HOMM3 AI Stuff..
Dean Christopher Farmer
Firstly I'd like to point out that I love this game, and am completely
addicted. Congratulations on yet another winner NWC. Many late nights
ahead....
However, here are some points I would like to mention.
Point1.
I know this has been discussed before, but it really does seem to me that
the Heroes 3 AI, even on the normal level is cheating. Someone mentioned
in an earlier post that the AI is now very good at buying all of its
creatures up to and including seventh level each week, well to me it seems
a little *too* good to me.
Unless I have missed something, It is not possible for a fully upgraded
castle to be self-sustainable, other than maybe a capital. Even if you
factor in all of the sites on the map that give extra gold each week,
2000 per day cannot buy all the creatures for a particular 7th level castle
particularly in the case of Tower/Castle and all the $5000 troops.
My example is the scenario Titan's Winter. Quickly in this scenario I found
myself in the position of owning an entire continent, whilst another CP
owned another continent. I had 3 fully upgraded Tower Castles on my
continent and owned 4-5 of each mine. Even with all of the town income,
map-site income and constant marketplace resource-to-gold income I was
struggling to buy every single creature at all 3 castles each week.
However, as I quickly learned when I moved onto his continent and fought
him, the computer player had no trouble in buying EVERY single creature
each week, including 12 Archangels (1 flagged POG and 4 castles.)
From browsing his continent, he had 4 castles (4K+2K+2K+2K = 10K Income)
and 2 Gold Mines. Thats 12K per day which is 84K per week. 12 Archangels
alone cost 60K. Factor in the Champions and you're up to76K. Factor in
the Druids and you're already over 84K. Yet he managed to purchase the
Crusaders/Griffons/Halberdiers/Marksman (all of them) from all 4 castles
each week. Not sure exactly how much more cash that is, but its a
lot extra.
Ok, am I missing something obvious here when I play this game, or is the
AI cheating?
Point 2:
Anyone else noticed how the formula for quick combat when the
computer is fighting stuff seems to be:
If A tougher than B
B dies
A remains unchanged
else
A dies
B remains unchanged
To me, this has become very evident from playing some games, and explains
why the AI cruises through many wandering stacks and reaches your base
with a decent army fairly quickly. It also ensures that an AI which is
doing well does not lose valuable armies eliminating other AI opponents.
I've seen AI superheroes defeat other AI superheroes then come for me
literally without a scratch..
Point 3:
How on earth does the wandering monsters joining algorithm work?
Picture this. You've built up a superhero, you've fought for several
hours to take control of most of the map. My hero:
17 Titans
50 ArchMages
58 Genies
43 Naga Queens
300 Gremlin Masters
some large amount of gargoyles
80 iron golems
I have one more enemy 'superhero' to defeat. I am after him and quite
confident of victory. Our stats are about the same and he has:
16 Anicent Behemoths
26 Cyclops
40 Thunderbirds
47 Ogres
similear large numbers of lower level units.
Then he walks into a stack of wandering Behemoths (not ancient) that
SHOULD have kicked his butt. I am very surprised to right click and see
that a HORDE of Behemoths just joined his tough hero.
Now I fight this hero, and face 55 Behemoths. Now my hero is tough, but
not THAT tough, and I lose quite badly.
How is it fair that a wandering army more powerful than the rest of his
entire stack (or close to it) should be able to join him. We both spent
the entire game building up our stacks, going into battle, building up
more, etc and then in one instant he grabs something which is equal to
everything I have. Now, I would not have a problem with this if he
already had 3xHP of the wandering stack, or similar, but clearly in this
case he didn't. In fact I think I worked out his stack had LESS HP than
the 55 Behemoths. I lost the game (4 hour game) because of this so
naturally now I am a little ticked. Maybe I will calm down and change
my mind on this one ... comments?
Any agreement/disagreement on the above?
Deano
--
---------------------------------------------------------
Dean Farmer.
5th Year Comms Eng/Comp. Science
RMIT University. Melbourne Australia.
--
---------------------------------------------------------
Dean Farmer.
5th Year Comms Eng/Comp. Science
RMIT University. Melbourne Australia.
Dean Christopher Farmer
Forgot to mention that, one week prior to me arriving on the AI's
island, there was a week of the ArchAngel. It said ArchAngel growth+5.
I counted up the number of these guys in his 'killer hero' and he
indeed had purchased all 20 extra Archangels. The CP, on top of buying
EVERYTHING must have had a spare 100K lying around. I eventually won
this scenario, and inspected all of his castles. None contained the Grail
and I severaly doubt every single one of his heroes had expert estates.
He would need to have about 8 heroes with expert estates to even begin
to explain his opulance anyway. He only appeared to have 4 heroes which
I killed (eventually.)
Deano
Who still thinks the AI cheats quite badly with money/buying things.
John Lee
you can see if a "wandering" stack will join you or not before you encounter
them.
Dean Christopher Farmer <d...@cs.rmit.edu.au> wrote in message
news:7db4an$m2s$1...@emu.cs.rmit.edu.au...
Bruno Wolff III
>
> the Druids and you're already over 84K. Yet he managed to purchase the
> Crusaders/Griffons/Halberdiers/Marksman (all of them) from all 4 castles
> each week. Not sure exactly how much more cash that is, but its a
> lot extra.
Is it possible that the computer player had found the Grail?
> Then he walks into a stack of wandering Behemoths (not ancient) that
> SHOULD have kicked his butt. I am very surprised to right click and see
> that a HORDE of Behemoths just joined his tough hero.
> Now I fight this hero, and face 55 Behemoths. Now my hero is tough, but
> not THAT tough, and I lose quite badly.
He might have had the diplomacy skill. That effects how things join. It is
also possible to set wandering monsters to always join.
John Lee
I've had superior forces join me too. With the visions spell you can see
whether they will join you or not. It works both ways. The computer plays
by the same rules as you except up through the hard level (and after that it
gets more resources). At normal level and easy it doesn't even build a
structure in its town everyday.
It seems that with every game, even if the game is fair, there is always
some complaint that the computer is cheating when it isn't.
> Point 3:
>
> How on earth does the wandering monsters joining algorithm work?
> Picture this. You've built up a superhero, you've fought for several
> hours to take control of most of the map. My hero:
>
> 17 Titans
> 50 ArchMages
> 58 Genies
> 43 Naga Queens
> 300 Gremlin Masters
> some large amount of gargoyles
> 80 iron golems
>
> I have one more enemy 'superhero' to defeat. I am after him and quite
> confident of victory. Our stats are about the same and he has:
>
> 16 Anicent Behemoths
> 26 Cyclops
> 40 Thunderbirds
> 47 Ogres
> similear large numbers of lower level units.
>
> Then he walks into a stack of wandering Behemoths (not ancient) that
> SHOULD have kicked his butt. I am very surprised to right click and see
> that a HORDE of Behemoths just joined his tough hero.
> Now I fight this hero, and face 55 Behemoths. Now my hero is tough, but
> not THAT tough, and I lose quite badly.
>
> How is it fair that a wandering army more powerful than the rest of his
> entire stack (or close to it) should be able to join him. We both spent
> the entire game building up our stacks, going into battle, building up
> more, etc and then in one instant he grabs something which is equal to
> everything I have. Now, I would not have a problem with this if he
> already had 3xHP of the wandering stack, or similar, but clearly in this
> case he didn't. In fact I think I worked out his stack had LESS HP than
> the 55 Behemoths. I lost the game (4 hour game) because of this so
> naturally now I am a little ticked. Maybe I will calm down and change
> my mind on this one ... comments?
>
> Any agreement/disagreement on the above?
> Deano
>
>
John Lee
monster appear throughout the map.
John Lee <joh...@rmrc.net> wrote in message
news:8U8K2.23$H85.2...@iagnews.iagnet.net...
> He might have picked them up and not bought them. With the visions spell
> you can see if a "wandering" stack will join you or not before you
encounter
> them.
>
> Dean Christopher Farmer <d...@cs.rmit.edu.au> wrote in message
> news:7db4an$m2s$1...@emu.cs.rmit.edu.au...
> > FOR POINT 1:
> > Forgot to mention that, one week prior to me arriving on the AI's
> > island, there was a week of the ArchAngel. It said ArchAngel growth+5.
> > I counted up the number of these guys in his 'killer hero' and he
> > indeed had purchased all 20 extra Archangels. The CP, on top of buying
> > EVERYTHING must have had a spare 100K lying around. I eventually won
> > this scenario, and inspected all of his castles. None contained the
Grail
> > and I severaly doubt every single one of his heroes had expert estates.
> > He would need to have about 8 heroes with expert estates to even begin
> > to explain his opulance anyway. He only appeared to have 4 heroes which
> > I killed (eventually.)
> >
> > Deano
> > Who still thinks the AI cheats quite badly with money/buying things.
> >
John Lee
John Lee <joh...@rmrc.net> wrote in message
news:8X8K2.24$H85.2...@iagnews.iagnet.net...
> Dean Christopher Farmer <d...@cs.rmit.edu.au> wrote in message
Martin Leslie Leuschen
<Liberal snippage>
Dean Christopher Farmer (d...@cs.rmit.edu.au) wrote:
: HI HOMM3 Fans,
: Point1.
: I know this has been discussed before, but it really does seem to me that
: the Heroes 3 AI, even on the normal level is cheating.
It does on the last two diff levels. According to the programmers who
occasionally slum on this ng, it doesn't for the first three. (A few
exceptions, but none of them involve gold, IIRC.)
: constant marketplace resource-to-gold income
AI does this too now. With enough mines, marketplaces, and resource
silos, it can be significant.
: From browsing his continent, he had 4 castles (4K+2K+2K+2K = 10K Income)
: and 2 Gold Mines. Thats 12K per day which is 84K per week.
Assume he also had 1 of each mine for each castle, and a resource silo in
each. Thats 28 mine-equivalents. Even at 100gp/resource (50 for wood/ore)
that's an extra 2800 per day, or 19.6K per week. IIRC, 4 marketplaces
actually is better than that.
Add in gold from heroes, magic items, map sites, and extra resources from
same converted to gold, and the income gets even bigger.
Extra note: if the AI has a Rampart anywhere, add in the effect of the
dwarven treasury. Nice toy, that.
: 12 Archangels
: alone cost 60K. Factor in the Champions and you're up to76K. Factor in
: the Druids and you're already over 84K.
Note: AAs- 60K
Champs- 16K
The lower level creatures are less expensive. It doesn't take much extra
gold (above) to buy them.
: Yet he managed to purchase the
: Crusaders/Griffons/Halberdiers/Marksman (all of them) from all 4 castles
: each week. Not sure exactly how much more cash that is, but its a
: lot extra.
I suspect that the AAs are more than half of his costs.
: Ok, am I missing something obvious here when I play this game, or is the
: AI cheating?
Depends on your Diff level, of course, but on "hard" or less, it isn't.
(At least, it's not suposed to be.) It's just using all it's resources
effectively.
: Point 2:
: Anyone else noticed how the formula for quick combat when the
: computer is fighting stuff seems to be:
: If A tougher than B
: B dies
: A remains unchanged
: else
: A dies
: B remains unchanged
Actually, my battles seem to go pretty much like this, and I have to play
them all out the long way. A 50% edge in power means a 90% edge in
casualties, in my experience.
: To me, this has become very evident from playing some games, and explains
: why the AI cruises through many wandering stacks and reaches your base
: with a decent army fairly quickly. It also ensures that an AI which is
: doing well does not lose valuable armies eliminating other AI opponents.
: I've seen AI superheroes defeat other AI superheroes then come for me
: literally without a scratch..
Easy enough to test this - set up an AI hero in the editor with no spell
book and a horde of slow walkers. (Walking dead or dwarves are s good
bet.) Place many hostile shooter stacks nearby, and give yourself an
observer hero to watch from a protected redwood nearby.
(There are a few useful items and spells that will tell you his exact
stack size.) He can kill a few titan with hordes of walking dead, but
he should take casualties.
: Point 3:
: How on earth does the wandering monsters joining algorithm work?
Not sure on exact details, but diplomacy really helps, and having
monsters from the same "building" (upgraded or not) also helps.
: Picture this. You've built up a superhero, you've fought for several
: hours to take control of most of the map. My hero:
: 17 Titans
: 50 ArchMages
: 58 Genies
: 43 Naga Queens
: 300 Gremlin Masters
: some large amount of gargoyles
: 80 iron golems
: I have one more enemy 'superhero' to defeat. I am after him and quite
: confident of victory. Our stats are about the same and he has:
: 16 Anicent Behemoths
: 26 Cyclops
: 40 Thunderbirds
: 47 Ogres
: similear large numbers of lower level units.
: Then he walks into a stack of wandering Behemoths (not ancient) that
: SHOULD have kicked his butt.
Maybe he checked with the visions spell? or was just desparate to
get away from you?
: I am very surprised to right click and see
: that a HORDE of Behemoths just joined his tough hero.
: Now I fight this hero, and face 55 Behemoths. Now my hero is tough, but
: not THAT tough, and I lose quite badly.
: How is it fair that a wandering army more powerful than the rest of his
: entire stack (or close to it) should be able to join him.
Perhaps they were swayed by his diplomatic tongue. Perhaps one of the
hero's Ancient Behemoths was the graddaddy of all the normal behemoths
and they joined out of respect. :) Perhaps that stack was *set* to
be friendly in the editor.
: We both spent
: the entire game building up our stacks, going into battle, building up
: more, etc and then in one instant he grabs something which is equal to
: everything I have. Now, I would not have a problem with this if he
: already had 3xHP of the wandering stack, or similar, but clearly in this
: case he didn't. In fact I think I worked out his stack had LESS HP than
: the 55 Behemoths. I lost the game (4 hour game) because of this so
: naturally now I am a little ticked. Maybe I will calm down and change
: my mind on this one ... comments?
Lesson: Expert diplomacy can be brutal.
: Any agreement/disagreement on the above?
: Deano
Martinl
michael wilson daniels
: EVERYTHING must have had a spare 100K lying around. I eventually won
: this scenario, and inspected all of his castles. None contained the Grail
: and I severaly doubt every single one of his heroes had expert estates.
: He would need to have about 8 heroes with expert estates to even begin
: to explain his opulance anyway. He only appeared to have 4 heroes which
: I killed (eventually.)
You might take a look at the scenario in the editor -- at least in homm2,
quite a lot of the single maps & campaign maps had weekly resource bonus
events that the computer players got. It might be that your map's author did
the same sort of thing. It's not really the AI cheating, but the map
designer giving the computer an advantage.
--
Michael W. Daniels | "Babylon 5...taught us that we had to create the
CS/Ling. Senior | future, or others will do it for us. It showed us
| that we have to care for each other, because if we
| don't, who will?"
Gus Smedstad
>
> <Liberal snippage>
> Dean Christopher Farmer (d...@cs.rmit.edu.au) wrote:
>
> : I know this has been discussed before, but it really does seem to me that
> : the Heroes 3 AI, even on the normal level is cheating.
>
> It does on the last two diff levels. According to the programmers who
> occasionally slum on this ng, it doesn't for the first three. (A few
> exceptions, but none of them involve gold, IIRC.)
>
> : constant marketplace resource-to-gold income
>
> AI does this too now. With enough mines, marketplaces, and resource
> silos, it can be significant.
You are correct. The AI uses the marketplaces to trade for resources -
especially gold. Once the castles are fully bought up, it can generate
a lot of extra cash by selling unneeded resources.
On the map in question, Purple has 1 alchemist lab, 2 sulfer Dunes, 2
Sawmills, 1 ore pit, 2 gem ponds, and 2 crystal mines, in addition to
the 2 gold mines. Add 4 resource silos, for +4 wood and +4 ore a day.
There's also a Water Mill, which is worth 1K a week, a Mystical Garden,
which gives 500 gold or 5 gems a week, and a Windmill, which gives 3-6
of a random resource each week.
Rough average extra income:
98 wood + ore / week
55 mercury + crystal + sulfur + gems / week
1250 gold / week
Wood + Ore trades at 62 gold with 4 markets, and others at 125 gold. So
extra income from this sort of thing is about 13K gold per week. Not
counting any income artifacts or special abilities.
Of course, if Purple happens to build the Grail, that's 35K more gold a
week right there.
> Actually, my battles seem to go pretty much like this, and I have to play
> them all out the long way. A 50% edge in power means a 90% edge in
> casualties, in my experience.
That's pretty much correct. The AI does use a shorter, simpler combat
method when there is no human involved, primarily to avoid possible
stalemates (example: attacker at a fort loses the catapult.) It's not
perfect by any means, of course.
Quick combat, however, uses the actual, regular combat logic, with no
animations. Unless the combat runs 20+ rounds, in which case it
finishes it with the above "fast" method.
> : Point 3:
>
> : How on earth does the wandering monsters joining algorithm work?
>
> Not sure on exact details, but diplomacy really helps, and having
> monsters from the same "building" (upgraded or not) also helps.
Correct. And no, the AI hasn't the slightest how diplomacy works. If
it touches a stack, it means it thinks it can win. Or its desperate.
> : How is it fair that a wandering army more powerful than the rest of his
> : entire stack (or close to it) should be able to join him.
If he was tough enough to take the stack, he had Expert diplomacy, and
he had similar creatures (which he did), there's a 50% chance the stack
will join. This is true for you, too.
> Lesson: Expert diplomacy can be brutal.
Yep!
- Gus
AI programmer and general handyman on Heroes III
New World Computing
Don't think that just because we're talking, I'm in customer service
:).
John DiFool
Dean Christopher Farmer wrote:
> HI HOMM3 Fans,
>
> Firstly I'd like to point out that I love this game, and am completely
> addicted. Congratulations on yet another winner NWC. Many late nights
> ahead....
> However, here are some points I would like to mention.
>snip<
I dunno about all that stuff, but the AI can definitely see thru the necromancer
shroud (plus the shroud dwelling on the adventure map...). Keep my killer
stack in town, AI runs away. Move killer stack c. 2-3 days march away
from town, AI moves toward my necropolis. VERY irritating...(esp. for
an Undead Fiend like me ;-).
John DiFool
Gus Smedstad
>
> You might take a look at the scenario in the editor -- at least in homm2,
> quite a lot of the single maps & campaign maps had weekly resource bonus
> events that the computer players got. It might be that your map's author did
> the same sort of thing. It's not really the AI cheating, but the map
> designer giving the computer an advantage.
Actually, this is quite rare in Heroes III, because it isn't required.
There is precisely one official map in Heroes III that gives the
computer a "care package" on the first day (Rumble in the Bogs), and
none that give it regular income. A couple of maps (Peacemaker,
Serpent's Treasure) give regular income but exclude it from AI players.
Titan's Winter, the map he's talking about, does give Purple a long time
to build up. However, once you get there, it's just a matter of time.
You have 7 castles to his 4, and he has no way to launch an offensive,
since you control all the shipyards.
The dicey part of that map is the first part, when you're fighting for
control of the 4 castles on the mainland.
- Gus
AI programmer and general handyman on Heroes III
New World Computing
Should I be caught or killed, New World will disavow any knowledge of
my actions.
Gus Smedstad
>
> I dunno about all that stuff, but the AI can definitely see thru the necromancer
>
> shroud (plus the shroud dwelling on the adventure map...). Keep my killer
> stack in town, AI runs away. Move killer stack c. 2-3 days march away
> from town, AI moves toward my necropolis. VERY irritating...(esp. for
> an Undead Fiend like me ;-).
This is one area in which the AI does cheat. Specifically, it ignores
the shroud when marking areas as "dangerous" because there's a powerful
hero there.
Otherwise, the shroud works. Perhaps too well - it should remember more
than it actually does.
- Gus
AI programmer and general handyman on Heroes III
New World Computing
Opinions expressed are mine, not NWC's, etc.
Stephen Lee
John Lee <joh...@rmrc.net> wrote:
>During the week of "any monster", a lot of wandering stacks of that said
>monster appear throughout the map.
Not week; month. Also, months of upgraded creatures never occur.
--
Boy I Love Losing Superbowls
Count On Losing This Sunday
Dorky Overrated Losers Play Hideously In Nasty Setback
Just End The Season
John M Clancy
play
> > them all out the long way. A 50% edge in power means a 90% edge in
> > casualties, in my experience.
>
> method when there is no human involved, primarily to avoid possible
> stalemates (example: attacker at a fort loses the catapult.) It's not
> perfect by any means, of course.
>
> Quick combat, however, uses the actual, regular combat logic, with no
> animations. Unless the combat runs 20+ rounds, in which case it
> finishes it with the above "fast" method.
>
Gus,
Why not use that method all the time? For computer player vs computer
player or computer player vs wandering monster stack? I to have wondered
why the AI never loses any creatures from battles. Even when attacking
large numbers of powerful shooters. They should lose 'some' troops in some
battles, even if it's just their low level stack.
If it's a question of the computer taking too long give us an option
to turn it on for higher end systems. I know my system could handle it.
John M Clancy
Gus Smedstad <g...@nwcomputing.com> wrote in message
news:36F97E24...@nwcomputing.com...
> Martin Leslie Leuschen wrote:
> >
> - Gus
> AI programmer and general handyman on Heroes III
> New World Computing
> :).
Gus,
You and George have done more for customer service for me
and the other posters of this newsgroup and fans of Homm than
some other 10 companies I could name. Trust me. :)
I'm glad NWC picked you up. Max was a cool game but I think
you were under-appreciated and under-utilized at Interplay.
Bruno Wolff III
>
> That's pretty much correct. The AI does use a shorter, simpler combat
> method when there is no human involved, primarily to avoid possible
> stalemates (example: attacker at a fort loses the catapult.) It's not
> perfect by any means, of course.
I have a question about this. How come there isn't some way out of human-human
battles that are effectively stalemates? If neither person wants to give,
this is going to be a very unsatisfying end to a game.
For human-computer stalemates, I am willing to give the computer the edge
and just lose the battle, but I see no reason to do this with another human
unless I will still clearly be winning if I give up on the battle.
John M Clancy
Bruno Wolff III <br...@cerberus.csd.uwm.edu> wrote in message
news:7ddo1e$ccp$1...@uwm.edu...
Maybe there should be a 50 move rule like in chess.
If the battle isn't over after 50 turns then both heroes
get dumped to the tavern. With their leftover respective
troops as if they had surrendered, but without the gold
cost.
Stalemates are pretty rare anyway. I've come across
maybe 5 cases of a stalemate with the computer and I
have been playing homm for a LONG time.
Bruno Wolff III
>
> If it's a question of the computer taking too long give us an option
> to turn it on for higher end systems. I know my system could handle it.
I would like to see more "realistic" combat as well. I would be very willing
to have pauses while resolving conflicts not involving humans and get more
reasonable battle outcomes.
It always irked me in HOMM2, that the computer could walk through ghosts with
stacks that should have left more ghosts than there were at the start of the
battle. There isn't any creature that has quite that impact anymore, but there
still is a rock paper scissors effect between unit types. And it would be nice
to see this reflected in all battles.
Another side effect will probably be more accurate spell point usage. I
haven't checked to see if leaders get charged spell points for battles not
involving humans. Even if it does, the amount charged may not be very
representative of how many would be used in particular battles.
Michael Kaspar
John M Clancy wrote in message <7ddr2r$j...@dfw-ixnews8.ix.netcom.com>...
>> I have a question about this. How come there isn't some way out of
>human-human
>> battles that are effectively stalemates? If neither person wants to give,
>> this is going to be a very unsatisfying end to a game.
>If the battle isn't over after 50 turns then both heroes
>get dumped to the tavern. With their leftover respective
>troops as if they had surrendered, but without the gold
>cost.
>
>Stalemates are pretty rare anyway. I've come across
>maybe 5 cases of a stalemate with the computer and I
>have been playing homm for a LONG time.
I don't play on the internet (yet), but perhaps you could elaborate on how a
human human stalemate might happen. I suppose if both heroes exhaust
spellpoints and neither side has shooters or fliers or balista this could
happen, but it seems there should still be someone with an advantage.
Another alternative is limit battles to 15 rounds or so (probably more than
you need except in the aformentioned stalemate situation), and if the battle
is not over both sides are "shackled" to each other, until the defender's
turn. At that point he can opt to retreat out of combat or stay for 15
rounds with him as the attacker. Additionally, if the defender has another
hero nearby with troops, the troops could be added to the foray by normal
procedures. This is similar to the "engaged" combat result from Battle of
the Bulge, a boardgame from Avalon Hill before they were disposed of by
Hasbro.
Michael J. Kaspar
ICQ 3025001
remove DONOTSPAM to reply thru email
anthony stuckey
>I don't play on the internet (yet), but perhaps you could elaborate on how a
>human human stalemate might happen. I suppose if both heroes exhaust
>spellpoints and neither side has shooters or fliers or balista this could
>happen, but it seems there should still be someone with an advantage.
A smaller faster group that doesn't want to get wiped out by a
bigger, slower group can run around the board essentially infinitely.
>Another alternative is limit battles to 15 rounds or so (probably more than
>you need except in the aformentioned stalemate situation),
Nope. Certain units have and can use more shots than that.
I used to do this all the time with Grand Druids etc versus very
slow monster stacks like Ogres and Dwarves. (non-upgraded)
>and if the battle
>is not over both sides are "shackled" to each other, until the defender's
>turn. At that point he can opt to retreat out of combat or stay for 15
>rounds with him as the attacker.
That makes some sense.
--
Anthony Stuckey stu...@uiuc.edu
System Administrator, students.uiuc.edu
George Ruof
> A smaller faster group that doesn't want to get wiped out by a
>bigger, slower group can run around the board essentially infinitely.
Hmmm... How would it be if after a certain number of rounds a unit
became tired and couldn't move as far? Would that allow another unit to
eventually catch them?
--
George Ruof
gr...@pacificnet.net
Senior Programmer
New World Computing
Christoph Nahr
wrote:
>Hmmm... How would it be if after a certain number of rounds a unit
>became tired and couldn't move as far? Would that allow another unit to
>eventually catch them?
For consistency you'd have to make the slower group tired as well.
The faster group would still have the edge. I think a chess-like
solution would be best: no attack (physical or spell) on any creature
for, say, ten rounds means that the attacking hero has lost the
battle. Coupled with a warning one turn before reaching the limit so
that you can surrender if you wish.
--
Chris Nahr (christo...@uumail.xxde, remove xx to reply by e-mail)
See my HoMM3 Manual Addenda at http://uuhome.de/christoph.nahr/
David Lam
I used to recall that in HOMM2 in I get a week of "PLAGUE", and when I tried
to reload the autosave from the day before, and try again, I would always
get "PLAGUE". Therefore, the outcome for the new week (whether it's Cyclops
Population +5, or Skeletons Doubles) is decided at the beginning of day 7
instead of start of Day 1. This prevents cheating if someone gets a bad
result.
In HOMM3 , this is different, and allows for cheating. IT is too tempting to
just load the autosave and click "end turn" again if I get a bad result.
It would be good if such results are predetermined and saved into the save
game.
Dave
QurqirishD
>for, say, ten rounds means that the attacking hero has lost the
>battle. Coupled with a warning one turn before reaching the limit so
>that you can surrender if you wish
I disagree with this. If you put in this rule, than a single defending
archangel can beat 10,000 ogres, simply because they are slow. In fact, any
number of stacks of creatures with speed less than 5 can be defeated by a
single creature of speed 16+ (or even less if that creature flies/teleports and
has obstacles on the battlefield)
The Qurqirish Dragon, of the Xanadu Dragons
--==<<{{ UDIC }}>>=--
Remember- my address is no laughing matter
Alexander E. Richman
was their stalemate resolution (of course, they were more common than would ever be
in HOMM). If the battle lasts a long time (say, 50 turns), or the 10 turn without
attack suggestion, both heroes withdraw, that is, the attacker and defender both
return to the positions which they held prior to the attack. At some later time,
the attack can be reinitiated, such as after a visit to a magic well (if it hadn't
been used recently). Anyway, I personally think this is the fairest resolution to
a true standoff.
Alex Richman
Bobstrosity
>> Crhsitopher Nahr (i think) :no attack (physical or spell) on any creature
>>for, say, ten rounds means that the attacking hero has lost the
>>battle. Coupled with a warning one turn before reaching the limit so
>>that you can surrender if you wish
>
>I disagree with this. If you put in this rule, than a single defending
>archangel can beat 10,000 ogres, simply because they are slow. In fact, any
>number of stacks of creatures with speed less than 5 can be defeated by a
>single creature of speed 16+ (or even less if that creature flies/teleports
and
>has obstacles on the battlefield)
>The Qurqirish Dragon, of the Xanadu Dragons
>--==<<{{ UDIC }}>>=--
>Remember- my address is no laughing matter
A proposed solution to end stalemates, after 25 (or whatever) rounds,
have the creature "overcome with battle lust" and everything gets maximum
morale (overriding all artifacts and stuff), and have "Chronos the wizard of
time" cast a
"remove obstacles expert level (and earthquake)" and a
"expert level mass haste" from afar, his castle in EverWhen.
That ought to do it. ( Of course undead don't get the morale, but the hast
should help,
and with no obstacles, they should be able to attack)
Presumably, a stalemate occurs when a strong stack cannot catch a weak
stack.
This should help.
Or have disease affect all units, and let them slowly die off (so the weak
stackk dies off
first, which is the way it should be)
Cheers,
bob
Bruno Wolff III
>
> I don't play on the internet (yet), but perhaps you could elaborate on how a
> human human stalemate might happen. I suppose if both heroes exhaust
> spellpoints and neither side has shooters or fliers or balista this could
> happen, but it seems there should still be someone with an advantage.
If one player's creatures cannot engage the other's, because of speed or
castle walls, and the other player does not want to engage, then the battle
is a stalemate.
It is also possible for units to be able to reach each other, but not be
able to kill each other. This can happen with small numbers of high hit
point units and tents. Trolls used to be the classic case in HOMM2. I think
there is something that regenerates in HOMM3 as well.
It is also possible to have extremely long battles that will eventually be
won by one side or the other, but people may get offly board sitting through
them. A 1000 round battle wouldn't be that hard to contruct with a good
(stats of 20) leader involved.
Bruno Wolff III
with each round of combat taking a small amount of the attackers movement
points. When he runs out of movement points the battle breaks off.
I think this might improve another aspect of play where a leader just barely
reaches another leader or castle with its movement. The leader would then
need to win a quick battle or the battle end without conclusion.
I have suggested this before, but NWC didn't seem to like the idea.
Christoph Nahr
<dejast...@my-dejanews.com> wrote:
[The Qurqirish Dragon whose post has not yet appeared on my server
wrote...]
>>I disagree with this. If you put in this rule, than a single defending
>>archangel can beat 10,000 ogres, simply because they are slow. In fact, any
>>number of stacks of creatures with speed less than 5 can be defeated by a
>>single creature of speed 16+ (or even less if that creature flies/teleports
>and
>>has obstacles on the battlefield)
You are right but I don't see the problem. If you try to attack an
archangel with only one stack of slow walkers (if you had seven then
you could eventually hit him somewhere, right?) and no Haste/Teleport
or attack spells then I think you have only yourself to blame. You
simply shouldn't do things like that.
Derek Benson
III) wrote:
Uh, I don't want to offend, but my expert diplomacy is turned off at
the moment, and I think NMC probably didn't like this idea because
it's dumb; or at least totally unworkable in the game. In the late
game, when I have a powerful hero/army and the computer likewise, then
I have to plan out parking my hero near a castle/enemy hero and
attacking him on the next turn, just to be sure that my hero has
enough movement points to be able to fight X number of combat rounds?
This is totally ridiculous.
Pertaining to other comments on this thread, the combat is fine for
single player against the computer; it's practically impossible to get
into a stalemate, although I've done this in HOMM 2, walking around in
a circle with my last unit to stay away from the comp's last slower
unit. Possibly this may be a problem in multiplayer games, but I
certainly hope that drastic changes are not made to the combat for
single player. All stacks getting slower each combat round because
they get tired, also a bad idea.
-Derek
Bruno Wolff III
> stu...@staff.uiuc.edu (anthony stuckey) wrote:
>
>> A smaller faster group that doesn't want to get wiped out by a
>>bigger, slower group can run around the board essentially infinitely.
>
> became tired and couldn't move as far? Would that allow another unit to
> eventually catch them?
That won't help the cases where units can't do enough damage to kill units
and there is healing going on or if one person is behind intact walls with
no archers and the catapult has been destroyed.
Assuming you don't want to have a battle end with no result (since there isn't
a way to do that), I would suggest just having one side or the other win
after a fixed number of rounds. You could allow the loser one last chance
to retreat or surrender if applicable.
As to who wins, you could use one of the following:
Defender Wins
Attacker Wins
Most hit points in creatures wins, with one of attacker or defender winning
ties
You could probably pick a limit of around 100 rounds.
Bruno Wolff III
] On 26 Mar 1999 17:28:08 GMT, br...@cerberus.csd.uwm.edu (Bruno Wolff
] III) wrote:
]
]>My thought on battle stalemates is that this issue would be best resolved
]>with each round of combat taking a small amount of the attackers movement
]>points. When he runs out of movement points the battle breaks off.
]>
]>I think this might improve another aspect of play where a leader just barely
]>reaches another leader or castle with its movement. The leader would then
]>need to win a quick battle or the battle end without conclusion.
]>
]>I have suggested this before, but NWC didn't seem to like the idea.
]
] Uh, I don't want to offend, but my expert diplomacy is turned off at
] the moment, and I think NMC probably didn't like this idea because
] it's dumb; or at least totally unworkable in the game. In the late
] game, when I have a powerful hero/army and the computer likewise, then
] I have to plan out parking my hero near a castle/enemy hero and
] attacking him on the next turn, just to be sure that my hero has
] enough movement points to be able to fight X number of combat rounds?
] This is totally ridiculous.
Why is that totally ridiculus? Anyway, a fixed upper limit to the number of
rounds that wasn't too large (say 100) would also work. The tough part is
deciding what to do when the limit is reached. I would prefer to see the
attack end with both leaders living (unless one only has troops that go
away after the battle). That isn't currently a valid result and may not
be something that could be easily added.
]
] Pertaining to other comments on this thread, the combat is fine for
] single player against the computer; it's practically impossible to get
] into a stalemate, although I've done this in HOMM 2, walking around in
] a circle with my last unit to stay away from the comp's last slower
] unit. Possibly this may be a problem in multiplayer games, but I
] certainly hope that drastic changes are not made to the combat for
] single player. All stacks getting slower each combat round because
] they get tired, also a bad idea.
It isn't that hard to get into stalemates with small forces. This isn't
seen as a problem in single player, because you, the human get screwed.
In HOMM 2 the computer AI never tried to get draws. I don't know whether
or not that is true in HOMM 3. When there are two humans playing, there
isn't any particular reason to screw one over the other. If the game is
close, neither person may want to give. So a close game may end in a draw
because of a relatively minor battle. This doesn't sound very satisfying
to me.
Sherilyn
I am very glad to see New World Computing is having someone help people with
text support in the newsgroups. I think it is great that your company
allows it.
I read in many prerelease talk that one of the options was in the new
HOMMIII was going to be a global buy all creature command. Once you get
many castles going, and you have plenty of gold, buying all your creatures
every week, well can be just quite boring. Was this command ever
implemented? I looked for it but couldn't find it :(.
Thanks for your help,
please respond by sending email to sher...@iag.net
Sherilyn
Gus Smedstad <g...@nwcomputing.com> wrote in message
news:36F98530...@nwcomputing.com...
> John DiFool wrote:
> >
> > I dunno about all that stuff, but the AI can definitely see thru the
necromancer
> >
> > shroud (plus the shroud dwelling on the adventure map...). Keep my
killer
> > stack in town, AI runs away. Move killer stack c. 2-3 days march away
> > from town, AI moves toward my necropolis. VERY irritating...(esp. for
> > an Undead Fiend like me ;-).
>
> This is one area in which the AI does cheat. Specifically, it ignores
> the shroud when marking areas as "dangerous" because there's a powerful
> hero there.
>
> Otherwise, the shroud works. Perhaps too well - it should remember more
> than it actually does.
>
> - Gus
> AI programmer and general handyman on Heroes III
> New World Computing
Stephen Lee
Bruno Wolff III <br...@cerberus.csd.uwm.edu> wrote:
[snip]
>It is also possible for units to be able to reach each other, but not be
>able to kill each other. This can happen with small numbers of high hit
>point units and tents. Trolls used to be the classic case in HOMM2. I think
>there is something that regenerates in HOMM3 as well.
Yep, wights/wraiths regenerate. If it is a tent that is causing the
battle to be protracted, though, you can just take a turn or two to trash
the tent.
Matt Kracht
>That's pretty much correct. The AI does use a shorter, simpler combat
>method when there is no human involved, primarily to avoid possible
>stalemates (example: attacker at a fort loses the catapult.) It's not
>perfect by any means, of course.
Why not just have the attacker flee when his catapult is destroyed?
It makes a whole lot of sense to me. If my catapult was destroyed and
all my flyers and ranged units were killed, I'd have to flee.
In fact, I think I'll try this next time I'm under seige. Catapults
have 1000 hit points, if I remember correctly, so it shouldn't take
much more than 6 Archangels to do in a catapult in one turn (assuming
you have a good Attack). Or maybe a horde of Rocs.
>If he was tough enough to take the stack, he had Expert diplomacy, and
>he had similar creatures (which he did), there's a 50% chance the stack
>will join. This is true for you, too.
Wow. I wish I got diplomacy as an option more often.
Christoph Nahr
<ask.me.if.you...@none.com> wrote:
>I read in many prerelease talk that one of the options was in the new
>HOMMIII was going to be a global buy all creature command. Once you get
>many castles going, and you have plenty of gold, buying all your creatures
>every week, well can be just quite boring. Was this command ever
>implemented? I looked for it but couldn't find it :(.
This would be nice but I found that it's not strictly necessary
because I can rarely ever afford buying all creatures at once.
Usually I have to make a decision how to spend my money.
dar...@my-dejanews.com
<C7132130A76C8D51.5042F873...@library-proxy.airnews.net
>, gr...@pacificnet.net (George Ruof) wrote:
> > A smaller faster group that doesn't want to get wiped out by a
> >bigger, slower group can run around the board essentially infinitely.
> Hmmm... How would it be if after a certain number of rounds a unit
> became tired and couldn't move as far? Would that allow another unit to
> eventually catch them?
This should apply equally to the pursuers, which doesn't solve the problem.
My suggestion: if nobody dies for 10 rounds in a row, the battle is declared
a stalemate, neither hero receives XP for kills, and the game reverts to the
adventure map with the heroes standing next to each other. Effectively this
blocks hero movement without loss of life; fine with me (I don't think it's
terribly unbalancing for a serpent fly to be able to distract a stack of
diamond golems indefinitely). Also if this applies to "wandering" (now
*there's* a misnomer) monsters, the next time they are attacked, they should
have a chance of coming up with a different formation. More spread-out
formations will be able to "corner" the speedy critters. This especially
applies to heroes; the big slow stack should be spread out to several stacks
to catch the "mouse".
Darryl.
--
Just my opinion.
This account is a spam-filter; real email d.greensill<AT>sg.qut.edu.au
-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
the count
> gr...@pacificnet.net (George Ruof) writes:
> >stu...@staff.uiuc.edu (anthony stuckey) wrote:
>
> >> A smaller faster group that doesn't want to get wiped out by a
> >> bigger, slower group can run around the board essentially
> >> infinitely.
>
> > Hmmm... How would it be if after a certain number of rounds a unit
> > became tired and couldn't move as far? Would that allow another
> > unit to eventually catch them?
>
> sides, so the pursuer would also be slowing down.
But presumably the slower guys, being tougher and having a higher
endurance (minotaurs and dwarves, for example, tend to stereotpyically
be thought of as having high endurance), wouldn't start slowing down as
soon as the fast guys...
John M Clancy
Bruno Wolff III <br...@cerberus.csd.uwm.edu> wrote in message
news:7dgv69$one$1...@uwm.edu...
> From article
<C7132130A76C8D51.5042F873...@library-proxy.airnews.ne
> >
> >> A smaller faster group that doesn't want to get wiped out by a
> >>bigger, slower group can run around the board essentially infinitely.
> >
> > Hmmm... How would it be if after a certain number of rounds a unit
> > became tired and couldn't move as far? Would that allow another unit to
> > eventually catch them?
>
> and there is healing going on or if one person is behind intact walls with
> no archers and the catapult has been destroyed.
>
> Assuming you don't want to have a battle end with no result (since there
isn't
> a way to do that), I would suggest just having one side or the other win
> after a fixed number of rounds. You could allow the loser one last chance
> to retreat or surrender if applicable.
>
> As to who wins, you could use one of the following:
> Defender Wins
> Attacker Wins
> Most hit points in creatures wins, with one of attacker or defender
winning
> ties
>
I would say make it:
Defender wins if it's a castle battle.
Attacker wins if it's a field battle.
> You could probably pick a limit of around 100 rounds.
That's a bit much. I think 50 is plenty.
Alexx S Kay
>stu...@staff.uiuc.edu (anthony stuckey) wrote:
>> A smaller faster group that doesn't want to get wiped out by a
>>bigger, slower group can run around the board essentially infinitely.
>Hmmm... How would it be if after a certain number of rounds a unit
>became tired and couldn't move as far? Would that allow another unit to
>eventually catch them?
I don't think so. Such a 'tiredness' penalty should apply to both
sides, so the pursuer would also be slowing down.
Alexx
Al...@world.std.com
http://world.std.com/~alexx
Everybody should regularly make a habit of reading things that
they disagree with vehemently. It's part of keeping your mental
pencils sharp.
Plexy
hired are halved, in Homm3 they just don't increase which is why I am less
likely to load the autosave.
David Lam wrote in message ...
Jason Berkan
>Unlike Homm2 when the week of the Plague occurs, the troops you have not
>hired are halved, in Homm3 they just don't increase which is why I am less
>likely to load the autosave.
Actually the week of the plague works the same way. No new troops and
any unbought troops are halved. (With the exception of troops w/ a
horde building which for some reason produce normally).
Jason Berkan
Regina, SK
Nan Wang
> My thought on battle stalemates is that this issue would be best resolved
> with each round of combat taking a small amount of the attackers movement
> points. When he runs out of movement points the battle breaks off.
> I think this might improve another aspect of play where a leader just barely
> reaches another leader or castle with its movement. The leader would then
> need to win a quick battle or the battle end without conclusion.
> I have suggested this before, but NWC didn't seem to like the idea.
Me neither.
valdu...@gmail.com
Calming Sounds
> HI HOMM3 Fans,
> Firstly I'd like to point out that I love this game, and am completely
> addicted. Congratulations on yet another winner NWC. Many late nights
> ahead....
> However, here are some points I would like to mention.
> Point1.
> I know this has been discussed before, but it really does seem to me that
> the Heroes 3 AI, even on the normal level is cheating. Someone mentioned
> in an earlier post that the AI is now very good at buying all of its
> creatures up to and including seventh level each week, well to me it seems
> a little *too* good to me.
> Unless I have missed something, It is not possible for a fully upgraded
> castle to be self-sustainable, other than maybe a capital. Even if you
> factor in all of the sites on the map that give extra gold each week,
> 2000 per day cannot buy all the creatures for a particular 7th level castle
> particularly in the case of Tower/Castle and all the $5000 troops.
> My example is the scenario Titan's Winter. Quickly in this scenario I found
> myself in the position of owning an entire continent, whilst another CP
> owned another continent. I had 3 fully upgraded Tower Castles on my
> continent and owned 4-5 of each mine. Even with all of the town income,
> map-site income and constant marketplace resource-to-gold income I was
> struggling to buy every single creature at all 3 castles each week.
> However, as I quickly learned when I moved onto his continent and fought
> him, the computer player had no trouble in buying EVERY single creature
> each week, including 12 Archangels (1 flagged POG and 4 castles.)
> From browsing his continent, he had 4 castles (4K+2K+2K+2K = 10K Income)
> and 2 Gold Mines. Thats 12K per day which is 84K per week. 12 Archangels
> alone cost 60K. Factor in the Champions and you're up to76K. Factor in
> the Druids and you're already over 84K. Yet he managed to purchase the
> Crusaders/Griffons/Halberdiers/Marksman (all of them) from all 4 castles
> each week. Not sure exactly how much more cash that is, but its a
> lot extra.
> Ok, am I missing something obvious here when I play this game, or is the
> AI cheating?
> Point 2:
> Anyone else noticed how the formula for quick combat when the
> computer is fighting stuff seems to be:
> If A tougher than B
> B dies
> A remains unchanged
> else
> A dies
> B remains unchanged
> To me, this has become very evident from playing some games, and explains
> why the AI cruises through many wandering stacks and reaches your base
> with a decent army fairly quickly. It also ensures that an AI which is
> doing well does not lose valuable armies eliminating other AI opponents.
> I've seen AI superheroes defeat other AI superheroes then come for me
> literally without a scratch..
> Point 3:
> How on earth does the wandering monsters joining algorithm work?
> Picture this. You've built up a superhero, you've fought for several
> hours to take control of most of the map. My hero:
> 17 Titans
> 50 ArchMages
> 58 Genies
> 43 Naga Queens
> 300 Gremlin Masters
> some large amount of gargoyles
> 80 iron golems
> I have one more enemy 'superhero' to defeat. I am after him and quite
> confident of victory. Our stats are about the same and he has:
> 16 Anicent Behemoths
> 26 Cyclops
> 40 Thunderbirds
> 47 Ogres
> similear large numbers of lower level units.
> Then he walks into a stack of wandering Behemoths (not ancient) that
> SHOULD have kicked his butt. I am very surprised to right click and see
> that a HORDE of Behemoths just joined his tough hero.
> Now I fight this hero, and face 55 Behemoths. Now my hero is tough, but
> not THAT tough, and I lose quite badly.
> How is it fair that a wandering army more powerful than the rest of his
> entire stack (or close to it) should be able to join him. We both spent
> the entire game building up our stacks, going into battle, building up
> more, etc and then in one instant he grabs something which is equal to
> everything I have. Now, I would not have a problem with this if he
> already had 3xHP of the wandering stack, or similar, but clearly in this
> case he didn't. In fact I think I worked out his stack had LESS HP than
> the 55 Behemoths. I lost the game (4 hour game) because of this so
> naturally now I am a little ticked. Maybe I will calm down and change
> my mind on this one ... comments?
> Any agreement/disagreement on the above?
> Deano
>
> --
> ---------------------------------------------------------
> Dean Farmer.
> 5th Year Comms Eng/Comp. Science
> RMIT University. Melbourne Australia.
> --
> ---------------------------------------------------------
> Dean Farmer.
> 5th Year Comms Eng/Comp. Science
> RMIT University. Melbourne Australia.
To your 3rd point. once you have expert Diplomacy and a strong army, most wondering creatures will be willing to join you.