What are some strategies to automate the maze?

[ConsumedChemicals]

I have been stuck for so long on thinking about it. if you have a pathing system, what was your thought process or idea?

[Lostrick]

imagine putting your right hand to the right wall, keep following that wall and you'll find the treasure since all the wall are connected

[umop-apisdn]

Originally posted by Lostrick:

imagine putting your right hand to the right wall, keep following that wall and you'll find the treasure since all the wall are connected

... not necessarily true, if you fertilize the chest when you get to it. That is to say, only the first maze in a given hedge works like that.

[samite_alchemist]

I like Koenich's guide.

[michael_kuerbis]

There are plenty of guides and discussions here around advanced maze solving algorithms such as Trémaux.

But let's have a look at the thought process behind the simplest method: the already mentioned right-hand (or left-hand) system.

Keep an (imaginary) hand on the right (or left, makes no difference) wall, and follow that wall. If you want to see it in action, look for some footage of the 3D Maze screensaver from Windows 95/98.

While it sounds simple and is simple to imagine in your head, the actual implementation for TFWR requires a little more thought.

Now if you were in some other game with a regular first or third person camera, this would be easy to follow: if you can go right, go right. Else if you can move forward, move forward. Else if you can go left, go left. Else turn around and go back.

But forward, left, right and backward are all relative directions. For our drone in TFWR, we need to work with absolute directions: North/East/South/West. The drone has no notion of rotation either. What this means is that we have to store this piece of information in a variable: in which direction are we going? What was the last direction in which we were moving?

Knowing that, let's assume we were last going North. With our hand to the right, we have to try going East now. If that is not possible, we have to go North again - forward. If North is not possible either and we have a wall in front of us, we must go West - left. If West is also unavailable, then there is no way to go but backwards - South. Note that the direction we finally picked becomes the new direction in which we are currently going.
The cases for the other directions are analogous.

A benefit besides the simplicity is the minimum of data required: you do not need to memorize any specifics of the labyrinth. All you have to know is the current direction in which you are going.

The key downside is, of course, that it stops working if the maze contains a loop. As mentioned by others already, the initial version of a maze in TFWR is always loop-free. But if you fertilize the treasure instead of harvesting it, then a random wall can get removed, the maze is no longer guaranteed loop-free and you will need a different algorithm.

[ConsumedChemicals]

Originally posted by michael_kuerbis:

There are plenty of guides and discussions here around advanced maze solving algorithms such as Trémaux.

But let's have a look at the thought process behind the simplest method: the already mentioned right-hand (or left-hand) system.

Keep an (imaginary) hand on the right (or left, makes no difference) wall, and follow that wall. If you want to see it in action, look for some footage of the 3D Maze screensaver from Windows 95/98.

While it sounds simple and is simple to imagine in your head, the actual implementation for TFWR requires a little more thought.

Now if you were in some other game with a regular first or third person camera, this would be easy to follow: if you can go right, go right. Else if you can move forward, move forward. Else if you can go left, go left. Else turn around and go back.

But forward, left, right and backward are all relative directions. For our drone in TFWR, we need to work with absolute directions: North/East/South/West. The drone has no notion of rotation either. What this means is that we have to store this piece of information in a variable: in which direction are we going? What was the last direction in which we were moving?

Knowing that, let's assume we were last going North. With our hand to the right, we have to try going East now. If that is not possible, we have to go North again - forward. If North is not possible either and we have a wall in front of us, we must go West - left. If West is also unavailable, then there is no way to go but backwards - South. Note that the direction we finally picked becomes the new direction in which we are currently going.
The cases for the other directions are analogous.

A benefit besides the simplicity is the minimum of data required: you do not need to memorize any specifics of the labyrinth. All you have to know is the current direction in which you are going.

The key downside is, of course, that it stops working if the maze contains a loop. As mentioned by others already, the initial version of a maze in TFWR is always loop-free. But if you fertilize the treasure instead of harvesting it, then a random wall can get removed, the maze is no longer guaranteed loop-free and you will need a different algorithm.

Thank you!!