[aaron695]

More interestingly , if it doesn't know meters then how does it know density, why would it pick a more dense section of triangles.

[qqqwerty]

The twitter thread isn't super clear about the exact mechanics. But If I had to guess, maybe the fleeing fox randomly chooses the path to follow, and because there are more possible paths in the dense parts of the map the odds of selecting one of those is higher. i.e. if I have 6 possible paths to my right, two to my front, and two to my left, if I randomly pick a path there is a 60% change of going right.

[aaron695]

This only works if the camps mesh is really wide, black hole style.

If you enter the camp and it gets high suddenly, nothing is changed, since as a player you know you are in a camp.

The fact the fox stays, chewing through triangles, rather than running though like it would if it was based on meters doesn't matter.

Although, since the fox stays in the camp, you would think it was telling you something, even though it was random chance.

[heavenlyblue]

As I understood that is that the fox is looking for the longest path (e.g. distance defined by breadth-first search) and longest paths are usually the ones that lead you to treasure because treasure is hidden deep within levels

[s1artibartfast]

That's funny, I understand it as the exact opposite. The fox is tasked with finding the shortest path that hits 100 points. This leads it to the closest high point density area.

[heavenlyblue]

> The fox is tasked with finding the shortest path that hits 100 points.

This is exactly what I said (breadth-first). Breadth first find shortest distances to all nodes in the neighbourhood.

[Sniffnoy]

But the whole point is that the fox doesn't know actual distance, it only knows triangle distance.

[s1artibartfast]

Yeah, I think the the path length is calculated in terms of distance, but the patching requirement isn't a set distance.

[aaron695]

> looking for the longest path

> "The Fox isn't trying to get 100 meters away - it's trying to get 100 triangles away."

Breadth-first search wouldn't be possible. If at each triangle the fox has ~2 options that's 2^100 to get to 100 triangles away

But if it did impossibly work out every journey, and then chose a random last node it would end up in camps more often.

Depth-first search shouldn't preference the camps, unless it gets dragged in from afar. But if it can only get dragged in really close to/in a camp, then it's about the same probability as the fox running past/through a camp anyway. And I would assume as a player you would see the fact it's pretty.

[heavenlyblue]

> If at each triangle the fox has ~2 options that's 2^100 to get to 100 triangles away

I don’t think you understand what Djikstra’s algorithm is. Breadth first doesn’t mean you have to always re-visit the same node more than once.

[thargor90]

it doesn't know density. density is a property of the triangle placement. More triangles in "dense" areas. Pick a random triangle, you are likely to fall into a dense area.