Heroes 4. Player versus Comp difficulty

The old Heroes games developed by New World Computing. Please specify which game you are referring to in your post.
ywhtptgtfo
Hunter
Hunter
Posts: 528
Joined: 06 Jan 2006

Heroes 4. Player versus Comp difficulty

Unread postby ywhtptgtfo » 16 Jan 2006, 02:27

Hi. I've been playing H4 again lately and played single player scenarios under the "Champion" difficulty. The games were exceedingly long because of the creeps and the comp players had pathetic AI. Is it supposed to be like this? Should I play single player at all in this game?

User avatar
Angelspit
CH Founder
CH Founder
Posts: 6720
Joined: 18 Nov 2005
Location: Angelspit
Contact:

Unread postby Angelspit » 16 Jan 2006, 03:26

Well, sorry to disappoint you, but multiplayer is even slower because of a weak data transfer routine.

Lowering the difficulty setting will, in some cases, speed up the game by reducing the size of the neutral stacks. It will also make the AI's life easier, and it might even succeed in launching an attack against you.

I usually played at Advanced.
I'm on Steam and Xbox Live.

ywhtptgtfo
Hunter
Hunter
Posts: 528
Joined: 06 Jan 2006

Unread postby ywhtptgtfo » 16 Jan 2006, 05:01

I don't mind the speed of the game, but I want some challenge. The H3 AI was a pain in an ass to compete with (well probably not to pros) but the H4 AI sucked. I started another game with intermediate difficulty and I dominated the map without breaking a sweat. That's pretty disappointing...

User avatar
DaemianLucifer
Round Table Hero
Round Table Hero
Posts: 11282
Joined: 06 Jan 2006
Location: City 17

Unread postby DaemianLucifer » 16 Jan 2006, 06:16

Well,the AI in HIV is the worst one ever.And unfortunatelly,nothing can be done about that.Equi team tried to upgrade it a bit,and succeded,but its just a tiny bit.

User avatar
Corribus
Round Table Knight
Round Table Knight
Posts: 4994
Joined: 06 Jan 2006
Location: The Duchy of Xicmox IV

Unread postby Corribus » 16 Jan 2006, 06:17

ywhtptgtfo wrote:I don't mind the speed of the game, but I want some challenge. The H3 AI was a pain in an ass to compete with (well probably not to pros) but the H4 AI sucked. I started another game with intermediate difficulty and I dominated the map without breaking a sweat. That's pretty disappointing...
If it's challenge you want, you won't find it in H4. The computer player in H4 might as well be controlled by a brick for how dumb the AI is. If I'm not being lied to, there are some H4 mapmakers who have used complicated scripting to at least make the AI competitive in some maps. But I find H4 so trying that I don't even have the patience to try many more of these, as most of the highly recommended H4 maps that I have played were much less enjoyable than even mediocre maps made for H3.

It's a shame, because there's a great game somewhere in there hidden beneath it's very (very) rough edges.

But I'm a known H4-hater, so take my biased opinion for what's it's worth.
Corribus
"What men are poets who can speak of Jupiter if he were like a man, but if he is an immense spinning sphere of methane and ammonia must be silent?" - Richard P. Feynman

User avatar
pepak
Round Table Knight
Round Table Knight
Posts: 195
Joined: 06 Jan 2006

Unread postby pepak » 16 Jan 2006, 06:35

DaemianLucifer wrote:Well,the AI in HIV is the worst one ever.And unfortunatelly,nothing can be done about that.Equi team tried to upgrade it a bit,and succeded,but its just a tiny bit.
You can't do anything with the AI itself, but a good mapmaker can create a map where the lack of intelligence won't be evident.

Besides, I believe H4 AI is almost exactly the same as in H3, except that the fewer choices of H3 (far less important hero development, fewer artifact combos, smaller battlefield...) meant that you would rarely see the stupidity in action. Still, it was possible - consider the battles against 3500 Nagas in the Faerie Dragon scenario of the dragon hunt campaign in Armageddon's Blade.

User avatar
Corribus
Round Table Knight
Round Table Knight
Posts: 4994
Joined: 06 Jan 2006
Location: The Duchy of Xicmox IV

Unread postby Corribus » 16 Jan 2006, 06:40

pepak wrote:Besides, I believe H4 AI is almost exactly the same as in H3...
Huh? :|

The H3 AI had deficiencies, but they're not even in the same galaxy of ineptitude.

Corribus
"What men are poets who can speak of Jupiter if he were like a man, but if he is an immense spinning sphere of methane and ammonia must be silent?" - Richard P. Feynman

User avatar
DaemianLucifer
Round Table Hero
Round Table Hero
Posts: 11282
Joined: 06 Jan 2006
Location: City 17

Unread postby DaemianLucifer » 16 Jan 2006, 06:44

Well there are just a few maps that offered real chalenge.And they took every last drop of mental energy of the mapmaker to make them.Sure,the AI in HIII was bad in some situation.But it would never let an army twice as weak in every aspect to beat it.It couldve been exploated in some ways,but very few of them.And it offered quite a chalenge throughout the whole course of the map,not just during the begining,even in some mediocre maps.

User avatar
Ururam Tururam
Scout
Scout
Posts: 163
Joined: 06 Jan 2006
Location: the Void
Contact:

Unread postby Ururam Tururam » 16 Jan 2006, 08:12

Corribus wrote:If it's challenge you want, you won't find it in H4.
Thy the campaign by Wimfrits called "A Wind of Thorns" on hard or higher, or my "Angels are back" or "New Found Land" scenarios on Expert or higher and then dare to say that again! :devil:
Hoc est opus!
UT homepage: http://urtur.webpark.pl

User avatar
pepak
Round Table Knight
Round Table Knight
Posts: 195
Joined: 06 Jan 2006

Unread postby pepak » 16 Jan 2006, 10:30

Corribus wrote:The H3 AI had deficiencies, but they're not even in the same galaxy of ineptitude.
Not really. H3 had pretty much the same issues as H4, except that they weren't so visible because H3 is much "smaller". For example:

- The much larger H4 battlefield makes tactics such as Slow + Poison + Run-Like-Hell-Until-The-Opponent-Dies quite easy. H3 don't seem to have the same problem, until you try it against slow creatures on a terrain with some kind of a rift - then you will notice that the AI behavior is exactly as stupid in H3 as in H4. Exacmple: The two fights against 3500 Nagas in the third (Faerie Dragon) scenario of the dragon hunting campaign in Armageddon's Blade.

- Also, fights against AI armies with a hero in them tend to be much harder in H3 than in H4. But once again it can be shown that the behavior is pretty much identical - except that the immortality of heroes in H3 lets them survive much longer (thus causing more damage through their bonuses).

- H3 has more balanced heroes than H4, simply because almost all H3 skills work "automatically". In H4, you have to adapt your tactics to your skills (or better yet, choose your skills to better match your tactics). Since the AI seems to more or less automatically go for a "preferred" skill (with the levels of preferrence being set in stone by the programmers), it's not quite surprising that H4 AI heroes are weak opponents compared to H3 heroes. You can see that the problem exists in H3, though, as soon as you try one of the skills that require some specific action (e.g. Tactics) - the computers starts behaving in a fixed pattern which you can easily exploit (just like in H4).

- The same is true for artifacts. The H3 artifacts are pretty much all the same, differing only in the actual bonus to the primary skills. It's not that difficult to write a code which will select the highest ranking artifact and give it to the hero to use. In H4, it is far less obvious whether a Sword of Quickness (+1 attack) is more useful than a Sword of Gods (+30% attack, I believe) or even a Wand of Healing. It seems the AI's choice of artifacts to use is largely dictated by their "general value", just like in H3 - and just like it was the case with skills. I mean, how many times did you see an AI hero to use a Book of Air and the supermage hat (don't remember its name) on the adventure screen to Fly/DD next to you, then exchange them for a Heavenly Helm and Book of Earth and attack you with them?

There are many more such examples. What I said about artifacts is pretty much true for the monsters, too - AI rarely makes use of special unit abilities, both in H3 and in H4, except when the abilities are automatic (require no change of tactics to use effectivelly). I am almost certain that for any example of stupid AI behavior in H4 I could show that this very same behavior is present in H3 - except that its impact is reduced by the reduced scale of H3.
DaemianLucifer wrote:Sure,the AI in HIII was bad in some situation.But it would never let an army twice as weak in every aspect to beat it.
You mean, something like allowing a hero with 20 Titans to annihilate 3500 Naga Queens? I am sure H3 AI would never permit that! :-)

User avatar
Psychobabble
Spectre
Spectre
Posts: 706
Joined: 06 Jan 2006
Location: Melbourne, Australia
Contact:

Unread postby Psychobabble » 16 Jan 2006, 10:55

Corribus wrote:
ywhtptgtfo wrote:I don't mind the speed of the game, but I want some challenge.
That's simply not true. Maybe it's a fair comment on the game/maps at the time of release, it's not at all accurate for the current crop of good maps. Check my map picks page for recommended SP maps, but you can certainly find a good challenge in this game.

User avatar
Gaidal Cain
Round Table Hero
Round Table Hero
Posts: 6972
Joined: 26 Nov 2005
Location: Solna

Re: Heroes 4. Player versus Comp difficulty

Unread postby Gaidal Cain » 16 Jan 2006, 11:37

ywhtptgtfo wrote:Hi. I've been playing H4 again lately and played single player scenarios under the "Champion" difficulty. The games were exceedingly long because of the creeps and the comp players had pathetic AI. Is it supposed to be like this?
First of all, Champion is not playable, because the stacks are increased as much for the ai as for you- which means that the AI is crippled.

Second of all, the H4 adventure map AI sucks. There can be no question of this. It's far less agressive than the H3 one, and the fact that it at least manages to build heroes that aren't totally sucky doesn't help much(H3 AI grabs whatever it can, H4 builds starting skill and combat). It's also very bad a town building, and tends to build whatever it can- Nagas and Ogre Mages for example (which was clearly inferior pre-equilibris). It also doesn't make use of Immortality potions, which is bad. To round it off, it uses Quick Combat all the time. If you've ever tried it, you know that it means wastly higher losses than what should be possible.

The good news is that the combat AI is good, at least on par with the one in H3- as long as it doesn't have to defend in a siege. This means that the game can be a good challenge on a SP map, as long as it isn't designed to be played anything like a MP map.
You don't want to make enemies in Nuclear Engineering. -- T. Pratchett

User avatar
DaemianLucifer
Round Table Hero
Round Table Hero
Posts: 11282
Joined: 06 Jan 2006
Location: City 17

Unread postby DaemianLucifer » 16 Jan 2006, 12:36

No,combat AI isnt good neither.Heres just one example:

One lord,level 5 with 150+ genies against about 100 skeletons and 100 zombies.Quick combat results:Lord dies,combat is won.Normal combat result:Lord survives,combat is won.The only thing it does better is that it doesnt always attack the weakest unit first.(although it tends to attack summons over other units)
pepak wrote: You mean, something like allowing a hero with 20 Titans to annihilate 3500 Naga Queens? I am sure H3 AI would never permit that! :-)
Sure,you can do it with lots of spells,but I was talking more about balanced units:if you have 20 titans,level 20 hero,100 genies,etc and AI has 40 titans,level 40 hero,200 genies,etc,you still can win in HIV.

User avatar
Gaidal Cain
Round Table Hero
Round Table Hero
Posts: 6972
Joined: 26 Nov 2005
Location: Solna

Unread postby Gaidal Cain » 16 Jan 2006, 12:48

DaemianLucifer wrote:No,combat AI isnt good neither.Heres just one example:

One lord,level 5 with 150+ genies against about 100 skeletons and 100 zombies.Quick combat results:Lord dies,combat is won.Normal combat result:Lord survives,combat is won.The only thing it does better is that it doesnt always attack the weakest unit first.(although it tends to attack summons over other units)
That's not really the combat AI. That's quick combat, which is a whole other story. The combat AI, as in what controls the enemy when you control your units yourself, is competent enough. It's not great, but it does give you a run for your money.
You don't want to make enemies in Nuclear Engineering. -- T. Pratchett

User avatar
DaemianLucifer
Round Table Hero
Round Table Hero
Posts: 11282
Joined: 06 Jan 2006
Location: City 17

Unread postby DaemianLucifer » 16 Jan 2006, 13:05

Gaidal Cain wrote:
DaemianLucifer wrote:No,combat AI isnt good neither.Heres just one example:

One lord,level 5 with 150+ genies against about 100 skeletons and 100 zombies.Quick combat results:Lord dies,combat is won.Normal combat result:Lord survives,combat is won.The only thing it does better is that it doesnt always attack the weakest unit first.(although it tends to attack summons over other units)
That's not really the combat AI. That's quick combat, which is a whole other story. The combat AI, as in what controls the enemy when you control your units yourself, is competent enough. It's not great, but it does give you a run for your money.
Ok,Ive just made a small test.Ive given the computer a level one lord and 144 genies,and Ive given myself 150 skeletons and 150 zombies,and heres how the combat went:

Round one:Genies cast myrth on themselves,lord walks towards my army,my armycreeps forward.

Round two:Genies cast ice bolt and kill my skeletons,lord attacks zombies and dies.

Round three and four:two ice bolts and no more zombies.

So quick combat is exactly the same as regular combat.

User avatar
Gaidal Cain
Round Table Hero
Round Table Hero
Posts: 6972
Joined: 26 Nov 2005
Location: Solna

Unread postby Gaidal Cain » 16 Jan 2006, 13:38

:no: The fact that you have found one situation where the results are the smae doesn't mean that they'll always be it. If you wish a better example: try having 50 vampires in one army, and 20 Ogre Mages or so. In quick combat, the Vampires would win, but loose some of their numbers (this would perhaps even be true if you controlled the vamps and actually tried to loose some of them), but in QC, you're going to loose vampires. Note: I haven't tested this exact situation, but one could easily see similar ones when playing death pre-equ and running around with only a stack of vampires. That stack would diminish in virtually QC fight, but it would be damn hard to loose one vampire even if one tried.
You don't want to make enemies in Nuclear Engineering. -- T. Pratchett

User avatar
Corribus
Round Table Knight
Round Table Knight
Posts: 4994
Joined: 06 Jan 2006
Location: The Duchy of Xicmox IV

Unread postby Corribus » 16 Jan 2006, 15:00

pepak wrote: Not really. H3 had pretty much the same issues as H4, except that they weren't so visible because H3 is much "smaller". For example:
None of your examples (the effects of artifacts on tactics, for example) have anything to do with the AI. Pure and simple, the adventure AI in H4 does not work. The combat AI, as pointed out, is better (except that it does some pretty peculiar, exploitable things in sieges), and if you do happen to meet up with a loaded enemy hero, you might just have an exciting battle, but the adventure AI is so bad that the virtues of the combat AI don't really matter.

Corribus
"What men are poets who can speak of Jupiter if he were like a man, but if he is an immense spinning sphere of methane and ammonia must be silent?" - Richard P. Feynman

User avatar
pepak
Round Table Knight
Round Table Knight
Posts: 195
Joined: 06 Jan 2006

Unread postby pepak » 16 Jan 2006, 15:17

Corribus wrote:None of your examples (the effects of artifacts on tactics, for example) have anything to do with the AI.
Oh yes, they do. They simply make the deficiencies, present in the series at least since H3, so much more visible in H4. Those aren't new issues - those are old issues made evident by the massive increase of complexity in H4.

User avatar
Corribus
Round Table Knight
Round Table Knight
Posts: 4994
Joined: 06 Jan 2006
Location: The Duchy of Xicmox IV

Unread postby Corribus » 16 Jan 2006, 15:32

Alright, one by one then:
pepak wrote: - The much larger H4 battlefield makes tactics such as Slow + Poison + Run-Like-Hell-Until-The-Opponent-Dies quite easy. H3 don't seem to have the same problem, until you try it against slow creatures on a terrain with some kind of a rift - then you will notice that the AI behavior is exactly as stupid in H3 as in H4. Exacmple: The two fights against 3500 Nagas in the third (Faerie Dragon) scenario of the dragon hunting campaign in Armageddon's Blade.
This is the combat AI, not the adventure AI. The H4 combat AI is much better than the adventure AI, as I've mentioned.
- Also, fights against AI armies with a hero in them tend to be much harder in H3 than in H4. But once again it can be shown that the behavior is pretty much identical - except that the immortality of heroes in H3 lets them survive much longer (thus causing more damage through their bonuses).
Again, combat AI.
- H3 has more balanced heroes than H4, simply because almost all H3 skills work "automatically". In H4, you have to adapt your tactics to your skills (or better yet, choose your skills to better match your tactics). Since the AI seems to more or less automatically go for a "preferred" skill (with the levels of preferrence being set in stone by the programmers), it's not quite surprising that H4 AI heroes are weak opponents compared to H3 heroes. You can see that the problem exists in H3, though, as soon as you try one of the skills that require some specific action (e.g. Tactics) - the computers starts behaving in a fixed pattern which you can easily exploit (just like in H4).
Combat AI again.
- The same is true for artifacts. The H3 artifacts are pretty much all the same, differing only in the actual bonus to the primary skills. It's not that difficult to write a code which will select the highest ranking artifact and give it to the hero to use. In H4, it is far less obvious whether a Sword of Quickness (+1 attack) is more useful than a Sword of Gods (+30% attack, I believe) or even a Wand of Healing. It seems the AI's choice of artifacts to use is largely dictated by their "general value", just like in H3 - and just like it was the case with skills. I mean, how many times did you see an AI hero to use a Book of Air and the supermage hat (don't remember its name) on the adventure screen to Fly/DD next to you, then exchange them for a Heavenly Helm and Book of Earth and attack you with them?
Hmm, combat AI again...
There are many more such examples.
Why don't you try providing one that demonstrates that the H4 adventure AI is exactly the same as the H3 adventure AI, as this is the part of the game in H4 that is severely deficient. I don't think you can.

Corribus
"What men are poets who can speak of Jupiter if he were like a man, but if he is an immense spinning sphere of methane and ammonia must be silent?" - Richard P. Feynman

User avatar
wimfrits
Round Table Knight
Round Table Knight
Posts: 2047
Joined: 06 Jan 2006
Location: Utrecht, the Netherlands

Unread postby wimfrits » 20 Jan 2006, 15:19

Corribus wrote:Why don't you try providing one that demonstrates that the H4 adventure AI is exactly the same as the H3 adventure AI
Wasn't the H4 adventure AI an exact copy of the H3 adventure AI?
I'm pretty sure I read that somewhere.

It does explain the AI's behavior on the adventure map, like inability to deal with mobile neutral stacks and threat areas of neutral stacks. Or AI armies walking around joyfully with tombstones instead of returning to a castle.
Are you suggesting coconuts migrate?


Return to “Heroes I-IV”

Who is online

Users browsing this forum: No registered users and 11 guests