Do I need to learn graph theory if Unity provides NavMesh AI

[MUSTAFA_RIAZ]

I’m a student learning about pathfinding in games. I know the basics of A* pathfinding and that it comes from graph theory (nodes, edges, heuristics). But Unity already provides a NavMesh system with NavMeshAgent and NavMeshObstacle.
I’m confused about a few things:

1. Do I really need to learn graph theory and algorithms (like A*) if Unity gives me NavMesh AI?
2. When is it better to use A*/grid pathfinding instead of NavMesh?

[Brian_Trotter]

It all depends on the type of game you’re going for.
NavMeshAgent is great for 3d games where a character can move to any available position. It’s actually an A* algorithm in which the nodes represent larger areas, but you don’t need to understand A* to use it. It doesn’t work well, however, with 2d games or games like Final Fantasy Tactics and XCom tile type games.

A* itself is is fantastic for tile based games whether they are 2d or 3d. It’s worth getting an understanding of the way this algorithm works.

[MUSTAFA_RIAZ]

I’m actually working on two games right now:

1. A horror game, kind of like Silent Hill mixed with Alan Wake.
2. An RTS game.
   So what do you think I should do for pathfinding in each case? Also, someone suggested that I should implement DOTS for better performance , is it really useful for these kinds of games?

[Brian_Trotter]

For both of these, assuming 3d, the Navmesh should do the trick.

In terms of DOTS, that’s a whole different programming paradigm, and if you’re using DOTS then you’ll likely need to roll an A* to handle things.