Pathing

[Rumpelstiltskin]

Just wanted to say that it's horrible.
Ships and subs get stuck near and around shorlines.
Land units get stuck in and around hills.
If two opposing groups of units clash physically, while killing each other, they will be unable to move.
No matter how frantically you are gonna click around them to make them move here or there they will usually be stuck there til one group dies.
Land units get stuck around wreckage for sooo long...
Its impossible to raid like before since raiding parties just STOP when a few front units die as they meet the defenses.
Units now tend to move in lines...
The bunch up and making them move in a more scattered fashion seems harder now.
Units sometimes bounce of buildings or enter a continuous cycle of bumping into a building than trying to move again and bumping in the building again and so forth till eternity...
Units STILL get stuck inside labs after they are built....

Maneuvering units in general has become so much more tedious and hard its almost unbearable to play if you are called upon to play a game where complex maneuvering is required.

[zerver]

I somewhat agree but pathing now is actually better than last few Spring versions. You just have to micro-manage your units a bit more so they don't get blocked by the enemy or wrecks and this I find to be good since it rewards those who really care for their troops and bring constructors to reclaim wrecks during the assault.

And wrt units moving in lines, you need to use customformations or front move. Regular move commands for a squad is going to be much less efficient for many reasons, lined up units are easy dgun targets etc.

[albator]

Why Ba is not using QTPFS ?QTPFS

[Beherith]

For why think BA not switch support what?

[smoth]

Played sc2 last night, makes spring look amazing

[albator]

For why think BA not switch support what?

albator wrote:
bugs that are still there :
- unit still cannot go through paths where you reclaim the dragon teeth, you still need to manually micro manage every single unit in order that work (meaning you better spend time optimize eco and play brutal air/nuke/...). New trick is to build DT and then to reclain them: you don't lose metal but still that has the same effect....


things you should know by now:

the default pathfinder system does not instantly update when a terrain change happens
the time it takes to update is proportional to the amount of changes happening across the map
the hole created by reclaiming some DTs can be smaller than the search resolution chosen for a path
it's pointless to keep reporting about these old bugs/design problems, there already exists a magic solution called QTPFS


Last edited by Kloot on 08 Sep 2012, 10:31, edited 1 time in total.

[Das Bruce]

You just have to micro-manage your units a bit more so they don't get blocked by the enemy or wrecks and this I find to be good since it rewards those who really care for their troops and bring constructors to reclaim wrecks during the assault.

So what if you don't want this kind of play in your game?

[Forboding Angel]

Played sc2 last night, makes spring look amazing

This. StarCraft 2's pathing is awful and frustrating.

I have pretty much no real issues in spring 91. Seems to work great. (Video Proof: http://youtu.be/MLMlCHuwZQ8 )

[Silentwings]

For me too, pathing in 91 seems to work nicely.

[Rumpelstiltskin]

Well, my experience does not go beyond BA so I cant really comment on pathing in other games.

[Kloot]

Ships and subs get stuck near and around shorlines.
Land units get stuck in and around hills.

problem: units have inertia
consequence: they might move onto impassable squares when trying to follow a path that goes near any
solution: make units obey Starcraft-style physics

If two opposing groups of units clash physically, while killing each other, they will be unable to move.

fix: allowPushingEnemyUnits=true

Units now tend to move in lines... The bunch up and making them move in a more scattered fashion seems harder now.

fix 1: ctrl+click
fix 2: enable heatmapping
fix 3: avoidMobilesOnPath=true

Units sometimes bounce of buildings or enter a continuous cycle of bumping

just a bug that won't break the universe

Units STILL get stuck inside labs after they are built

which units?

[Rumpelstiltskin]

The last one i have seen getting stuck was a Weasel, in BA.
there are others.
i'll add them in next time i see it.

@getting stuck on shorelines and hills
They always had inertia yet they didnt always get stuck like this.
all the other comments are for the exulted leader Kim Jong beherith.

EDIT
Game of BA yesterday, advanced bombers were stuck in their T2 lab after being built.
had to manually tell them to move for them to rise up fomr the lab and let others be built.
Happened several time.

Dont know if its an issue of individual units...Probably some underlying problem.
Maybe something to do with how you manage your lab rally point?

[FLOZi]

kloot, is there at all a selection of 'optimal settings' you'd recommend for path-finding, including things such as footprint sizes, heat map values etc?

Or alternately pitfalls to avoid e.g. footprint sizes that are radically different to movetype sizes

[Beherith]

I was about to ask the same thing, flozi :)

[MadRat]

TA always sent units at a 45° angle until they reached the proper x or y axial line with their destination. It really never flanked a defense like a human would.

The demo of the pathing out on YouTube seems to imply spring does try to flank. It just seems like some units would be best attacking direct, some indirect, and some perhaps away from the enemy. Too bad units cannot queue time in their path to avoid collisions, too.. Its probably a MTR item so I won't beat a dead horse.

[tzaeru]

Now that BA has QTPFS as a modoption, it really doesn't seem like a magic solution. Thing is, it's way too heavy. 80 units moving at once are enough to freeze this computer for a full second and this is quite high-end one.

I can't imagine how bad it'd be in 6v6 that went to end-game.

However, pathfinding as it is, is far from as bad as people claim it to be. In fact, compared to many commercial games, Spring's pathfinding is pretty excellent. It has flaws, that's for sure, but it's nowhere near unplayable.

[Beherith]

Tzaeru, BA 7.72 has a lot of tweaks to improve unit handling. Try them!

[tzaeru]

Tzaeru, BA 7.72 has a lot of tweaks to improve unit handling. Try them!

Yeah, actually units move much better now and particularly get less stuck to each other and wrecks.

I'm all happy with it, it's better than in most games and fully playable.

[Licho]

Why was ship beaching/units stuck on hill problem so rare in the past?

How was it different, inertia always existed, right?

[Kloot]

Game of BA yesterday, advanced bombers were stuck in their T2 lab after being built.

Mantis http://springrts.com/mantis/view.php?id=3257, completely unrelated to pathfinding.

kloot, is there at all a selection of 'optimal settings' you'd recommend for path-finding, including things such as footprint sizes, heat map values etc?

Or alternately pitfalls to avoid e.g. footprint sizes that are radically different to movetype sizes

Certain combinations of parameters are better than others, yes. I might do a writeup about them and how they interact with each other, but the constant nagging about pathing being "terribad" isn't making me run faster.

Why was ship beaching/units stuck on hill problem so rare in the past?

How was it different, inertia always existed, right?

1) turninplace=0 did not exist, units stopped for every turn between user waypoints
2) brakerate values were interpreted "liberally", units stopped _much_ faster than they were supposed to
3) bugs related to directional slope checking in more recent versions that I didn't spot until yesterday (eg. ships care about the facing of underwater slopes)

(the first two of these you could easily verify by setting useClassicGroundMoveType=true)

And just as a reminder: what unit movement was like <= 0.82.7), what it is like now (83.0 - 91.0) and where it will be going (>= 92.0).