1 vs. 1: Frog vs. AI

By on September 12, 2012 6:11:37 PM from JoeUser Forums JoeUser Forums

Frogboy

Join Date 03/2001
+1484

FrogBoy

AI

Faction: Altar
Strategy: Strangulation
Capital City: Athica (4/3/2)
Intelligence: Not much
Faction: Kraxis
Strategy: Aggressive
Capital: Goyarh (4/2/2)
Intelligence: Challenging

Turn 10

Turn 10

Capital City has Enchanted Hammers.

Capital City has Meditation

Sovereign is level 3

First Pioneer is done

Capital City has Inspiration

No second enchantment

Sovereign is level 1

Pioneer under construction

Turn 50

Turn 50

  • Cities: 3
  • Outposts: 2
  • Sovereign: Level 6 (with Courage, Stoneskin, Burning Blade, Cloak of Thorns)
  • # of Champions: 1
  • None of my new cities have essence
  • Cities: 2
  • Outposts 2
  • Sovereign: Level 2 (with Burning Blade)
  • # of Champions: 2
  • Capital city now has inspiration and meditation (Karavox doesn’t have Enchanted Hammers)
  • Second city has 1 essence (inspiration cast)

Turn 75

Turn 75

Has decided it’s time for the AI to die.

Doesn’t know they’re going to die.

Turn 100

Turn 100

  • Conquered their colony city
  • Deb (Sovereign) is level 10.
  • Use superior mobility to bypass screening armies to attack capital. Capital taken.
  • By turn 110, they are dead.

  • Karavbox is only level 5
  • Capital is defended well. But they send out armies to try to thwart my build out.

 

(To be continued)

Screenshots:

image

I want that scroll!

image

AI bug found, they don’t put any concern on player build up unless THEY dislike the player.

image

Roughly even holdings in the world.

image

First enemy city taken.

image

They see my build up and try to send forces to take me out. I bypass those units and take their capital.

image

They want their capital back

image

Their final, desperate battle. It is the end for them.  They might have won but the AI failed to treat my ranged *spells* intelligently and as a result, they crawled towards me to try to get the first Melee hit. I took advantage of that to use flame dart on them.

image

…and so to work I go…

Locked Post 87 Replies
Search this post
Subscription Options


Reason for Karma (Optional)
Successfully updated karma reason!
September 14, 2012 2:32:21 PM from GalCiv II Forums GalCiv II Forums

To be honest, there should be some restriction to what you can rush build, and a soft limit in rushing in general.  Let's say you rush three things in one turn, the second and third should cost more because you're basically paying for double and triple overtime.  You can do it, but it will be an inefficient use of your treasury.

Reason for Karma (Optional)
Successfully updated karma reason!
September 14, 2012 2:41:54 PM from Elemental Forums Elemental Forums

You should not be able to rush more than one thing per turn, ever. 

But I see no reason not to be able to build multiple things per turn if you have the production capacity.

Reason for Karma (Optional)
Successfully updated karma reason!
September 14, 2012 6:39:18 PM from Elemental Forums Elemental Forums

I do agree the rushing mechanic should be changed. A small change making this a lot smaller issue is to allow last turn's unused production to go to the next item in the list.

To me it seems the AI is currently left behind in the early game. If it manages to get to middle & late stages of the game, then it will play OK. But, usually at that stage it is already left behind and there is no real fight.

My "rules for winning":

  • Never build pioneers if there is no known use for it.
  • Build military units only when you have economic backbone to do so without raising taxes. In the early game, do not build them at all.
  • Try to minimize unrest, maximize production, then population, and finally economy and research in cities (in this order).
  • Rush-build.
  • Leveling up your supporting units is _important_. The HP bonuses are important for your success. Autoresolve will kill your fast units often by rushing them to enemy lines. Put high-def high HP units in the first line, then attack flanks with the glass cannons once engaged.
Reason for Karma (Optional)
Successfully updated karma reason!
September 15, 2012 2:30:10 AM from Elemental Forums Elemental Forums

The current rushing mechanic exploit (I'm going to call it an exploit) is unintuitive and well, stupid.

 

One idea: could min turn build time be a function of city level (6-city level?), and could production be capped at x% of the production needed to build an item- Those things might help solve a couple of problems at once. (pop not meaning enough, and rush exploit.

 

Maybe have extra min build for some things, and make them unrushable.

 

Reason for Karma (Optional)
Successfully updated karma reason!
September 15, 2012 2:56:14 AM from Elemental Forums Elemental Forums

I really fail to understand why people love minimum build times.  A city already has a measure of its production capacity.  It is called, wait for it... production! 

If a city can produce something in three months why the Hell should it take nine instead?  Why do we need to artificially cap and slow down things?  You build an outpost that can control four resource, and now your city has to stop and spend three years on developing them?  No time to dig a well meanwhile, they are too busy?

If a city has X production, it should be able to produce X worth of buildings and units in a turn.  You need a reason not to do it this way, and I have not heard one yet.

Reason for Karma (Optional)
Successfully updated karma reason!
September 15, 2012 3:17:41 AM from Elemental Forums Elemental Forums

>Intelligence: Challenging

where you can see lists of AI? that would look from what level of the AI cheats. and find out how much it cheaters

Reason for Karma (Optional)
Successfully updated karma reason!
September 15, 2012 5:58:34 AM from Stardock Forums Stardock Forums

How about the rush feature only adding a percentage to the production, reducing the time instead of insta-builds.

Reason for Karma (Optional)
Successfully updated karma reason!
September 15, 2012 11:55:49 AM from Elemental Forums Elemental Forums

Quoting CyanPile,
>Intelligence: Challenging

where you can see lists of AI? that would look from what level of the AI cheats. and find out how much it cheaters

test the game. Expert  is 3-4 times stronger (points game). that is, we get the same Challenging, but he comes to your house with a good army.

Reason for Karma (Optional)
Successfully updated karma reason!
September 15, 2012 12:00:57 PM from Elemental Forums Elemental Forums


CoreDifficultyLevels.xml holds the modifiers for difficulty.

Reason for Karma (Optional)
Successfully updated karma reason!
September 15, 2012 12:15:24 PM from Elemental Forums Elemental Forums

Quoting CyanPile,

Quoting CyanPile, reply 57>Intelligence: Challenging

where you can see lists of AI? that would look from what level of the AI cheats. and find out how much it cheaters

test the game. Expert  is 3-4 times stronger (points game). that is, we get the same Challenging, but he comes to your house with a good army.

On Expert the AI has a 25% bonus to income and 25% bonus to the sovereign HP. Also starts with 10 times as much money as you and may have other cheats, as it uses <AIIntelligenceFactor>2.0</AIIntelligenceFactor>, and according go Brad, any value bigger than 1.5 causes the AI to cheat.

Reason for Karma (Optional)
Successfully updated karma reason!
September 15, 2012 12:16:43 PM from Elemental Forums Elemental Forums

Quoting seanw3,
Expert

Thank you.

Quoting bbr91,
On Expert the AI has a 25% bonus to income and 25% bonus to the sovereign HP. Also starts with 10 times as much money as you and may have other cheats, as it uses 2.0, and according go Brad, any value bigger than 1.5 causes the AI to cheat.

 

AIIntelligenceFactor - I thought that this means that he is calculating the moves longer than usual

Thank you.

Reason for Karma (Optional)
Successfully updated karma reason!
September 15, 2012 12:38:40 PM from Elemental Forums Elemental Forums

Quoting CyanPile,
AIIntelligenceFactor - I thought that this means that he is calculating the moves longer than usual

Here's a quote from Brad:

So I’m pumping the computer AI to “challenging”. At this level, he’s identical to the human player still except there’s a few more APIs that get used (a little more CPU gets used).  Those APIs largely revolve around recruiting champions more aggressively and more intelligently handling quests. In the XML, these algorithms are based on  <AIIntelligenceFactor>1.5</AIIntelligenceFactor> which is my own voodoo I use to determine what functions I’ll use.  Go over 1.5 and it might include cheats so for future modders, never go over 1.5 if you don’t want the AI to potentially be “Cheating” (where cheating would include an imbalanced map, getting champions placed in slightly closer to AI starting locations, etc.).

But it from December 29, 2011, so things might have changed in the meanwhile.

Here's the link: http://forums.stardock.com/414671/page/3

Maybe Brad will be kind enough to confirm that this is still true, or to further clarify what exactly AIIntelligenceFactor does.

Reason for Karma (Optional)
Successfully updated karma reason!
September 15, 2012 8:40:17 PM from Elemental Forums Elemental Forums

In my "expert" in cheating automatic battle. And after the defeat of his champions without physical injury.

In general, the computer is not very good at using magic bus. For example I put Haste on the main unit. At the beginning of the fight is far outweighs the balance. He can not do this, when the manna little. When a lot of mana, can take a very expensive spells (100 worth at a time), on which aid is less than Haste.

Вообще компьютер не очень хорошо использует магию в автобое. На пример я наношу Haste на самых главных юнитов. В начале боя это сильно перевешивает баланс. Он же может не сделать этого, когда манны мало. Когда манны много, может применять очень дорогие заклинания (по 100 стоимостью за раз), от которых помощи меньше чем от Haste.

It is desirable that the main characters, he puts a spell. I mean global spell.

Желательно что бы на основных персонажей он наносил защитные заклинания. Я имею ввиду глобальные заклинания.

Reason for Karma (Optional)
Successfully updated karma reason!
September 15, 2012 8:47:45 PM from Elemental Forums Elemental Forums

Yes indeed the tactical combat AI needs further refining, especially in the use of spells. It does tend to be quite wasteful of mana, with little gain. But this is the exact type of issue that will be addressed during the beta 5 period.

Reason for Karma (Optional)
Successfully updated karma reason!
September 16, 2012 5:38:18 AM from Elemental Forums Elemental Forums

A few more important things. AI does not retreat wounded troops. This is especially true for melee fighters. There are a couple of fighters or teams with full health. Is there an advantage in battle. Why did he send in the midst of the battle of wounded soldiers? It is necessary to send in the midst of units with the highest level of health! Wounded champion lose very sorry (for example, the hero up to 50 HP. If he left 10 HP, then he should withdraw. If archers shoot at him, he should retreat and at 20 HP. And he had to hide behind trees). The same goes for other teams. Injured a stone golem. Wounded troops melee. It is necessary to assess the power of the enemy. Archers good estimate of the enemy. Melee is not. If the fight can be won, it must at all costs keep alive Champions. If possible, it is necessary to save recruiting beings with good (like a stone statue of the Golem. Very good unit to start the game). And the first step is to kill the enemy archers. And those who do more damage (most characters). In second place finish off the weakest.

AI know who the man? I read that there is discrimination against the players. For example neutrals do not attack AI. I myself had ever seen. But now I have persuaded the Allies (Allies team. They had to give a lot of their resources. Money, metal, horse, etc.) to attack the common enemy. In this case, the enemy on which all the power is on me. For which he does not attack my allies AI? By the fact that he started the war with me first? Or he knows who before him (AI or human)?

 

Ещё несколько важных вещей. ИИ не отступает ранеными отрядами. Это особенно касается рукопашных бойцов. Есть пара бойцов или отрядов с полным здоровьем. Есть преимущество в битве. Зачем он посылает в гущу сражения раненых бойцов? Надо посылать в гущу юнитов с максимальным уровнем здоровья! Раненого чемпиона потерять очень жалко (допустим у героя максимум 50 HP. если у него осталось 10 HP, то ему надо отступать. если по нему стреляют лучники, то ему надо отступать и при 20 HP. и ему надо прятаться за деревьями). То же самое касается других отрядов. Раненый каменный голем. Раненные отряды ближнего боя. Надо оценивать силы врага. Лучники хорошо оценивают силы врага. Рукопашные же не очень. Если бой можно выиграть, то надо любой ценой оставить чемпионов живыми. По возможности, надо спасти призывных существ с хорошими показателями (вроде каменного голема из статуэтки. очень хороший юнит для начала игры). И первым делом надо убивать лучников у врага. И тех, кто наносит больше урона (чаще всего героев). На втором месте добивание самых слабых.

ИИ знает кто человек? Я читал что есть дискриминация игроков. К примеру нейтралы не нападают на ИИ. Сам я такого не видел. Но вот сейчас я уговорил союзников (союзники по команде. пришлось им отдать кучу своих ресурсов. деньги, металл, лошади и прочее) напасть на общего врага. При этом враг по чему то все силы ведёт на меня. По чему он не нападает на моих союзников ИИ ? По тому что он начал войну со мной первым? Или он знает кто перед ним (ИИ или человек) ?

 

Reason for Karma (Optional)
Successfully updated karma reason!
September 16, 2012 5:57:05 AM from Elemental Forums Elemental Forums

CyanPile, the formula to choose who to attack first in a tactical battle I believe is the same that was used in GalCiv to prioritize targets and is something like Total Attack / (Total Defense + Hit Points). The higher the result, the higher the target priority. Maybe someone can confirm this, I've only seen this in one post before. This formula works well actually, since your weak units will be targeted first, so you will lose your archers quickly.

Perhaps the AI will be further refined to know when to retreat if wounded and so on, we have to wait for the beta 5 for that.

Reason for Karma (Optional)
Successfully updated karma reason!
September 16, 2012 6:33:51 AM from Elemental Forums Elemental Forums

>Total Attack / (Total Defense + Hit Points)

And to whom was he attack first? I just do not quite understand the formula. I attack those who are most prejudicial. In the long battle - is the key to success.

А на кого именно он нападает первым? Не совсем понял формулу. Я нападаю на тех, кто больше всего урона наносит. В длительной битве - это залог успеха.

AI is generally not too bad. I for example a couple of times did: first spell correctly applied the amplifier (after which the MP I had no), put closer to the enemy cannon fodder, and included automatic battle. If the AI at the end of battle has severely wounded away, it would be all perfect (on a fully manual battles long gone). And I'm talking about the battles that in "fast automatic battle" always leads to defeat (ie, all my troops defeated. The enemy occasionally one of the units killed). And with full manual control could they win without losses (but most losses were unimportant 1.2 units). Needed here more elaborate system of priorities. With the division of the enemy's army in two. The use of topographic features. Very often, the victory is that the enemy army droblyu apart. And each of these parts are quickly killed. That is cannon fodder running from the enemy. The strongest teams (champions on horseback. Haste.) To destroy this moment the most important part of the enemy. This is especially effective if the champions of high level (with more than 100 HP. Evage) and in the unit more than two. After that, the city is restored. And again attack the approaching enemy.


ИИ в целом не так уж плох. Я к примеру пару раз делал так: вначале правильно применял усилительные заклинания (после чего маны у меня не оставалось), ставил ближе к врагу пушечное мясо, и включал автоматический бой. Если бы ИИ ещё в конце таких битв отводил подальше тяжело раненых, то было бы вообще идеально (на полностью ручные битвы много времени уходит). И это я говорю о битвах, которые в "быстром автоматическом бою" всегда приводят к поражению (то есть все мои войска разбиты. у врага изредка 1 из отрядов убит). А при полностью ручном управлении удавалось их выиграть даже без потерь (но чаще были потери 1-2 маловажных отрядов).Тут нужна более продуманная система приоритетов. С разделением армии врага на две части. Использование особенностей рельефа . Очень часто победа заключается в том, что вражескую армию дроблю на части. И каждую из этих частей быстро убиваю. То есть пушечное мясо бегает от врага. Самые сильные отряды (чемпионы на конях. + Haste.) в этот момент уничтожают самые главные вражеские части. Особенно это эффективно, если чемпионы высокого уровня (с более чем 100 HP. + Evage ) и их в отряде больше двух. После этого в городе восстанавливаюсь. И снова нападать на приближающегося врага.

 

p.s. Archers can sometimes be overlooked. If your most important units in good condition.

p.s. Лучников иногда можно игнорировать. Если самые важные ваши отряды в хорошем состоянии.

 

 

Reason for Karma (Optional)
Successfully updated karma reason!
September 16, 2012 11:18:17 AM from Elemental Forums Elemental Forums

Please add me to the list of people calling for (somewhat) standardised starting locations. I

generally only ctrl+n until I get an interesting, competetive starting location (so, a shard, decent tiles, and another resource or two), and I'm sicking of spending ages doing so and investing many hours into the resulting game, only to find out when I finally defeat them that my opponents were crippled from birth. Starting location matters so much, and to randomize it entirely for AI players while most human players are effectively guaranteed a good start just throws a huge unnecesarry spanner into the works from the get go. I find the game better when you have a competetive, interesting starting location, and i don't see why the AI shouldn't be afforded the same luxury.

And what about the new players, who don't know about ctrl+n, or what even constitutes a good start location? Are we to handicap them as well? Frogboy always complains about the difficulty of designing games for users with completely different PCs: well isn't it going to be equally tough to design and balance the mechanics for players with completely different starting locations?

Reason for Karma (Optional)
Successfully updated karma reason!
September 16, 2012 11:56:50 AM from Elemental Forums Elemental Forums

I understand the desire for competitive starting locations, but I prefer the randomness of the starts. I am a purest when it comes to gameplay, I take the roles as they land and deal with the limitations. Of course it yields a more difficult game, but I play to play and not to win. Some of my most enjoyable games are when I lose and the game turns to how long can I survive, even after 6 hours of gameplay to a loss, I had fun playing. I don't believe in imparting this gameplay to others and I so I like the compromise already in place crtl+n.

As for the AI starting position, I hope that harder difficulties (higher than challenging I suppose), the AI can cheat the starting location, to always have tile yield of at least 4 3 2 for the starting city (as difficulty goes up the yields move to a 5 4 3 tile yield for starting city), and so the AI just needs to find a river or forest nearby to take advantage (Not sure how difficult that is to program, given the current state).

The idea of standardizing the starting locations, leaves me with fewer options of gameplay. I'd prefer if they didn't limit my gameplay. I like starting the game with no shards nearby or other resources. It leaves me to explore to find the resource and then compete withe AI to get that resource quickly. Sometimes my capital city is not my biggest city or the most productive, feels more realistic. I enjoy playing the game where I have no essence for my first city.

This is not like a poker game, where I only play if I think I can win. But I do not begrudge those who only want to play if they think they can win with their current hand. But I don't want this to be a game where everybody starts with basically the same resource cards in their hand and only the draw deck is randomized. Even if the resource hand is randomized, it is still of the same format. Must have at least 1 shard card, 1 resource card, and tile cards of at least 3 4 2 (or variant). It removes some of the replay value, especially if you grow to acquire new tastes in gameplay mechanics.

Reason for Karma (Optional)
Successfully updated karma reason!
September 16, 2012 12:20:05 PM from GalCiv II Forums GalCiv II Forums

I used to be like, wow, let's found an entire civilisation from one settler unit, this is awesome.  Then some genius added Advanced Starts to Civ 4 and I was like, this is awesome.

I mean, given the choice between running for a hundred turns doing the stuff that you did the past dozen games, or setting up your first three or four cities now complete with resource improvements and buildings, what would you choose?

Reason for Karma (Optional)
Successfully updated karma reason!
September 16, 2012 1:12:19 PM from Elemental Forums Elemental Forums

Quoting MarvinKosh,
I used to be like, wow, let's found an entire civilisation from one settler unit, this is awesome.  Then some genius added Advanced Starts to Civ 4 and I was like, this is awesome.

I mean, given the choice between running for a hundred turns doing the stuff that you did the past dozen games, or setting up your first three or four cities now complete with resource improvements and buildings, what would you choose?

 

I choose option A.  Fallen Enchantress is shaping up to have quite a bit of replay value due to the amount of variety found in the beginning of each new game, and the degree to which this effects it's outcome.   The transition from feeling like a RPG to playing like a strategy game is very smooth.

Reason for Karma (Optional)
Successfully updated karma reason!
September 16, 2012 2:56:55 PM from Elemental Forums Elemental Forums

Quoting parrottmath,
I understand the desire for competitive starting locations, but I prefer the randomness of the starts.   ... 
                                                                                         o   o   o 

The idea of standardizing the starting locations, leaves me with fewer options of gameplay. I'd prefer if they didn't limit my gameplay. I like starting the game with no shards nearby or other resources. It leaves me to explore to find the resource and then compete with the AI to get that resource quickly. Sometimes my capital city is not my biggest city or the most productive, feels more realistic. I enjoy playing the game where I have no essence for my first city.

I agree completely with the above.  But, in particular, I believe in preserving Choice ...

On this issue of "Balanced" starting positions, versus (completely) "Random" starting positions, I would like to see an OPTION, whereby both styles of play (both sets of preferences) can be accomodated.  In this regard, I have previously advocated the availability (in the player-choices/option screens at the start of each game) of a:

Selectable Switch to chose between "Balanced" starting positions, and Random starting positions .

I can't see why this would be difficult to program; and it would maximize player choice, and Replay-ability.  Balanced Starting Locations (for both the human players and the AI) could be 3,3,2's (or better); with some additional resources nearby such as (two of the following) iron ore, shards, grain (or apiaries) and a recruitable champion; no overwhelmingly powerful monsters within  X number of tiles; and non-cul-de-sac starting locations, not hemmed in by mountains or seas.  Obviously, the formula/template for Balanced Starting Locations could be further defined/refined, however Stardock chooses.  By contrast, Random Starting Locations would be exactly that; and they can be chosen by the human player for variety (and subject to how many times you want to use the Cntl-N feature). 

Having variety, and Meaningful Choices, are hall-marks of great 4X games, and great Strategy games.  Adding the Selectable Switch would solve this issue for both camps, and increase E:FE Replay-ability.

[ Final Note:  I apologize for having addressed this issue in at least 3 discussion threads, in the last couple months; but I am constantly amazed by seeing new replies, every few days, in which someone  INSISTS  that either  Balanced Starting Positions, or Random Starting Positions are best; and no one else, who favors the opposite, should have the ability to choose between the two.  People should have choice; and not have their favored style of play forbidden by others ... ]  

Reason for Karma (Optional)
Successfully updated karma reason!
September 16, 2012 3:35:41 PM from Elemental Forums Elemental Forums

I'm with OrionM42 on this one.  Give me a checkbox.  Random (as it is now) and Balanced (whatever beta testers think is best)

My choice for Balanced would 3/3/2, a shard matching the highest path of the sovereign, a hero, a weak quest, and a few weak lairs.

Reason for Karma (Optional)
Successfully updated karma reason!
September 16, 2012 5:20:41 PM from Elemental Forums Elemental Forums

That is as I said before, give options to make all happy.

I prefer randomness, I feel that is the whole point of a random map, the true randomness.

Then others can have the predefined maps based on size (number of players) with predefined and balanced starting locations just like the maps from many a great rts games of the past.

Let this great game be everything to everyone so they can play the way that makes them happy.

Everyone's idea of fun is different and limiting that fun in what seems like a minor way to some will be significant to others. Many a potentially great game has been ruined because  those who value such things had them taken away ( I have always been bitter towards games that had great beginnings then abandoned those great core principles in sequels to fit the mainstream).

I was so thrilled when I heard FE would have TRUE RANDOM MAPS. So please do not take away the true randomness.

if a change is made, please make it an option of choice so all can be happy!

 

Thank you for this great game!!!

Reason for Karma (Optional)
Successfully updated karma reason!
September 16, 2012 8:39:46 PM from Elemental Forums Elemental Forums

Just a point regarding these so called "true random maps"...

As I said earlier, most of the land is impassable and empty. A 100X100 tile maps hold actually only 4000 that you can/will walk on, and they are one giant maze usually. This is a type of map, but it's really different from the normal maps. Atleast give us rough options such as found in civ.

Reason for Karma (Optional)
Successfully updated karma reason!
Stardock Forums v1.0.0.0    #101114  walnut1   Server Load Time: 00:00:00.0000391   Page Render Time: