A* is not the hard part for us, since we are using a 3rd party library. Our complexity is in the scanning system that looks at the world and makes the layers of nodes linking each grid cell to its 4 neighbors plus any extra links for jumping. We have to deal with multiple grids, that move around the world and each can be partially updated. There are partial blocks and moving (door) blocks. There are links through blocks if you could stand next to the blocks, since you can destroy anything, so anything could be a path if it is reachable. Costs also have to account for double blocks, since a path goes through a low and high block. Tall buildings can have grids with 100k-200k nodes.Holy crap. I'm reading up on the A* pathfinding, and what? I'll have to read it again. I get the gist of it, but the different lists, and adding more lists to decide actually shortest path. Though based on what I've read, it goes through the path of least resistance that is adjacent, but what if another path actually becomes the shortest later on, but would only be accessible if they would have taken a different path earlier? Such as the walls, doors, etc. Is there a way to make it determine different paths then choose the shortest overall?
Paths need to be generated rapidly. About every 1-2 seconds. Imaging a zombie chasing you. Every 10 seconds and he would always be going to an old location.Finally, a real question.
The zombies recalculate the shortest path every so often, 10 seconds for 7dtd if I remember correctly. The pathing is relative, it's determined from wherever the zombie happens to be at the time, they have no memory or ability to predict shorter paths.
We do have diagonals, but that is a post process step after A* makes a path.1) I can expand on it, I've got a reasonably good understanding of it.
2) It's not really a case of how far a Z can see, it depends on the total movement cost of getting to you.
Essentially every block within a certain area has a point cost to move across it. The lighter the block (including air blocks) the fewer points it costs to move there. There's an exception for doors which cost less than they should to make doors more likely to be attacked. If the total point cost of walking around the maze is less than the total cost of breaking through the wall then they will walk around.
Hidden blocks are calculated as well.
Pretty smart for something whose prefrontal cortex is dribbling out of their nose.
Anyone interested in the new pathfinding system should watch this...
I'm just trying to keep it simple for the non code-monkeys.Paths need to be generated rapidly. About every 1-2 seconds. Imaging a zombie chasing you. Every 10 seconds and he would always be going to an old location.
Yes it is necessary to update it. Humoring you here. Imagine a big fat Dishong tower and a zombie that just spotted you halfway inside through a crack on the floor 80 blocks away. He would then just come through the stairs, rooms and obstacles in your general direction right? Nope, banging at a wall spinning around is what he would probably do. Faatal could just only use that general direction optimization thing in a flat biome with no obstacles because a flat forest doesn't have obstacles, or does it? .Voxel game mate, there is no easy way out of this and AI is a pain in the btt so big that you will need to splint your fingers everyday just to have a simple "go there and make a sandwitch" notion added to a zombie. I'm actually quite comfortable that he will achieve great things. I mean, his name literally means "God is gracious" lol. :02.47-tranquillity:I'm just trying to keep it simple for the non code-monkeys.
Honestly? I thought you would have varied the update time depending on distance from the nearest player. Pin-point accuracy isn't really necessary when the Z is far away. I've no idea if the reduction in system resource usage would be worth the extra time and code though. Always got to plan for worst case scenario which in this case would be lots of zombies very close with obstructions in the way, something like a maze perhaps?![]()
The cooler looks awesome in my opinion (regardless of where the phantom electricity is coming from).New post on Twitter:
https://twitter.com/joelhuenink/status/1049812749742555137
Wait what?9) NO. Immersion Warriors have cause to smile and nod their heads in agreement with each other and smirk at those who don't agree with them due to the recent changes.
You called? Ha - kidding.Wait what?
What's going on? So many backpack questions. Where's all this coming from?
Have we just established an 'Encumberance' system?
Oh christ..
It's a kill or be killed scenario in game. People will usually prefer to keep to themselves. Trust is a big component for the serious players.Why is it that the ONLINE community is so toxic, I tried to play online on multiple different servers and none of the servers would accept someone to comer play with them when they were new. Ive only played solo play for the past 4 years, and went to try online and no one would accept. This is why the game has no potential outsidfe of playing with friends. Trash community.![]()
wat? A gaming center? Like, an Internet Cafe?the game is amazing, A17 will be even better, Im opening a gaming center in Texas and 7 days will be one of the main games I host, but the overall community, being afraid to inv new players, is just gonna kill any chance it has to go mainstream. main concern.
Sounds inefficient if it's not done during the A* calculations?We do have diagonals, but that is a post process step after A* makes a path.
Does the backpack stick to zombies when shot at them?I would like to but....twenty questions would be so much more fun and draw it out so much longer.![]()