Home Trending Latest threads New posts Developer diaries

Naga DLC campaign is pathetic

Jump to latest Follow Reply
  • We have updated our Community Code of Conduct. Please read through the new rules for the forum that are an integral part of Paradox Interactive’s User Agreement.
T

TheDeaddude

Private
2 Badges
Mar 28, 2014
11
0
  • Warlock 2: The Exiled
  • Warlock: Master of the Arcane
Campaign played on "Impossible Difficulty" (small map setting)
Naga is a weak pitiful race that just gets face rolled by catapults/warriors (Humans) and restore ardania (Krolm). Naga tower appears out of no where? No problem restore ardania will destroy the tower and the water at once. Send in your warriors/veterans with catapults to clear units out.[Downside to not destroying naga towers with units/spells is no gold/mana].

AI's x2
They are being slaughted by random naga spawns and water terraforms. They are too stupid to restore their own lands. 1 AI has only 1 city and dies as soon as they hit the water. Another AI has 3 cities and water all around/inbetween their cities.

You think if they add a race solely dependent on water; they would make the AI's able to stand their own against the random naga/water terraforms? Guess not.

The great battle
Ok so i got to battle the naga boss. This boss is freaking near impossible to kill, every 5 turns it heals all 750 hp. Boss levels as you fight it and ithas 110, melee 110 ranged and 75+ to the rest of resistances. Nothing short of 4 Legendary pimped up heroes or an Army of lich's is going to put this boss in his place. [Or alot of level 5+ catapults and Krolm's wrath spell].

Game said the united one and naga are working together yet i see no united one or his minions?
Ok so after i beat the naga boss there is no united one or any of his minions you have to fight or kill. IMHO getting to the naga boss is easy, killing the boss requires at least 2 legendary lords and another 2 lords. The lords need to be crazy buffed and you need an army of temple units such as lich's that can hit 3 tiles away. Why? because the naga boss rains 84 AoE spirit damage on everything within 6 tiles. Nothing short of pimped up lords with all those resistance perks and spells you've been stacking is going to sruvive 84 spirit damage. Fog of war and long range units are your friends. For humans i substituted lich's with leveled catapults and Krolms wrath spell for the extra punch on every unit i own (which includes my lords that deal 200+ damage already).

Naga boss (level 6)
Health: 750
melee resist 105 Death resist 90
ranged resist 115 Elemental resist 90
life resist 175 Spirit resist 75
Sight range +2, immune to banes, can see invisible units and restores full health every 5 turns.
Damage = 40 spriit, 28 elemental, 12 death.
Attack skill = Rain of water 84 spirit damage on target and around 6 tiles around it. Says it has 3 turn c/d but it spammed that ability every turn. Whatever i still killed him

Water anybody?
Achievement status is currently 0.03% on steam, not alot of people can beat this boss or simply didn't buy dlc huh? I wonder why? (sarcasism).

Full opinion of DLC, campaign and all?
It's horrible for $20 price range. Only worth $5 IMHO because that's how little content they added to the dlc. 1 race and 1 half-♥♥♥♥♥ campaign with no challenge even on impossible difficulty. I never once saw a naga AI building a city or attacking me outside of random trigger spawns. How the hell can you call this an epic battle? the whole game felt like a giant vacuum of nothing from your shard all the way to ainadra.

What are your thoughts about the naga dlc campaign? Do you love or hate it? Why?
 
Last edited:
The greatest problem for this game was the ai and it seems it will be the greatest problem in the future.
The devs can't create an ai which is able to handle simplest things.

Thats very sad as the game and the ideas are wonderful, but without an better ai thats not enough.

Very sad!
 
Well crappy AI's are the norm for just about every game you can think of. Age of wonders 3, disciples 1, 2 and 3, Heroe's of might and magic 1-3 (4+ are not worth naming). I've seen first person shooters and hack & slash games evolve way more than the turn based strategy game's. They literally didn't improve anything from games made in 1990, just made it look prettier. Even RPG's have evolved to the point a new genre was created such as MMORPG. If game dev's don't step up and make real changes i'm not buying another turn based tabletop game ever again short of a board game played in real life with friends.

Most of the people who know of and play turn based strategy games are from the late 19th century and if they don't do any real improvements to the old game genre they will have no fans. Take legends of grimrock for example, they used way old games nostalgia and just recreated it and made it... BETTER, why should this be so hard to pull off? We want BETTER not the same old boring crap that made us rip our hair out or frown because the game had so much potential but lacked so much initiative.

For exmaple
Legends of grimrock made the old game slightly more forgiving, kept the difficulty and added save/load so players didn't have to restart at level 1 every time they died. Why? because in the 20th century time is valuable and not alot of us have time to spend 4 hours in one sitting to "beat" a game. This is evolution and this is my example of what warlocks 2 should have improvised upon the problems of warlocks 1. Listen to your fans and make those changes we are crying for and you will be 3x more successful. Ignore us and fail.
 
Last edited:
I don't think you're being entirely fair.

Paradox very clearly did listen to their fans. Warlock 2 has highly requested improvements in spell research, spell casting, a more useful unit listing, and especially modding support. The diplomacy is virtually untouched and the AI seems very similar to the first game's, but pretending that they just didn't listen at all is obviously untrue.

You might check out Galactic Civilizations II for a better AI experience, but really AI is not a solved problem in any system that's remotely open. We haven't even perfected Chess AI, let alone a complex strategy game.

Holding up the FPS genre as one that has advanced other than graphically makes me wonder whether I even understand what you're talking about.
 
Grygus said:
....We haven't even perfected Chess AI....
Click to expand...

WHAT???

Thats not true, the chess computer ai are now at a very high level and can beat grandmasters!

Read this:

Pocket Fritz 4 (2009)

In 2009 a chess engine running on slower hardware, a 528 MHz HTC Touch HD mobile phone, reached the grandmaster level. The mobile phone won a category 6 tournament with a performance rating 2898. The chess engine Hiarcs 13 runs inside Pocket Fritz 4 on the mobile phone HTC Touch HD. Pocket Fritz 4 won the Copa Mercosur tournament in Buenos Aires, Argentina with 9 wins and 1 draw on August 4–14, 2009. Pocket Fritz 4 searches fewer than 20,000 positions per second. This is in contrast to supercomputers such as Deep Blue that searched 200 million positions per second. Pocket Fritz 4 achieves the same performance as Deep Blue.

Pocket Fritz 3 using version 12.1 of Hiarcs won the same event the previous year with six wins and four draws, running on a 624 MHz iPaq hx2790. The 2008 Mercosur Cup was a category 7 tournament. Pocket Fritz 3 achieved a performance rating of 2690.
Click to expand...

So there are huge improvements in ai programming, but game designers seems to prefer "nice graphics".....
 
jk_bonn said:
the chess computer ai are now at a very high level and can beat grandmasters!
Click to expand...
And they took forever and a half to develop to that point. And the rules of chess are pretty simple, compared to the complex systems running your average video game these days.

The AI would not only have to deal with unit positioning, it would have to manage resources, city placement, research, unit and building construction, diplomacy, and the fact that it cannot see the entire board at a time. And that's just off the top of my head.

Yes AI in most games suck, yes they could probably be better. But there is no way it could even be anywhere close to a chess AI.
 
Chess AI is very good, but that is not counter to what I said. I said it was not perfected. Chess AI still loses sometimes. Your example got a draw in only 10 games. That is excellent! But not perfected. It can improve.

And as Shidan points out, this is a greatly simplified game upon which AI programmers have been working for decades. AI is a very hard problem. Even in your example, that represents 12 years of progress!
 
Grygus said:
I don't think you're being entirely fair.

Paradox very clearly did listen to their fans. Warlock 2 has highly requested improvements in spell research, spell casting, a more useful unit listing, and especially modding support. The diplomacy is virtually untouched and the AI seems very similar to the first game's, but pretending that they just didn't listen at all is obviously untrue.

Holding up the FPS genre as one that has advanced other than graphically makes me wonder whether I even understand what you're talking about.
Click to expand...

I guess i let my anger cloud my judgement and refused to look at what they did do as opposed to what they didn't do.
I'll give you that much, but if you look at FPS games like F.E.A.R, halo 1-3, Left 4 Dead the AI is competent enough to kick even the most skilled player's. In said games it's actually a challenge and it's worth it to keep playing and punishing yourself to get better at the game and the feeling of accomplishment when you finally "win" is the best feeling playing a video game imho.

I'm no game designer and i don't know what needs to happen in the background in order for an AI to be smart. If they can take a couple or many years to build areally good chess ai, they can do the same with a video game. It takes 2 or 3 years or even more? so what? Most games need that much time to be created, polished and tested anyways.

One fact remains solid for turn based games and that is they are using a copy and pasted AI from 24 years ago. Why not improve on it? Probably because they are looking for a quick buck which is fine, we all need to eat. Just don't expect people to buy your crap if you don't improve and evolve your products better suited to the times.

I'm not looking for perfection i'm looking for a challenge. If ai's are dumb there won't ever be an element of challenge outside beating the many OP neutrals and monsters laying about the shards/map. Honestly if the AI could mass produce endgame units and
temple units even dumb as it is, it would be a challenge or even near impossible to kill. The ai's have no upkeep and infinite mana that they don't even use to their fullest advantage. Even if you "taught" the ai to auto dispel any unit with magic buffs that would make the game that much more challenging. Outside of 50 turns if the ai doesn't kill you, you win so no point playing afterwards.
 
Last edited:
TheDeaddude said:
I guess i let my anger cloud my judgement and refused to look at what they did do as opposed to what they didn't do.
I'll give you that much, but if you look at FPS games like F.E.A.R, halo 1-3, Left 4 Dead the AI is competent enough to kick even the most skilled player's. In said games it's actually a challenge and it's worth it to keep playing and punishing yourself to get better at the game and the feeling of accomplishment when you finally "win" is the best feeling playing a video game imho.
Click to expand...
Now, I'm not really an avid FPS player, but I've played my fair share. Typically they way they "improve" the AI for the higher difficulties, is making them more resistant to damage, deal more damage, or more accurate. While yes, that makes the game harder, it doesn't constitute intelligent AI IMO.

Now there are things that I think could be improved in W2, such as actually enchanting their units with combat enchantments, or dispelling the ones on opponents, as you said. Though, these types of things would massacre players who don't fully understand the game mechanics. I know for my first several matches in W1, I didn't ever really use enchantments, as I didn't realize how good they would actually be. If the AI had actually used them, I would have lost those matches. Now, that's fine with me, but there are many players who would get put off by that, and pissing off customers is not a way to get people to buy your games.

Currently, what I do to make the game more challenging, is don't allow myself to use enchantments or temple units. And don't upgrade my units to their tier 2 versions until well after the AI has. I'm finding it quite fun to fight an AI who outnumbers me, with higher quality units.
 
Shidan said:
Now, I'm not really an avid FPS player, but I've played my fair share. Typically they way they "improve" the AI for the higher difficulties, is making them more resistant to damage, deal more damage, or more accurate. While yes, that makes the game harder, it doesn't constitute intelligent AI IMO.

Now there are things that I think could be improved in W2, such as actually enchanting their units with combat enchantments, or dispelling the ones on opponents, as you said. Though, these types of things would massacre players who don't fully understand the game mechanics. I know for my first several matches in W1, I didn't ever really use enchantments, as I didn't realize how good they would actually be. If the AI had actually used them, I would have lost those matches. Now, that's fine with me, but there are many players who would get put off by that, and pissing off customers is not a way to get people to buy your games.

Currently, what I do to make the game more challenging, is don't allow myself to use enchantments or temple units. And don't upgrade my units to their tier 2 versions until well after the AI has. I'm finding it quite fun to fight an AI who outnumbers me, with higher quality units.
Click to expand...
This thing about enchantments not being used by the AI puzzles me. Seems like an easy thing to program.
If (storedGold > 100 & manaIcome>0 & goldIncome>0) then
pickRandomUnitAmongTopThreeBestUnits(unit.giveEnchantment)
End if

:) Its not quite that easy, but you get my point.
 
For AI to be competitive in 1UPT game like this, it does not need just bonuses to gold and mana. That just gives more cannon fodder.

They need bonuses that would increase unit strength per tile. Bonuses to unit experience gain, extra experience levels a start, free perks. As well as ability to quickly get higher tier of units (which actually does need some AI programming and not ramming extra bonuses). Anyway, AI army needs to simulate that was "hardened" veteran army, something that player army becomes after some time, and AI army will never become in regular play, since it will always lose to human player or mismanagement, so all troops will be rookies.
 
Well crappy AI's are the norm for just about every game you can think of.
Click to expand...

I think the key is to design the game for the AI. A bit like board games, the AI play his own game which is different than the player for the only reason of challengin and threathning the player. But they don't use make the same decisions, they do not play the same game. This kind of mechanism is used in coop board games.

That is a technique I intend to use for my games. I would have done in a mod as events that power up the AI but I am not sure it is possible.

late 19th century
Click to expand...

WOW, I did not know video games existed in the 19th century. Television was not even invented.

So there are huge improvements in ai programming, but game designers seems to prefer "nice graphics"
Click to expand...

It's true that there have been improvement in AI, and the little development of TBS video game did not help make the "science" progress.

And they took forever and a half to develop to that point. And the rules of chess are pretty simple, compared to the complex systems running your average video game these days.
Click to expand...

Baiscally, chess AI are easy to make, you just calculate all the possibilities and take the best. Sure many refinement are possible to improve that. But video game is much more complex. In one of my game a colonist was moving back and forth all the time. Probably that it did not remember were it had to go, and got reasigned to a new task each turn instead of memorising the task.

But if there was no unit movement in the game, that would not have been an issue. So it is possible to change certain game mechanics to make it more AI friendly.

Take also not that most AI cheat in the background to make you believe that you have a fair challenge when in fact they are all actors plotting with each other to make you think you have a challenging AI.

-----------------------------------------

Anyways, I am curious to know if the DLC added new functionalities to modding?
 
If anything, the AI's in base game often conduct uncoordinated wars between themselves that result in no cities changing hands, wasting units for nothing. With the long periods of city siege, the game feels like less than the sum of its parts to me, although I haven't really mastered all its systems.
 
larienna said:
I think the key is to design the game for the AI. A bit like board games, the AI play his own game which is different than the player for the only reason of challengin and threathning the player. But they don't use make the same decisions, they do not play the same game. This kind of mechanism is used in coop board games.

That is a technique I intend to use for my games. I would have done in a mod as events that power up the AI but I am not sure it is possible.
Click to expand...
Ever played AI War? That game heavily relies on a different type of game being played by the human players and the AI players.

In any case, expecting a computer controlled opponent to behave like a human is not smart for the forseeable future. Outside simple games at least (and yes, for this discussion chess is simple). Scientists still don't quite understand what Intelligence is, so making something that behaves intelligently is going to be very tough in any kind of complex situation.
 
It's not the worst AI I've faced; it doesn't look terribly incompetent on its own but is just way too lunatic to do anything good. An ally that breaks an alliance for no sensible reason (in my view at the very least), declares war, doesn't even send a single unit for some 20ish turns just to realize that they're already fighting other AIs already, so perhaps we might as well just peace out and ally again a couple turns before to break it again and declare again doesn't seem to be uncommon. I personnaly would probably not do better if I really were to hold 4 wars simultaneously and aggressively.

There are many strategy games, but disapointingly, most fail in the most important part - the AI.
 
I don't think the AI fails... It's just not adapted to everyone. Personnaly I like the different opponents to have different characters and behaviours, their own personnality.

And if you want an AI that acts like a player... play against a player. The AI has not to be competitive. It needs to be a challenge, if by challenge we understand: something that makes the player willing to be better, and that could beat him. But the very good player will not find a serious opponent in a good AI. The serious opponent for a very good player needs to cheat or it needs an unfair game. (that's how it works in the Long War mod for X-COM, or in Civilization V).

Here the problem is that the AI doesn't have a personnality. They just don't have any ways to do something against the Nagas. I don't think they should all be able to restaure their terrain, but they should have different behaviours towards that problem... The AI doesn't even aknowledge there is a problem! That's not a good thing and I don't think we can find an other explanation than "The Ai is not the focus of the game. There is one because it needs one but it's not the true challenge of this game, which is more about finding a way to kill powerful units."
 
Making an actually competent AI in an open arena is nearly impossible these days. The only games that actually have hard AIs either have a very simplistic set of rules(say Chess) or are limited to small(usually turn based) skirmishes(say X-Com, Expeditions Conquistador, etc). The person above who said Civilization V has a tough AI made me laugh. Civilization V's AI is awfully incompetent. It's just that on Immortal and Deity(the last 2 difficulties) the AI receives so many bonuses and the player so many penalties that it's actually somewhat challenging to win, because the AI is actually capable of winning in a reasonable time frame making the player actually need a well refined strategy and execute it at least decently to successfully race the AI. This is the same case in FPS games, more bonuses for AI monsters, larger quantity of AI monsters and large penalties on the player. None of these things have anything to do with AI.

The thing is that for a company to make an actually good AI they need to first have employees that mastered the game in order to know how to program the AI to do well. Than invest a ton of man hours into it and go through a rigorous testing process. This process requires an already completed game before improving the AI and it will only work for that 1 game, other games from said company and even successor games(for example Warlock 1 to Warlock 2) will require massive adjustments to the previously perfected AI or even starting entirely from scratch. Considering the fact that most games are being released at the beta/early access a perfected AI would basically have to be in a DLC as the AI changes from beta/early access to release are minor and mostly do with fixing bugs and forcing the AI to use various available features it wasn't using, not with actually improving it to make it tougher. Now, consider how much time developers typically spend on making a DLC and how long it will take to perfect an AI for a game and you'll realize that it simply isn't financially feasible as the improved AI DLC would have to cost almost as much as the game itself.

Lastly you have the problem of CPU resources. Consider the first Chess AIs were developed on super computers and only recently(after decades of development) moved to less resource intensive CPUs like smart phones, an AI for a single game with an average shelf life of around 5 years will never reach the optimization the Chess AI has now and even if it somehow becomes optimized it will still require a lot more resources than a smart phone as all strategy games on the market are significantly more complex than Chess. So even if you ignore all the other problems, a good workable AI optimized(which I must stress again, is extremely unlikely) for a game will double if not more the resource requirements of the game, and games these days already have pretty steep resource requirements.

So as much as I personally would love to have a good challenging AI for strategy games, I realize the current impossibility of it happening. Thus the best we can hope for is a gross AI bonuses+player penalties system like in Civilization V, as that's the most we're currently going to get. Personally I hate this type of system but it's the only way so yeah...
 
You must log in or register to reply here.
Top Bottom
Back