5127 replies to this topic

[Speeder]

AI will use all those support powers (Drakuv, Spinblade) properly in the next update.

[BotRot]

(Only applies in Skirmish/Multiplayer) Since Scorpion Cell AI always builds Psychic Dominator first and Fake Psychic Dominator later, it becomes predictable which Dominator is constructed first. (unless player does not pay attention to Domination countdown).

 

P.S.  Soviet AI does not seem to build Dustdevils at all, making them the only non-StolenTech jet that AI does not build. Perhaps they don't have a set squad for it, or is there another reason?

 

P.S.  Extra Irkallas from Mental AI Boost tend to idle and rarely shoot nearby enemies, making HQ AI defense much weaker than they should be. 

Edited by BotRot, 24 February 2020 - 01:05 PM.

[Daffa the Mage]

I'm not sure if this has been discussed about, but I think Armadillo is quite powerful for it's cost value. At least from my runs in skirmish, these things are broken as hell when they first get unlocked (T2).

 

With like 6 these things can rush in and wreck a ConYard + some buildings in between. I noticed particularly in Skirmish that if a Chinese AI decides to Armadillo Rush early, whoever got struck by that attack (other AIs mostly) is more likely to get wrecked. It doesn't help the fact that Armadillo is T2. Borillo is somewhat broken as well, but much more manage-able to handle due to their less bulky armor than these pesky Armadillos.

 

Gharial seems to be pretty dangerous too, but since it's cost value is high (slow training and expensive cost), I suppose it's sort of balance itself.

[Malekron]

I'm not sure if this has been discussed about, but I think Armadillo is quite powerful for it's cost value. At least from my runs in skirmish, these things are broken as hell when they first get unlocked (T2).

 

With like 6 these things can rush in and wreck a ConYard + some buildings in between. I noticed particularly in Skirmish that if a Chinese AI decides to Armadillo Rush early, whoever got struck by that attack (other AIs mostly) is more likely to get wrecked. It doesn't help the fact that Armadillo is T2. Borillo is somewhat broken as well, but much more manage-able to handle due to their less bulky armor than these pesky Armadillos.

 

Gharial seems to be pretty dangerous too, but since it's cost value is high (slow training and expensive cost), I suppose it's sort of balance itself.

Maybe add t2 anti-tank units/infantry for the factions to deal with the Armadillo?

[Opus Custom Tank]

I'm not sure if this has been discussed about, but I think Armadillo is quite powerful for it's cost value. At least from my runs in skirmish, these things are broken as hell when they first get unlocked (T2).

With like 6 these things can rush in and wreck a ConYard + some buildings in between. I noticed particularly in Skirmish that if a Chinese AI decides to Armadillo Rush early, whoever got struck by that attack (other AIs mostly) is more likely to get wrecked. It doesn't help the fact that Armadillo is T2. Borillo is somewhat broken as well, but much more manage-able to handle due to their less bulky armor than these pesky Armadillos.

Gharial seems to be pretty dangerous too, but since it's cost value is high (slow training and expensive cost), I suppose it's sort of balance itself.

Armadillos not a treat if you have T1 tanks. Borillo and Armadillo already enough nerfed on precious patchs. Btw I love Raven.

[Tathmesh]

The Gharial vehicle itself is only worth $800 actually. Each Knightframe is $400. Gharials are ridiculously cost-effective lol. 

[Daffa the Mage]

 

I'm not sure if this has been discussed about, but I think Armadillo is quite powerful for it's cost value. At least from my runs in skirmish, these things are broken as hell when they first get unlocked (T2).

 

With like 6 these things can rush in and wreck a ConYard + some buildings in between. I noticed particularly in Skirmish that if a Chinese AI decides to Armadillo Rush early, whoever got struck by that attack (other AIs mostly) is more likely to get wrecked. It doesn't help the fact that Armadillo is T2. Borillo is somewhat broken as well, but much more manage-able to handle due to their less bulky armor than these pesky Armadillos.

 

Gharial seems to be pretty dangerous too, but since it's cost value is high (slow training and expensive cost), I suppose it's sort of balance itself.

Maybe add t2 anti-tank units/infantry for the factions to deal with the Armadillo?

 

Could be a solution, though I think almost all factions has T2 anti-tank from what I noticed, but doesn't use them effectively as AI at least.

 

 

I'm not sure if this has been discussed about, but I think Armadillo is quite powerful for it's cost value. At least from my runs in skirmish, these things are broken as hell when they first get unlocked (T2).

With like 6 these things can rush in and wreck a ConYard + some buildings in between. I noticed particularly in Skirmish that if a Chinese AI decides to Armadillo Rush early, whoever got struck by that attack (other AIs mostly) is more likely to get wrecked. It doesn't help the fact that Armadillo is T2. Borillo is somewhat broken as well, but much more manage-able to handle due to their less bulky armor than these pesky Armadillos.

Gharial seems to be pretty dangerous too, but since it's cost value is high (slow training and expensive cost), I suppose it's sort of balance itself.

Armadillos not a treat if you have T1 tanks. Borillo and Armadillo already enough nerfed on precious patchs. Btw I love Raven.

 

Interesting. Never know it has been nerfed multiple times before. Borillo is still manage-able, so I think they're not that imbalanced, just need proper response. Armadillo problem seems to result from their health or armor (or perhaps both, not sure). An easy fix that I figured would be to add walls to AI ConYard to give them extra protections against Armadillo + Iron Curtain + Irradiation ultimate combo attack (last night I played, I need to cover allied Conyard with Iron whenever this combo happened, just lucky I got random Soviet last night), like they cover their labs. At least that helps me in my tests against China's rushes.

 

I'll be honest, I never noticed Borillo and Armadillo are ridiculous until I played an island map and lost allied base in like 5 first minutes into the game xD

 

Ah yeah, personally Raven is my favorite Teen Titans character

 

The Gharial vehicle itself is only worth $800 actually. Each Knightframe is $400. Gharials are ridiculously cost-effective lol. 

Still, it's T3 so by then people should be aware stuff like them are coming (especially in water maps). But now you mentioned it, it might be extremely cost-effective. I noticed the training is quite slow though, so yeah. Usually a proper map spying and some front line cover would help. The annoying side of Gharial is how it's 'equipped with as many anti battlefield hazard equipment as possible' rather than how quickly they deal damage actually

 

Oh, and I assume Knightframe+Gharial+Sync tactic is actually how people would use Gharial? At least somebody said that on one of my videos.

 

---

 

I'm curious for Scorpion Cell players, do you guys use Hijackers? I can't seem to find the ideal situation of why I would use this unit, so please give some insight. I at least get the Mortar Quad somewhat, but still can't wrap my head for Hijacker.

 

For Arsonists, how long actually they need for the fire to set blaze, and how it exactly works? (I mean spread and damage wise). I cannot discern exact damage so I usually just burn as much as possible and put Arsonist on loop attack so they can keep the fire spree without knowledge of how much damage I can dealt with each attack.

 

Is Hailstorm nerfed a lot? It's anti-structure attack seems to extremely pales in comparison to other anti-structure tanks from what I noticed. Kinda makes me confused how to siege with Pacific late game, cause I used Hailstorm in 3.3.0 patch and back then, as Hailstorm said it, "It's super effective". Right now, felt like that I need 2 Hailstorm for each siege artillery my enemy has (perhaps range advantage is the only possible thing I can say being the nerf reasoning).

[BotRot]

This is probably in the to-do list already, but still mentioning it anyways.

 

Rage and Shadow Ring radius has been increased since v3.3.0, but the visual effects are not updated. Perhaps it can be made larger in the next version to match the actual radius of these support powers?

[Verthunder]

Hi, lets bring them back original cameos for Navy Seal and for Chrono Legionaire. Its just like Conscript and Flack Tropper in patch 3.3.3. Love original ones

;

Edited by Verthunder, 29 February 2020 - 05:14 PM.

[BotRot]

If memory serves, Epsilon in PvP excels in infantry spam. As the current stolen tech unit, the Dybbuk-Seizer, does not really synergize with this strategy (e.g. effectiveness vs monster tanks is limited; Mastermind is self-explanatory, Tyrant can dig underground to avoid mind control and Colossus AA with enough numbers can take down Seizers before they make an impact), why not replace it with a unit that does synergize in an Epsilon vs Epsilon matchup?

 

Enter this unit suggestion, the return of the Chaos Drone, but with new role/s that might mix well in PvP gameplay better than the Dybbuk-Seizer (in theory).

 

Instead of chaos gas like in the original game, it releases Terminus gas instead, making it a fast anti-infantry blob unit. Its robotic quality already makes it a counter to Adept/Elite spam. 

 

The catch is that it is Cloaked to catch the Epsilon enemy by surprise. An alternative is give it cloak or disguise detection to counter enemy attempts to infiltrate your base.

---

 

To add more to its functionality and make it different from the Dybbuk-Evolver,  a deploy function that causes it to release Rage to nearby friendly units (on paper sounds more impactful than Dybbuk-Seizer single target Rage), BUT:

 

- Option 1: Reduced radius (v3.0 radius), halved duration and causes Chaos Drone to explode on deploy. Also makes the Chaos Drone a one-time AOE infantry healer.

 

- Option 2: Instantly kills your friendly infantry (also from v3.0) but deploy has a long cooldown and Chaos Drone cannot attack at that time, making it an armor battalion supporter instead. It also keeps the "meme" status of Epsilon stolen tech, an anti-infantry blob unit but press D at the wrong time and you lose YOUR infantry blob instead

 

In either case, it reduces the possibility of a scenario of perma-Rage on Epsilon heroes.

 

In conclusion, a stolen tech unit that is part stealth mutator, part Rage inducer which, in theory, can turn the tide in Epsilon vs Epsilon matchup far better than the Dybbuk-Seizer, which is better off against other faction matchups.

Edited by BotRot, 02 March 2020 - 09:50 AM.

[Nox667]

Correct me on this if im wrong, but wasnt the term "terminus gas" used back in 3.0 for dreamweaver and deviator, the confusion weapons of Epsi? Since you mean a mutation weapon it would be Terranova i believe.
Regarding the idea: i genuinely like it. A stealthed, mc immune ST for EvE matchup sounds hella useful and most superior to Dybbuk-S. Unfortunately, it also sounds too similar to Archelon :/
Just its fast and fragile rather than slow and tanky, and instead of leaving healing areas for your own infantry it mutates enemy infantry.

Edited by Nox667, 01 March 2020 - 09:58 PM.

[BotRot]

Correct me on this if im wrong, but wasnt the term "terminus gas" used back in 3.0 for dreamweaver and deviator, the confusion weapons of Epsi? Since you mean a mutation weapon it would be Terranova i believe.

The term Terminus remains used in Epsilon based off the official descriptions of the Rage Inductor and Genomines.  Strangely Terminus is not mentioned anywhere in the campaign but Terranova is.

[Opus Custom Tank]

Correct me on this if im wrong, but wasnt the term "terminus gas" used back in 3.0 for dreamweaver and deviator, the confusion weapons of Epsi? Since you mean a mutation weapon it would be Terranova i believe.

The term Terminus remains used in Epsilon based off the official descriptions of the Rage Inductor and Genomines. Strangely Terminus is not mentioned anywhere in the campaign but Terranova is.

Looks like you are right. Genomines using Terminus. Then what is Terranova? I mean if Terminus is genetical mutation, what is difference between Terminus and Terranova? I thought Epsilon using Terranova term for genetic mutation.

[MCFLYYY]

I don't play PvP, just like to play the AI in some skirmishes now and again with a buddy. But the Mental AI needs some tweaking. Giving them $4m, 3x production ability, and more frequent attacks makes facing them very one-dimensional: just turtle up.

 

1. You can't possibly match them with ground units given their financial and production boosts. So instead just mass base defense at choke points because you can repair those for cheap.

2. Mass AA defense to stop the waves of Kirov assaults

3. Mass some kind of air unit for mobile base defense (Warhawk, Wolfhound, etc).

4. Once you're completed defending from the total onslaught that is 3-4 Mental AI attacks, you can begin massing your own attack force. BUT, this attack force must always be air units because it is very difficult to seige a Mental (or 2 if they're spawned near each other) because of the river of units coming out of their base. So instead just mass some kind of air unit (Warhawk, Wolfhound, Queztel etc) and go to work.

5. Target their power plants first, because they always run right at max power. So taking down 1 nuclear reactor or a couple power plants will put them under power and de-power all air defense.

6. From their it's very easy to wipe out the rest of their power and delete the rest of the base, while they waste their time cranking SCUDs out of their 4 War Factories.

 

I get that making good AI is difficult. But giving them the ability to churn out 12x as many units as you can (with all their production bonuses) and unlimited funds means you can't use varying strategies to defeat them. Some suggestions:

 

1. They constantly use spy planes - why not actually put the intel gained from this to use. They should be spamming counter units to whatever it is you're building. Doesn't AOE do this?

2. They shouldn't constantly attack every time they build "X" number of units. Let them mass up bigger attacks that can actually defeat base defense. Have AI "types" that do different things at random. Sometimes they go air heavy, sometimes they go tank heavy. Sometimes they rush T2, sometimes they turtle up and wait for you to attack before countering.

3. Group units that defend each other together. Makes no sense to see them fly in 3 Kirovs undefended to get shot down every 5 minutes, and then 3 Wolfhounds come in 2 minutes later each time. Why not send them together? Why not send 10 Wolfhounds with those Kirovs? Why not wait and send 9 Kirovs at once?

 

I also feel like the Foehn Reprocessor is way too broken against Mentals, with the constant stream of units they send into your base defense it gives you unlimited funds while the other factions are entirely dependent on ore/gem deposits.

 

/rant

 

Love this mod. Can't wait to see what's in store for future updates. Just would like to see some balance and creativity used in the Mental AI program.

[Handepsilon]

AIs can't be upgraded very far as far as I'm aware of. Their AI is basically constructed through sets of scripts which only goes as far as 'Attack this type', 'Wait for Tech Tree X' etc., Giving them limited money would end with them selling their own buildings just to keep their army pumping up.

 

Also, massing up armies are generally TERRIBLE idea to apply to AI. They will generally clump up in front of the base, trying to occupy one singular cell while massing their assembled army and keeps failing and thus lagging the game and never have them attacking at all. I had this happening in a lot of my triggers in custom maps that involves AI creating their own army, so.... yeah

[MCFLYYY]

So I may be wrong on some/all of this due to my limited knowledge of this mod in general but...

 

AIs can't be upgraded very far as far as I'm aware of. Their AI is basically constructed through sets of scripts which only goes as far as 'Attack this type', 'Wait for Tech Tree X' etc., Giving them limited money would end with them selling their own buildings just to keep their army pumping up.

 

This isn't true for regular CNC is it? If you kill their ore trucks, they rebuild them. And if you keep killing them til they run out of money, they stop building ore trucks and stop attacking as well. They simply sit idle because they're out of funds. I'm not sure I've ever seen an AI sell buildings in order to afford another ore truck.

 

 

Also, massing up armies are generally TERRIBLE idea to apply to AI. They will generally clump up in front of the base, trying to occupy one singular cell while massing their assembled army and keeps failing and thus lagging the game and never have them attacking at all.

 

Why is this not an issue currently then? If they build 3-4 tanks before sending them to attack in a group, by your logic all 3-4 tanks would be trying to occupy one single cell and failing and lagging the game and never attacking? How is it they're able to mass up 3-4 tanks before attacking but they can't mass up 5-6 or 7-8? Or why can't they mass 3-4 of multiple unit types and send them in a single group? I'm not asking for the AI to mass 100 tanks in front of their base, but I'd like to think they could do a little more than 3-4 at a time? 

 

And what about the other suggestions? Having the AI build counters to whatever you're building, having AI "types" which cause them to build more aircraft or tanks or rush/turtle. Can we add more resources to the maps or maybe have an option to increase the amount gained from capturing derricks? That way you can use factions other than Foehn when facing 6 Mentals?

 

Also, why does the game start to lag when we play together online? When I play 1v3 Mentals offline the game runs smooth and fast, but when we've previously played 2v2 online together the game slows to a crawl? If it's too many units onscreen wouldn't that be an issue in Offline Skirmish too? Is there something we can do to fix the online experience?

 

Thanks for your response btw!

[Nox667]

As far as im aware, the possibilities for AI are very limited in this engine, as hande pointed out already. Pathfinding and how waypointing works is screwy to say the least, a difference of one unit in the selection or presence of one unwalkable tile can mess things up big time already. For ai this seems ot be even worse.

They are already "countering" your units, since they have scripts that trigger certain teams to deploy when the player gets certain stuff, but its in no way dynamically feasible, mostly extends to them sending dogs whenever you get some spies. Even if such counter teams were more pronounced, the way attack scripts for the ai work would still leave the units braindead in the field. They attack their originally designated target (a miner for example) until something shoots them, and if they are attacked, they retaliate and ignore anything but the unit that attacked them first. This in turn would result in the counter teams becoming pointless since the units will still run into defenses and die and we are back to square one.

Another heavy limitation is that AI can not be made to respect shroud at all, they are always omniscient. Stealth is only respected to a limited degree. As soon as a stealthed unit has been revealed, the ai will send attack teams/throw support powers at it, and this stuff might arrive/attack the target long after the unit has been cloaked again. Disguised units can and will be attacked by ais supportpowers even though their units respect the usual engagement rules on them.

Dont know if different AI personalities akin to CnC3 are possible in the engine, but since i dont know any mod that attempted implementing smth like it im suspecting its outright impossible, or just not worth the effort.
About ressource issues, one word: spies. AIs near infinite cash makes it the best money source in the game. Outside of that, bear in mind that the game is and always will be balanced around pvp, not pve. Thus any changes to economy will have to be taken according to pvp balance, and as such, stronger eco for long slog type games is rather improbable. Reprocessor in general is a troublesome thing, for both pvp and pve though. I'd wish its bonus would just be completely changed for smth else, though i dont know what the replacing bonus could be.

Online play will always run less smooth than offline, since data has to be transferred over the net between the players. It gets worse the more units are on screen/on the map, since all these units, their current state, position, etc. etc. all needs ot be synced between the players.
Only things that can be done about that is picking a stable tunnel server thats preferably close to all participating players, avoiding to play with people on the other end of the world and improving the local internet connection of each player, though thats usually not in the affected persons hands. Even with ideal conditions an online match will always run slower than an offline one.

Edited by Nox667, 03 March 2020 - 11:49 PM.

[Zhang Jian]

I'm not sure if this has been discussed about, but I think Armadillo is quite powerful for it's cost value. At least from my runs in skirmish, these things are broken as hell when they first get unlocked (T2).

 

With like 6 these things can rush in and wreck a ConYard + some buildings in between. I noticed particularly in Skirmish that if a Chinese AI decides to Armadillo Rush early, whoever got struck by that attack (other AIs mostly) is more likely to get wrecked. It doesn't help the fact that Armadillo is T2. Borillo is somewhat broken as well, but much more manage-able to handle due to their less bulky armor than these pesky Armadillos.

 

Gharial seems to be pretty dangerous too, but since it's cost value is high (slow training and expensive cost), I suppose it's sort of balance itself.

I suggest that Armadillo have less HP while Pangolin have less fire. See details here: https://forums.revor...36#entry1103653

Edited by Zhang Jian, 04 March 2020 - 03:50 AM.

[Zhang Jian]

If memory serves, Epsilon in PvP excels in infantry spam. As the current stolen tech unit, the Dybbuk-Seizer, does not really synergize with this strategy (e.g. effectiveness vs monster tanks is limited; Mastermind is self-explanatory, Tyrant can dig underground to avoid mind control and Colossus AA with enough numbers can take down Seizers before they make an impact), why not replace it with a unit that does synergize in an Epsilon vs Epsilon matchup?

 

Enter this unit suggestion, the return of the Chaos Drone, but with new role/s that might mix well in PvP gameplay better than the Dybbuk-Seizer (in theory).

 

Instead of chaos gas like in the original game, it releases Terminus gas instead, making it a fast anti-infantry blob unit. Its robotic quality already makes it a counter to Adept/Elite spam. 

 

The catch is that it is Cloaked to catch the Epsilon enemy by surprise. An alternative is give it cloak or disguise detection to counter enemy attempts to infiltrate your base.

---

 

To add more to its functionality and make it different from the Dybbuk-Evolver,  a deploy function that causes it to release Rage to nearby friendly units (on paper sounds more impactful than Dybbuk-Seizer single target Rage), BUT:

 

- Option 1: Reduced radius (v3.0 radius), halved duration and causes Chaos Drone to explode on deploy. Also makes the Chaos Drone a one-time AOE infantry healer.

 

- Option 2: Instantly kills your friendly infantry (also from v3.0) but deploy has a long cooldown and Chaos Drone cannot attack at that time, making it an armor battalion supporter instead. It also keeps the "meme" status of Epsilon stolen tech, an anti-infantry blob unit but press D at the wrong time and you lose YOUR infantry blob instead

 

In either case, it reduces the possibility of a scenario of perma-Rage on Epsilon heroes.

 

In conclusion, a stolen tech unit that is part stealth mutator, part Rage inducer which, in theory, can turn the tide in Epsilon vs Epsilon matchup far better than the Dybbuk-Seizer, which is better off against other faction matchups.

I create a new kind of Dybbuk-Seizer. See details here: https://forums.revor...36#entry1103653

Edited by Zhang Jian, 04 March 2020 - 03:50 AM.

[Handepsilon]

This isn't true for regular CNC is it? If you kill their ore trucks, they rebuild them. And if you keep killing them til they run out of money, they stop building ore trucks and stop attacking as well. They simply sit idle because they're out of funds. I'm not sure I've ever seen an AI sell buildings in order to afford another ore truck.

That ain't part of the attack script. I think they just do that in another part of the script. I believe that would be dictated in rules.ini
 

Why is this not an issue currently then? If they build 3-4 tanks before sending them to attack in a group, by your logic all 3-4 tanks would be trying to occupy one single cell and failing and lagging the game and never attacking? How is it they're able to mass up 3-4 tanks before attacking but they can't mass up 5-6 or 7-8? Or why can't they mass 3-4 of multiple unit types and send them in a single group? I'm not asking for the AI to mass 100 tanks in front of their base, but I'd like to think they could do a little more than 3-4 at a time?

My bad, I believe they can at least put 9 tanks in and around that one waypoint, as long as they stand still afterwards, but that's the maximum and even then, it's probably sketchy enough not to attempt
 

Also, why does the game start to lag when we play together online? When I play 1v3 Mentals offline the game runs smooth and fast, but when we've previously played 2v2 online together the game slows to a crawl? If it's too many units onscreen wouldn't that be an issue in Offline Skirmish too? Is there something we can do to fix the online experience?

  This is the terrible state of TS Engine's multiplayer code (Red Alert 2 uses TS Engine, yeah). While other RTS games will usually cause the lagging client to speed up in order to catch up with the game's current state on the host's own end, such as for example, Warcraft 3, TS Engine games like Tiberian Sun and Red Alert 2 will slow down all clients depending on who's the slowest plus the state of your network connection. So if one of you lag so bad, so does the rest of you. That's not something anyone can change as far as I know
 
And the lag is not just from on-screen units either. The unit's movement pathfinding itself is quite laggy when done en-masse, which is why maps like Juggernaut mission and many many many custom challenges that involves massive enemy reinforcements wave are very laggy by default, even if you're not looking at the moving unit

I suggest that Armadillo have less HP while Pangolin have less fire. See details here: https://forums.revor...36#entry1103653

....What is this 'Pangolin'? Or were you talking about Borillo?

Edited by Handepsilon, 04 March 2020 - 07:33 AM.