Unit Movement Pt. II

[Zangeeph]

Unfortunately the last topic about path finding was locked. Hopefully we can have a more civil discussion this time around. Here is a picture which illustrates the problem I was talking about.



In this image it's clear what the issue is. The path found (in bold blue) is not a good one. There is a much shorter path. Going directly towards the way point would be significantly faster. The reason it doesn't do this is because it only find paths that use angles divisible of 45, making it impossible to find the correct angle of travel. Another issue with the current system is that it's unpredictable. Ordering the samson to go directly back to where it came from returns a radically different path.



This unpredictable pathing could be the difference between life and death. How can I know which route my units are going to take? One path could be clear while the other could be straight through an enemy's base.

No matter which way you look at it, the pathing system in BA is not satisfactory. Something needs to be done about it.

[Jazcash]

No matter which way you look at it, the pathing system in BA is not satisfactory. Something needs to be done about it.

It's an engine issue, not a BA issue ¬_¬ At least try to understand what does what before you go to any sort of effort trying to show devs (who are much more knowledgeable about this than you), what the problem is.

I could pick up on dozens of points you've made over time and highlight where you're mistaken but I feel it would be a waste of time judging by your previous responses.

[smoth]

You have an issue with 8 way movement.
either submit patch or get over it.

This is not something ba has much control over, yet. Ba can do a few things by tweaking slope tolerances and other stats but little more beyond that.

I am not being mean or cruel. This is the way of things.

[Kazori]

You have an issue with 8 way movement.
either submit patch or get over it.

This is not something ba has much control over, yet. Ba can do a few things by tweaking slope tolerances and other stats but little more beyond that.

I am not being mean or cruel. This is the way of things.

lets be honest though, player base is pretty much split between BA and ZK

That means these are the games that should get priority, they are simply more important

I don't know what the devs are thinking about current pathing but just about any player will say it was better about 1+ year ago... e.g. a bunch of old BA players returned recently and each one eventually said "wtf has happened to pathing"

Cons are getting stuck ontop of solars regularly, or if you try to get one to build a spaced out wind farm it will always eventually get stuck inside and just spend rest of game walking into the side of the wind. Try to give com an order to go around lab, it will bounce off the side of the lab a few times before doing anything. Often if the terrain isnt perfectly flat and you set a waypoint out of a lab, each unit produce will bounce away from the lab in some random direction as they are created.

and there just seems to be less control over units, BA is quite competitive and there is plenty of tournaments and high-skilled 1v1s so it is very annoying when you're trying to carefully micro your units to take out an llt for example and then they just drive off in the wrong direction or do a loop for no reason or randomly get stuck inside a wreckage

I have no idea how to fix pathing and I sure as hell wouldnt trust anything Zangeeph says lol but I think it should just be returned to how it was a year ago if that's possible

[Jazcash]

The devs have been aware of these issues for some time and are working to fix them. BA really needs to stop being mentioned here.

[smoth]

Better explanation:

Spring = engine BA uses.
Spring = where the 8 way pathing is.

BA = game that uses engine.
BA = can alter unit defs. NOT PATHING.

[Kazori]

not sure if you trying to explain that to me or not, I already know all that

I was just saying that if changes were made to make benefit some random other game (I mean they're obviously not beneficial to ZK/BA) then that would just be very stupid

if not then why has it not been reverted yet

I mean Kloot might have had good intentions when starting this pathing change nonsense but it has only been detrimental, why he cant just accept failure and hit the undo button

instead the players gotta suffer

honestly its people like this that give open source projects a bad name, always some douche who messes things up for everyone

i did hear some people talking about maybe less power for engine devs in future though

[gajop]

Better explanation:

Spring = engine BA uses.
Spring = where the 8 way pathing is.

BA = game that uses engine.
BA = can alter unit defs. NOT PATHING.

Think we may want to relax the scrutiny on the problem descriptions people come to us with - these are not bug reports after all. Sure, most people outside the engine devs and those quite familiar with their work (mostly content/AI/etc. devs) won't give exact bug descriptions or ways to reproduce problems, but they can give us ideas on what the player community feels is a problem (i.e BA pathfinding) , we can deduce the rest (if it's not something mod-specific such as some sort of pathfinding configuration in a config file, it's probably an engine thing).

[Zangeeph]

No matter which way you look at it, the pathing system in BA is not satisfactory. Something needs to be done about it.

It's an engine issue, not a BA issue ¬_¬ At least try to understand what does what before you go to any sort of effort trying to show devs (who are much more knowledgeable about this than you), what the problem is.

I don't need to know any of that to notice that the algorithm they chose to implement is a poor one. An algorithm which only ever uses 8 directions is always going to be a poor one. It's not appropriate to limit directions like this.

The length of the path in pixels used by the samson is (143^2+143^2)^1/2+216=418. The optimal path is ((143+216)^2+143^2)^(1/2)=386. So the samson is traveling 8.3% further than it needs to. When the optimal solution is as simple as it could be - a straight line - it's surreal to see it take a path that's 8.3% longer.

[Google_Frog]

I doubt any work is going to go into the current pathfinder. The next pathfinder, QTPFS, does not have the 8 movement directions issue but is not done. In particular there is the unacceptable limitation that if you give 50 units a move order it can be a few seconds before all units start moving. But other than that QTPFS is looking good.

[Kazori]

I doubt any work is going to go into the current pathfinder. The next pathfinder, QTPFS, does not have the 8 movement directions issue but is not done. In particular there is the unacceptable limitation that if you give 50 units a move order it can be a few seconds before all units start moving. But other than that QTPFS is looking good.

still dont understand what was wrong with old pathfinder

why does work need to go into current path finder, why cant it just be put like it was before? O.0

[MamiyaOtaru]

among many other games, Starcraft Broodwar for example has this same limitation. And it works fine. So units are taking suboptimal paths.. both sides are subject to the same limitation so it's not like you are getting screwed.

And when a pathfinder that doesn't have this limitation is being worked on, what is even left to discuss?

[Licho]

It cannot be returned to old pathfinder because lots of code already changed..
it would be hard job reverting the change..

If you are able to do that change, please submit patch, at least ZK would gladly use engine with old pathing until better pathing appears.

[FLOZi]

It cannot be returned to old pathfinder because lots of code already changed..
it would be hard job reverting the change..

If you are able to do that change, please submit patch, at least ZK would gladly use engine with old pathing until better pathing appears.

https://github.com/spring/spring/commit ... 3621f74c53

People who don't read the git log make baby jesus cry.

[smoth]

among many other games, Starcraft Broodwar for example has this same limitation. And it works fine. So units are taking suboptimal paths.. both sides are subject to the same limitation so it's not like you are getting screwed.

And when a pathfinder that doesn't have this limitation is being worked on, what is even left to discuss?

because users love to make unreasonable demands and expect them for free.

[Google_Frog]

When is that snapshot from? 82.0 is traditionally the "good" version.

[Johannes]

among many other games, Starcraft Broodwar for example has this same limitation. And it works fine. So units are taking suboptimal paths.. both sides are subject to the same limitation so it's not like you are getting screwed.

And when a pathfinder that doesn't have this limitation is being worked on, what is even left to discuss?

because users love to make unreasonable demands and expect them for free.

Making demands is not unreasonable (you'd expect programmers to care about their userbase to some extent even if they're working for free), just the manner it was made in here was.

[smoth]

Making demands is not unreasonable, just the manner it was made in here was.

correct, making demands isn't necessarily unreasonable. Sometimes they are justified. But as you said, he wasn't reasonable. Thanks for the assist!

(you'd expect programmers to care about their userbase to some extent even if they're working for free)

So tell me, great sage, if software is flawed does that mean the programmer didn't care?

[FLOZi]

When is that snapshot from? 82.0 is traditionally the "good" version.

Everyone has a different idea of the 'good version', as evidenced by the fact that whenever a new version comes out, 5 people immediately start praising the improved pathing before 10 start to say omgwtfthisisturrible.

[smoth]

flozi, I have tank questions for you when you got the chance dawg,