it would be better to develop this feature as a plugin rather than directly into the source tree. (and feasibility)

 

I'm really excited to hear your input on this. Thanks for your time!

Edited March 29, 2023 by Kyler

[Broxzier]

Hi Kyler!

Sounds like a fun project to work on. It's definitely possible. When attempting to open the ride, a function can be called to verify that it's solvable, and doesn't have to be saved, so you can use any resources necessary to verify it. Have a look at how guests determine where they can walk, and use that same data to try and find a path to the exit.

I've not seen a feature request for this specifically, and there are other simpler things you could work on instead. There are some issues on GitHub labeled 'good first issue', meant to get new contributors familiar with the code. Depending on your experience and confidence, you might want to have a look at those.

[andy55699]

 

 

Edited April 4, 2023 by andy55699

[andy55699]

Wondering if you have watched Marcel Vos's video about the maze already....I bet yes? 😀

 

With the path finding algorithm in the maze, the guests would always have 75% chance to turn left because they always check and choose the path in a clockwise direction counting from the foward direction and that's why it's likely that the guests would return to the starting point of the maze in the unsolvable layout as Marcel's example of an impossible maze.

My thought is if the guests have hit the same point e.g. starting grid of the maze for more than, say, 10, times, we can assume they get lost and the path finding algorithm could change the counting direction from clockwise to anticlockwise. Then the guests would be able to solve the impossible maze created by Marcel in his video. Does this make sense?😵

[andy55699]

Well but if you watched the video RCT2 Ride Overview - Maze by Marcel Vos, in 9:00 he mentioned in 2020 summer the path finding algorithm was modified so the left indented maze would behave like the right indented maze, which I don't feel strongly that there is a big improvement because occasionally i still got guests stuck in some mazes.

[Broxzier]

Guest pathfinding is very silly in mazes, but that doesn't mean the maze is unsolvable, which is what this thread is about.