75 replies to this topic

[dreddnott]

In the tactic files, BeginHoldState will do a move (not attackmove) if further than 60 away from the destination. This very often leads to units moving to hold a position, and actually being in range of the enemy units they're supposed to be fighting, but just running past them while getting gunned down, instead of just fighting them immediatedly. A check to see if enemies in range are in range of the destination, and if they are then do an attack move instead, might be the solution.
This issue becomes even worse when you take into account that the logic deciding on chokepoints to defend and such continously flips back and forth, so the units always finds themselves moving back and forth between points they'll never get close to.

Perhaps this is a silly question, but why not change the offending DoMove from HoldState to AttackState?

[Smokeskin]

I don't really know how AttackState differs from HoldState and I haven't tested that exactly. I think you'd still get problems from the indecision in the code, where it switches what point to defend repeatedly. AttackState would also have to distinguish between situations where the troops are locally in trouble and should move to the designated point because there are real problems there (something unnatural for an attackstate, they'd just not attack), situations where the problems at the destionation are so grim they should just drop it altogether, and situations where they have local superiority and should do attack moves. In the end, I think these decisions can't be made at unit level.

I'm thinking along the same lines though, but at the Strategy/Plan level. If someone could tell me where the code that selects what Plan gets carried out and what units are assigned to it, then that would be the right way to fix it. If your units are mixed in with or near enemy units, you don't use Defend and DefendChokepoint on points near the enemy unless you really mean it. Either you do it to attack and repel or kill the enemy because you can, or you go to attack one of his undefended points or just move to Hold some uncontested area. Get rid of the indecisions and only do Hold if you can actually hold it. I don't think the HoldState code is wrong, it is just being used at inappropriate times where you shouldn't try to hold the position.

So far I haven't been able to figure out where the code selects what Plan to carry out though, it would be great if someone could point me in the right direction

[ArkhanTheBlack]

The attack strategy creates and controls the attack plans and the defend strategy creates the defend plans and the defend choke point plan.

The problem of switching choke points is correct. The harassing code had a similar problem with constantly switching target positions. That's why I changed it. I think the attack and defend plans have specific control mechanisms to avoid the fast switching, but I agree, the defend strategy should have a switching delay for the defend choke point plans. That could solve at least 'some' problems. But just like you said, the problem of the movement code to get 10 new problems for every solved problem makes changes in this area a real nightmare, especially the tactics. However, the switching delay for the choke points should be easy to do without the risk of unpleasant surprises.

Edited by ArkhanTheBlack, 29 April 2008 - 03:43 PM.

[RobertW7928]

hmm yes, I have noticed that the Ai is somewhat lacking & all other the place, compared to Dark Crusade some good mostly bad, the Imperial Guard and Necrons are in good working order not many problems there, I wish i can say the same for Space Marine Allies and Eldar/Dark Eldar, any-who I believe you guys will sort something out, best of luck and hopefully a successful one at that. funny thought you guys try to get the sisters servitor to use dismantle on enemy buildings? they would be pissed lol. another thing in assassinate the Ai does some pretty stupid things, like running the Commander back and forth from ally base to other being attacked in the process and eventually killed an cant you do something that will encourage them to level up if hero's is enabled? you know to make them a tougher adversary, what about if there base is attacked, when there forces are attacking yours or others somewhere else I didn't see any of them return to defend there base, should they?

Edited by RobertW7928, 01 May 2008 - 09:23 PM.

[Smokeskin]

The attack strategy creates and controls the attack plans and the defend strategy creates the defend plans and the defend choke point plan.

The problem of switching choke points is correct. The harassing code had a similar problem with constantly switching target positions. That's why I changed it. I think the attack and defend plans have specific control mechanisms to avoid the fast switching, but I agree, the defend strategy should have a switching delay for the defend choke point plans. That could solve at least 'some' problems. But just like you said, the problem of the movement code to get 10 new problems for every solved problem makes changes in this area a real nightmare, especially the tactics. However, the switching delay for the choke points should be easy to do without the risk of unpleasant surprises.

A delay won't solve it. Moving back and forth between points at 10 sec intervals would be just as bad.

As I can see it, you don't have code in there that checks for the local situation, is my army able to beat the enemy where it is now? If yes, do an attack there. If not, you can either move to attack an undefended LP of the enemy, or move back to take up a defensive (waiting) position (preferably at an LP2), and that position should be mostly enemy free.

I could do this by just messing with your code, so Defendchokepointplan would also launch attacks and such. But to keep things right, I would much rather like to know, how (or where) is it decided what plan gets carried out? If I wanted to put in some checks to decide what plan to go for, where in the code should I be looking?

[ThetaOrion]

In the last game I played, the Necron AI ally did some weird things on Mountain Trail. After helping me stay alive, the Necron AI ally went up into the central critical point area with their whole army, and just sat there with their Restored Monolith and army and did nothing, until the Tier 4 Chaos AI enemy came and pushed them out of the area. Then the Necrons defended their homebase area and wiped out the Chaos invaders, and then the Necron AI ally went and took out the Chaos AI enemy on their side of the map. Then after taking out the Chaos AI enemy, the whole Necron AI ally army just sat there on their side of the map, while I as the IG and my Baneblade finished off the Necron AI enemy on my side of the map alone. Once again, when it was down to the last listening post or the upper north corner there of the AI enemy's base, then the Necron AI ally all moved together across the river and finished off the enemy. But, there's no reason that they shouldn't have been able to come sooner. That's the third time, now, that I have seen that phenomenon -- the first time was with the Eldar AI ally when they finished off the Chaos AI enemy on their side of the map and then just sat there in their Brownian motion soup doing nothing.

It's interesting to see that some of you guys might have some idea as to what is happening and what might be the solution. Every now and then I get into the situation where the AI ally just freezes into place with their whole army at Tier 4 and just do nothing except pace or adjust -- it's like they are stuck in some sort of infinite loop. It certainly doesn't happen every time, but enough to have been noticed.

Edited by ThetaOrion, 06 May 2008 - 01:43 AM.

[RobertW7928]

heh I've not long read bug report on a patch in progress for the game & I tell ya I'll be glad when they fix the many mistakes/problems this game has suffered, Ai Mod Forever =]

[thudo]

Found an issue with a tactic file..

This line in the Dark Eldar infantrytactic.ai file looks like:

-- I might have these attached
	if (self.squad_ai:IsAttached()) then

		if (self.squad_ai:HasSquadAttached("dark_eldar_squad_haemonculus")) then
			HaemonculusTactic.InitAbilities( self )
			HaemonculusTactic.DoAbilities( self )

But it should look like:

-- I might have these attached
	if (self.squad_ai:IsAttached()) then

		if (self.squad_ai:HasSquadAttached("dark_eldar_squad_haemonculus")) then
			DarkEldarHaemonculusTactic.InitAbilities( self )
			DarkEldarHaemonculusTactic.DoAbilities( self )

I'm so shocked this hasn't come forward with no one making a comment where there game caused a Fatal Error when the Haemonculus tries to attach to a squad. :s

Btw, we really need to look at that Builder jumping code when builders jump to buildpos when LPs need to get constructed and instead they warp to it but instead of building then walk back towards the HQ startpoint. :( This only happens with builders that can teleport or jump.

[Zenoth]

I'm so shocked this hasn't come forward with no one making a comment where there game caused a Fatal Error when the Haemonculus tries to attach to a squad. :s

I cannot speak for others but I do know that in my countless tests and in mere games I've played for fun with the Dark Eldar, many of those had four DE teams in them, I've never ever encountered any Fatal Errors about them, and more importantly I've never noticed the Haemonculus ever attaching himself to a single squad, so it should explain why no one has never seen that happen (a Fatal Error) since the condition for its trigger are not met. I myself am actually surprised that you guys had put the code in there to allow him to attach, because as I said I never saw him attached at a squad, that guy is always alone in my games.

On a side note I came back to read some posts around here and I'm very surprised that there's still no news about the first patch. I don't know what Relic's doing but at this rate StarCraft 2 will be released before it, it's a little ridiculous if you ask me, and the Advanced A.I fans around are waiting as much as us for that patch so all of the bugs found in this thread can find their way in a newer build. I find it a little depressing depending on a patch on which we don't have news concerning its status (anyone has any news?).

[thudo]

I'm surprised noone has experienced the Haemonculus attach bug or am I missing the script logic because its not right. The Tactic call isn't the same as the Haemonculus own tactic file itself thus, if the Haemonculus tried to attach to a squad it would cause an error (not a crash however). I wonder then if the Haemonculus attach code is broken as the Haemonculus then never attaches?

[Zenoth]

I'm surprised noone has experienced the Haemonculus attach bug or am I missing the script logic because its not right. The Tactic call isn't the same as the Haemonculus own tactic file itself thus, if the Haemonculus tried to attach to a squad it would cause an error (not a crash however). I wonder then if the Haemonculus attach code is broken as the Haemonculus then never attaches?

Well I cannot say for sure if the code is broken since I know nothing about coding. But as a tester and as a player I can assure you that I've never seen it attaching himself a single time since I've started testing this mod for Soulstorm, and believe me I've played a lot of games, both for testing purposes and just for fun since the latest build was released. By your post I am the one who's surprised indeed that "normally" he should be able to attach, so, perhaps yes logically that'd mean that it might be broken. It's ironically a good thing, since I never got a single Fatal Error with the Dark Eldars, or any other races for that matter (the only major bug I've seen was related to Heroes not working early on in the builds and the infamous Necron CTDs, but CTDs don't qualify as Fatal Errors).

[thudo]

Hmmm then maybe we should look at the Haemonculus attach code to see what gives. If you haven't ever seen the Haemonculus attach to a squad thats a prob

[ThetaOrion]

Nope, I have always seen the Hommonculus alone, never attached. Not ever seen the AI make and attach Apothecary and Mad Dok either!

And, when I see other SM Commanders and Eldar Farseer, they are always unattached.

The attaching thing might not be working well in 3.0.

--

I have only had one crash with SoulStorm, and that was while running the AI Mod 3.0 and the randomizer had given me a thing of all Orks, 2vs2, 4 factions of Orks. I had bunkered Gretchin in a listening post, and when I detached them or unbunkered them, then it crashed on me.

That's the only SoulStorm crash I have had so far, and I have played SoulStorm Campaign for a month and this AI Skirmish Mod a whole bunch, and had only one crash. Not bad!

I have been very pleased with the stability of the thing compared to some of the other junque that other game places have released this past year for their first release.

Edited by ThetaOrion, 27 May 2008 - 11:35 PM.

[mindjuicer]

Firstly, a big thank you to the Dawn of Skirmish team for turning the limited single player games into something great. You guys should be employed by Relic.

I've unofficially collected the fixes and suggested improvements from this thread into one archive. These are:

• thudo Haemonculus fix - darkeldarinfantrytactic.ai
  http://forums.revora...&...st&p=597132
• thudo scout cost fix - marinestrategyinfo.ai
  http://forums.revora...&...st&p=587766
• Larkin dance improvement (2nd one) - tactic.ai
  http://forums.revora...&...st&p=586633
• dreddnott Fleet of Foot improvement - eldarinfantrytactic.ai
  http://forums.revora...&...st&p=587174

You can download the file here. Just skip the 2 ads.

You must extract the file into your Dawn of Skirmish (Dowai_mod) folder. It will ask you if you want to overwrite your files. If you don't want to use any particular "fix", simply don't overwrite that file.

Don't blame the team if anything doesn't work. Use at your own risk.

[thudo]

Wow thanks.. only little prob with this is each player the user connects to via online/LAN has to ensure they have all these OR have the same code per connected machine otherwise it will not sync or fail after starting a game. So just be careful.

[mindjuicer]

Glad you appreciate it. The dance bug is a bit of a game-killer for me and I almost didn't notice Larkin's code on the first page. It seems to do the trick.

Off to get my arse kicked by your AI.