23 replies to this topic

[thudo]

Up for the testers! We're getting REALLY close now to a release candidate!

- IG hellhound adjustment
- Reduced assault marine/raptor rating
- Fixed capturing bug
- Improved anti stealth code
- AI won't build behind the enemy anymore
- Added Ork Trukk before Kan
- Cultists switch to ranged in tier 3+
- Tau commander got emergency teleport from Necron Lord
- (Plus the other stuff I already said in the forum)

[LarkinVB]

Great job. Lots of stuff I was hoping for is included.

[thudo]

Larkin - Did you want Trakks or Trukks before the KillKan? I thought it was Trakks as the Trukk isn't really great compared to the Trakk. Let us know.

Also.. there is a slight issue with the 4th IG buildprogram. Sending Arkan correct files.

[LarkinVB]

Well, pros usually do use trukks whereas I rarely see trakks. I don't know the difference in needed req and banners but since the AI is not good in doing transports I think trakks are fine due to their disruptive powers. A second oppinion is welcomed. I'm no real player but watched lots of dowsanc replays and tested AI vs AI.

[ArkhanTheBlack]

I added Trukks since they are anti infantry. Trakks are anti vehicle which makes not much sense to me for an early T2 encounter.

[LarkinVB]

Just tested a 2 vs 2 at mountain trail. Tau at position 3 did only build 2 generators at gamestart and no barracks at all for some minutes. It had choosen strategy 3. This was the same problem as in a game from inquisitor I watched some days ago. I restarted, tau did choose strategy 2 this time and they immediatly did build their barracks. I didn't check their strategies but it looked suspicious.

Same game. IG CS was build and tried to harass but the general was alone and did not stand a chance vs chaos lord. Meanwhile IG did build a single commissar. I think CS should try to reinforce a priest as soon as possible, that is as soon as the CS is ready for action.

[LarkinVB]

Raptors were jump attacking IG vehicles. Not clever.

Are gretchins now infiltrated faster ? Whatever, it did really help orks at gamestart.

Otherwise : Wowww !

The game went on for 1h and it was a stalemate when I did quit. Ork + Chaos vs Tau and IG. Slaughter and mayhem and it was like watching a very good movie. Excellent job Arkhan ! I lost track of writing down stuff as it was most fascinating to enjoy the show. We had all races T4 with baneblade, gnarloc, bloodthirster, demon lord, squiggy and most of them fighting each other multiple times. As soon as one of them was in danger to get anihilated his ally came to rescue him. rinse and repeat, back and forth. I think this will be a great release.

[ThetaOrion]

Meanwhile IG did build a single commissar. I think CS should try to reinforce a priest as soon as possible, that is as soon as the CS is ready for action.

--

1) It is essential for the IG to build all three Commissars and attach each one to a guardsmen squad, if 1.2 still allows the building of all three Commissars.

2) If you are going to have an IG Command Squad, a Priest and a Commissar reenforced seem to be helpful, although I always seemed to lose my Command Squad in 1.1, so after awhile I just stopped building it as it was a waste of resources. Three Guardsmen Squads with three Commissars attached was a better value.

3) You need to make sure that attached AI Commissars can use the Execute command, because my IG games changed when I started using the Execute command. I started winning instead of always losing.

4) Later, it is actually very useful to build a Priest and attach that to Kasrkin. Add the plasma upgrade, and those guys are often taking on tanks successfully. Although the potency of this combo does seem to change a lot from one Relic version to another.

5) Looks like from Larkin's input that you need to look at Tau Build Program #3 and figure out what is wrong with it.

6) Make sure that the IG AI Priest can use the fanatacism ability, as that fanaticism thing seemed to make the attached IG squad temporarily indestructible. In fact, I have often thought that it would be nice if you could give that IG Priest's fanatacism ability an area effect, so that when the Priest casts the spell, nearby IG squads benefit from it as well as the attached squad. Hope I said that right. But, an area effect for fanatacism might not be under AI control but might be a rebalance thing and thus something for DoWXP? I don't know. I do know, though, that you should at least make sure that the AI IG Priest can and does cast fanatacism and does attach up to an IG squad at its earliest convenience. Once again, it can make a big difference, so 'dat 'dem humies aren't 'da disgrace that Thud has been describing them as being.

And, from the occasional lurking I have done, it does indeed look like this is going to be a great release!

Edited by ThetaOrion, 28 April 2007 - 12:17 AM.

[Maktaka]

Only one Commie is necessary now, so the AI is doing good. Since Execute isn't broken and effecting plasma/GL anymore, it's nowhere near as useful and by extension neither is the Commissar. The regen bonus is mostly useful for melee, something GM should avoid anyway since they suck at it. Get one Commie for T2 to keep morale at full, then just use Priests from there on out. In T3 the Priest for GM should be moved over to the Kasrkin and Ogryn and replaced with a Commie. Heck, you can do quite well with just Priests and the upgrades as far as morale is concerned.

Fanatacism is already awesome, 10 seconds of 10 invulnerable Kasrkin is incredible. With just the Hellguns alone the squad will dish out about 2400 points of damage to infantry heavy medium during that time and cannot be stopped. You usually only have one or two squads under fire at a time anyways.

I look forward to your work guys, I've resorted to playing other people (and all their idiocy) online until your next release. At least your AI knows how to build power generators.

[dctrjons]

Just to add my 2cents about IG vs. necrons. (Tier 1 - Tier 2)
I'm not sure if this is cheese not conducive to good AI or not.

-I go for controlling <50% of the listening posts as quickly as possible. Pumping out squads just for capping, and dancing to keep them alive and slow the necrons.
-Barracks over armory for the squad cap bonus and requirement for turrets.
-Enough gens to keep turret production up and tier 1 LP upgrades.
-Problem I have with commisars is execute is tier 2.
-I sometimes will create another base before tier 2 depending on how may LP's I have capped still. For turrets and more gens if I forsee a long fight.
-Dance a few builders on repair duty for LP's and turrets.

I keep roving squads that try and keep LPs out of scarab hands and slow

Even if I make some mistakes and after a while I am in control of less than half the CPs it usually was enough to get me a good resource base and has kept their squad cap down enough to do minimal damage.

Then I move the armor out.

I don't know you're new AI, but my prefered strategy is zerging the LP's and turtling. And keep the LP's out of their hands as long as possible as cost effectively as possible. Spliting their army between attacking the base and destroying my LP's and their scarabs split between building gens and capping.

I somewhat feel this is cheese since I would not do this against any other race, only because of their weak/slow capping abilities.

Edited by dctrjons, 28 April 2007 - 02:44 AM.

[thudo]

Okie so what I got is:

1) Orks - What should be push out for Orks first? Trakks (mostly anti-vehicle) or Trukks (purely anti-infantry transport)? KillKans will be kept "as is".

2) Orks - Should we force the Grot Infiltrate to all buildprograms? Right now its set as a dynamic build for Tier1 so when 2 Grot squads are present and the AI has 50req it should build it. Not always but a forced build would always build it.

3) Tau - BuildProgram #3 is set to build the Barracks after the 2nd gen. This was done for a reason but should this be changed? Its the only buildprogram where the Barracks is later than the others.

4) IG - should we force out a Priest from the CS immediately? I have to check what leaders are available to the CS at gamestart.

5) IG - should all build programs force out Commissars so Guardsmen can use them to survive early tier1?

6) Are we concerned about having any attachable units accidentally attach to infiltrating squads? (ie. Apothacary attaching to Scouts with Infiltration researched)?

7) Should we have a check so that Jumpable Assault squads won't jump into CC vehicles? AI can't pull this off effectively and chop down vehicles in melee?

[ThetaOrion]

Thud!

To me it looks like a great list of things to check and look into before the release candidate.

--

IG - should all build programs force out Commissars so Guardsmen can use them to survive early tier1?

--

I do agree with Maktaka that the IG Priest can do as much good attached to a squad as a Commissar can do.

But, for Tier 1, for the opening three Guardsmen Squads, the Commissar is better simply because you can make three of those Commissars immediately after producing your three Tech Enginseers, without having having to upgrade the Command Center or do any time-consuming research, or at least in 1.1 that's the way it worked.

I actually went for Three Commissars attached to three guardsmen squads before going after the Command Squad, in the IG games I was winning, because attached Commissars was a better value. I would lose any of the extra Guardsmen squads that had nothing attached, but the three Guardsmen squads with the attached Commissar were all still standing after the initial enemy rush, although, I would sometimes lose one of the Commissars. The NW and other enemies shoot the Commissar, and he takes the damage, and the rest of the squad goes unhurt most of the time -- put the whole thing on overwatch, and use Execute whenever it comes available, and my Guardsmen/Commissars were taking on Necron Warriors with no problem in 1.1. That's the way I remember it. And, when one or two of your attached Commissars are dead, research Priest and replace them with Priests.

Maybe only try the Commissar experiment in half of the build programs? In the infantry build, whatever it is called, skip the Command Squad or put it further down the list, and go for all three attached Commissars followed up by the Priest Upgrade, and then start building and attaching Priests. That's basically what I always did whenever I won as the IG in 1.1. Whenever I built Commissars and attached them, it just always seemed to get me through Tier 1 just fine, and it also increased the value of each one of those Guardsmen reenforcements if you put that attached Guardsmen squad on overwatch, and of course gave you plenty of Guardsmen to Execute your way to victory. I just skipped the Command Squad, and if I built it, I built CS in Tier 3 when I had funds I didn't know what to do with. You could to the same trick with one or two of the IG build programs, to get you past Tier 1. I wouldn't do it in the Build Program where the tank rush is the primary goal, but some of the other Build Programs would probably benefit greatly from the trick, as long as 1.2 is pretty much the same as 1.1 where the Commissars and Guardsmen are concerned.

--

I even sometimes go on to build Psychers and attach them. I wasn't as pleased with the Psychers, but they do give you the option of having every IG squad that you build to have something command level attached to them.

It made a much bigger difference than I thought it would, the difference between winning and losing, in 1.1. I hope and assume that it's the same with 1.2 as well.

Summary:

For Tier 1 in 1.1, you didn't have to do any research at all to pump out Commissars and attach them up. They are just waiting to be made, and they proved to be of more value to me attached to Guardsmen than the expensive Command Squad that always died on me in CC with a Necron Lord or NW.

I found it to be a nice little trick for Tier 1 and possibly the answer or the solution that you were searching for?

Edited by ThetaOrion, 28 April 2007 - 07:42 AM.

[LarkinVB]

3) Tau program 3 will build barracks only after the 3min mark which is certainly not what was intented. Perhaps the reason is the 600 army requirement which it can only satisfy with stealth suits.

Seems eldars FOF sometimes is a bit late (research, not usage). I just watched a game were they had FOF ready after the 4min mark.

Edited by LarkinVB, 28 April 2007 - 07:42 AM.

[ArkhanTheBlack]

3) Tau program 3 will build barracks only after the 3min mark which is certainly not what was intented. Perhaps the reason is the 600 army requirement which it can only satisfy with stealth suits.

Yes, I checked it yesterday and Tau's third build program is broken, and yes it's because of the 600 requirement. Kroot nest was moved to tier 2 in patch 1.2 and Karnovore are now built from the barracks. I didn't know that, but now it's possible to include the Tau barracks to the starting build order and play the Kroot strategy with a commander. That's a definite improvement. Tau build programs should be fixed now...

Otherwise : Wowww !

The game went on for 1h and it was a stalemate when I did quit. Ork + Chaos vs Tau and IG. Slaughter and mayhem and it was like watching a very good movie. Excellent job Arkhan ! I lost track of writing down stuff as it was most fascinating to enjoy the show. We had all races T4 with baneblade, gnarloc, bloodthirster, demon lord, squiggy and most of them fighting each other multiple times. As soon as one of them was in danger to get anihilated his ally came to rescue him. rinse and repeat, back and forth. I think this will be a great release.

Candy for my ears...

1) Orks - What should be push out for Orks first? Trakks (mostly anti-vehicle) or Trukks (purely anti-infantry transport)? KillKans will be kept "as is".

Since not even the pros build them I think trakks are out of discussion for early T2 control.

2) Orks - Should we force the Grot Infiltrate to all buildprograms? Right now its set as a dynamic build for Tier1 so when 2 Grot squads are present and the AI has 50req it should build it. Not always but a forced build would always build it.

We don't need this. The reason why grot infiltration was researched so fast is because dynamic researches have special rights now. They are built even if resources are saved for a build item. This solved a lot of problems with researches not be built fast enough like the NL abilities.

4) IG - should we force out a Priest from the CS immediately? I have to check what leaders are available to the CS at gamestart.

I think I make a check for infiltrated enemies. If there are any on the table I force a psyker out and if not we take a priest.

6) Are we concerned about having any attachable units accidentally attach to infiltrating squads? (ie. Apothacary attaching to Scouts with Infiltration researched)?

This is weird since we've already a attachment check which should avoid this...

7) Should we have a check so that Jumpable Assault squads won't jump into CC vehicles? AI can't pull this off effectively and chop down vehicles in melee?

Hmm, melee jumpers only jump attack non-walker vehicles. That forces a lot of vehicles to dance since they can't fight in melee. Raptors might not be the perfect anty vehicle unit but assault marines have melta bombs against them.

[LarkinVB]

Why are tau drones forced in all build programs ? I rarely watched a good tau player using them at all.

What is the idea behind the Special Kauyon Strategy ? I don't think it does work at all yet. It has no Rush unleash.

About the CS squad : My argument was that the CS should reinforce as fast as possible as soon as it is build, regardless of other restrictions. A lone harassing general isn't good enough to do any real harm.

Edit: Perhaps enforce fast CS reinforcement only while harass is checked in control panel ?

Edited by LarkinVB, 28 April 2007 - 11:20 AM.

[LarkinVB]

Hmm, melee jumpers only jump attack non-walker vehicles. That forces a lot of vehicles to dance since they can't fight in melee. Raptors might not be the perfect anty vehicle unit but assault marines have melta bombs against them.

The raptors did jump a hellhound and were quickly dispatched by the two guardsmen squads 15 behind.

[ArkhanTheBlack]

Why are tau drones forced in all build programs ? I rarely watched a good tau player using them at all.

Make a better suggestion! I don't know the current killer combi for early T2 vehicles.

What is the idea behind the Special Kauyon Strategy ? I don't think it does work at all yet. It has no Rush unleash.

It's a melee strategy. Two karnivore squads are forced. T2 counts as 'natural unleash' for T1 rushs. Rush unleash commands are only used for fast vehicle techs now.

About the CS squad : My argument was that the CS should reinforce as fast as possible as soon as it is build, regardless of other restrictions. A lone harassing general isn't good enough to do any real harm.

I've done that already. Will be in in next build!


A general question: Did anybody recognize that the AI sometimes doesn't build posts on their strats? It's a huge problem on battle marshes with several race (Eldar seem to be immune to the problem). This is definitely a new DC 1.2 problem. I've never seen this before and I'm not able to fix it since the AI seems to get a wrong information from the capturing_flag:HasListeningPost() function. I'm not able to fix that! It's simply broken in the core routine. I hope the problem is at least restricted to only a few maps. I didn't see it on Kasyr for example.

Edited by ArkhanTheBlack, 28 April 2007 - 11:19 AM.

[thudo]

A general question: Did anybody recognize that the AI sometimes doesn't build posts on their strats? It's a huge problem on battle marshes with several race (Eldar seem to be immune to the problem). This is definitely a new DC 1.2 problem. I've never seen this before and I'm not able to fix it since the AI seems to get a wrong information from the capturing_flag:HasListeningPost() function. I'm not able to fix that! It's simply broken in the core routine. I hope the problem is at least restricted to only a few maps. I didn't see it on Kasyr for example.

Yes! It only happens on some maps but its not too consistent. It only seems to affect one strat point on a map near to the startpoint. I find that if I do this to the enginee's tactic file:

-- Update engineer tactic
	--EngineerTactic.Update(self)

It works much more reliably especially when loading a whole new map. Very unusual. Its NOT 100% but the way the Engineers behaved seemed a little more fluid and were more quicker to building LPs when Strats were around. Also... saw something interesting..

Tomb Spyders do not attack enemy LPs. They just sit there beside it being hit by them whereas a Wraith came up and just pummelled right into it. Odd? The Tomb Spyder, however, will attack all other enemy buildings fine with devastating melee attacks! I find this happens for a decent # of units. The NecronLord almost never melee attacks enemy LPs for example.

Edit: Ok never mind.. I played 4 games and each one the Tomb Spyder melee'ed the enemy LP. Unsure why its working in the map 2p_BloodRiver. Also saw the Necron Lord melee an enemy LP as well. hehe..

Edit #2: The best map to see the problem with the builders not building LPs on their own captured strats is 6p_Mortalis. There are alot of incidences where AIs won't build LPs on their own captured strats (ie. Place an Ork AI on Startpoint #5 and he won't build an LP on the bottom-right strat). Also.. place an SM player @ StartPoint#1 and make it controlled by the AI -- the SM AI will not build either between 1-2 LPs locally. Very odd but this has been going on since DC 1.2.

[thudo]

Additional observations to above...

It seems the SM, IG, and Orks are the most affected by the "stop building LPs on available local strats" bug. I cannot entirely reproduce the same with CSM, Eldar, Necrons, and Tau -- these 4 seem to work fine. Now what I did was to test the Orks on 6p_Mortalis as when you place them in Startpoint#5 they do not build an LP on the bottom-right strat almost all the time. So to fix that I did this real hack job as a test scripted in the dynamic build:

-- Force build generators since AI has issues keeping up with them decaying all the time
		  if (resource_manager:GetResourceAmount():Get( ResourceAmount.RT_Requisition ) > 90) then

		local tBuildType = CpuBuildType()
		tBuildType.btype = CpuPrerequisites.BT_Building
		tBuildType.name = "ork_gork_totem"
		if (self:TryBuild( tBuildType )) then
			aitrace("BuildController: Dynamic build of "..tBuildType.name)
		end
	end

And now it works. Its just as a test. Now the Orks always build LPs although obviously without the threat checks I believe.

IG are still ok now with building all LPs on available strats with my code ommision mentioned above to the Enginseer's tactic file. Their LP to the right of the HQ is only slightly delayed rather than delayed for long periods. Very wierd... more testing continues..

[ArkhanTheBlack]

Yes! It only happens on some maps but its not too consistent. It only seems to affect one strat point on a map near to the startpoint. I find that if I do this to the enginee's tactic file:

-- Update engineer tactic
--EngineerTactic.Update(self)

It works much more reliably especially when loading a whole new map. Very unusual. Its NOT 100% but the way the Engineers behaved seemed a little more fluid and were more quicker to building LPs when Strats were around. Also... saw something interesting..

This will break the builder code. No repairing, build assisting, retreating, etc. anymore. Definitely not good!

It seems the SM, IG, and Orks are the most affected by the "stop building LPs on available local strats" bug. I cannot entirely reproduce the same with CSM, Eldar, Necrons, and Tau -- these 4 seem to work fine. Now what I did was to test the Orks on 6p_Mortalis as when you place them in Startpoint#5 they do not build an LP on the bottom-right strat almost all the time. So to fix that I did this real hack job as a test scripted in the dynamic build:

An interesting approach! I tried a common approach for all races but it's very unreliable. On battle marshes it only worked sometimes.