Pathfinding: What are the chances...
Moderator: Moderators
- Forboding Angel
- Evolution RTS Developer
- Posts: 14673
- Joined: 17 Nov 2005, 02:43
- bobthedinosaur
- Blood & Steel Developer
- Posts: 2700
- Joined: 25 Aug 2004, 13:31
Re: Pathfinding: What are the chances...
You son of a...
Re: Pathfinding: What are the chances...
Low to none.
Re: Pathfinding: What are the chances...
it wouldn't be that hard to do, but it needs a dedicated developer, probably with some knowhow.
Re: Pathfinding: What are the chances...
It looks cool.
I haven't played supcom enough to know how good it is,but It might have all sorts of restrictions..
I haven't played supcom enough to know how good it is,but It might have all sorts of restrictions..
Re: Pathfinding: What are the chances...
the thing is, they are using fixed distance between units formations.
we do not have that in spring, and that allows for more variety (eg, it is not purely a bad thing). if you leave some space between units in formations, it is of course much easier to move two formations through each other, or move them through narrow passes (and it looking nice). it also requires units/models to be generally smaller though, compared to the map, to get the same gameplay.
is this right?
we do not have that in spring, and that allows for more variety (eg, it is not purely a bad thing). if you leave some space between units in formations, it is of course much easier to move two formations through each other, or move them through narrow passes (and it looking nice). it also requires units/models to be generally smaller though, compared to the map, to get the same gameplay.
is this right?
Re: Pathfinding: What are the chances...
we dont need flow crap or whatever it is...
We just need to sort what we have a bit..
I mean why is it than when i have a line of units facing right and i stretch a move line at a location to the left of my units all my units will get criss crossed...
a\ / a
b \ b
c/ \ c
a form right will go to c from left and c from right will go to a from left...I mean WTF...
We just need to sort what we have a bit..
I mean why is it than when i have a line of units facing right and i stretch a move line at a location to the left of my units all my units will get criss crossed...
a\ / a
b \ b
c/ \ c
a form right will go to c from left and c from right will go to a from left...I mean WTF...
Re: Pathfinding: What are the chances...
This happens randomly. Also, I would imagine this is a problem with Custom Formations rather than with general pathfinding. I've expressed this before in the form of pretty diagrams but nothing got done about it.Gota wrote:we dont need flow crap or whatever it is...
We just need to sort what we have a bit..
I mean why is it than when i have a line of units facing right and i stretch a move line at a location to the left of my units all my units will get criss crossed...
a\ / a
b \ b
c/ \ c
a form right will go to c from left and c from right will go to a from left...I mean WTF...
Re: Pathfinding: What are the chances...
It's the same If you remove custom formations.
Re: Pathfinding: What are the chances...
i remember loosely that the problem was the sorting algorithm taking too much CPU time if you have a lot of units in the formation, therefore, an upper limit for the perfect sorting/arrangement of units is used. maybe you should also remember the why sometimes, when nothing is done about your requests.
Re: Pathfinding: What are the chances...
2 units isnt a lot imo.hoijui wrote:i remember loosely that the problem was the sorting algorithm taking too much CPU time if you have a lot of units in the formation, therefore, an upper limit for the perfect sorting/arrangement of units is used. maybe you should also remember the why sometimes, when nothing is done about your requests.
Re: Pathfinding: What are the chances...
Spring pathing is borderline broken.
Can I set the weight the pathfinder gives to blocking units yet?
Can I set the weight the pathfinder gives to blocking units yet?
Re: Pathfinding: What are the chances...
I'm sure Custom Formations is relevant in some way though. It orders each unit a move command individually, so each unit has it's set move order to go to, each unit should have the most optimal path in terms of time it would take to get there, compared to any of the other units using that same path.Gota wrote:It's the same If you remove custom formations.
Re: Pathfinding: What are the chances...
Why not just make a lot of dust and cheat on collission detection, while to friendly formations cross each other?
Re: Pathfinding: What are the chances...
include this as a add. function for the MT-spring exe?
Re: Pathfinding: What are the chances...
You dont need many units to get this result...hoijui wrote:i remember loosely that the problem was the sorting algorithm taking too much CPU time if you have a lot of units in the formation, therefore, an upper limit for the perfect sorting/arrangement of units is used. maybe you should also remember the why sometimes, when nothing is done about your requests.
just a few..they will criss cross and cross each other's paths for no apparent reason,when they could just walk straight.
Take 5 units and test...
If 5 units are too much already than we definitely need some new system.
Re: Pathfinding: What are the chances...
This is a synced game. MT cannot work differently than ST.manolo_ wrote:include this as a add. function for the MT-spring exe?
I agree the pathfinder has some issues. We need a flow field expert :)
- Forboding Angel
- Evolution RTS Developer
- Posts: 14673
- Joined: 17 Nov 2005, 02:43
Re: Pathfinding: What are the chances...
Also, watch the damn video if you haven't, BEFORE replying here. A few people didn't and it shows and is royally irritating.
Re: Pathfinding: What are the chances...
what if you break a formation down into lines- like ants stupidly following the headleader- and if the headleader changes direction because of obstacles, all the tool has to calculate is that no leaders path crosses each other - would look mighty intelligent, and minimize calculation for every single unit.. - sort of snakes passing by each other..
also royally annoying forb were:
********
******
*********
********* ********
If your name is not on the list, consider yourself awarded.
PS: I also heard that SoupCom has >>ProblemsofItsOwn-thatarewaymoarImportanttosolve<<- can spring have >>ProblemsofItsOwn-thatarewaymoarImportanttosolve<< too?
also royally annoying forb were:
********
******
*********
********* ********
If your name is not on the list, consider yourself awarded.
PS: I also heard that SoupCom has >>ProblemsofItsOwn-thatarewaymoarImportanttosolve<<- can spring have >>ProblemsofItsOwn-thatarewaymoarImportanttosolve<< too?
Re: Pathfinding: What are the chances...
We need to show Chris that Spring pwns supcom so we need to employ ├╝ber algorithms. Maybe hueg neural network. A pathfinding that sucks at first, but eventually outperforms the flowfield