The AI in 1.12 is flat out broke when it comes to waging war. Its fine fighting other AIs as they are broke too so it evens things out, but against the player it is just crippled. On my current map I setup a me vs team of 4 vs team of 4, all AIs set to Expert. I figured id sit in the corner building up my empire while the 4 vs 4 duked it out, and when one of them was winning (or had won) they would then come after me. That part worked out beautifully and eventually after many hours I had the 3 strongest nations that were teamed all declare war on me within like 3 turns. So here I thought I would get this awesome siege on my territory that I could repel (or at least try to). Instead...
After declaring war on me, none of the three AIs, all of whos territory and cities bordered my kingdom, had even a single unit at my border ready to attack. This was like 1939 poland declaring war on germany and then trying to prep some cavalry to go across the border. This is all the while my whole armies are stationed at the borders to defend my frontier. I waited and waited and eventually the first enemy stack arrived, 5 units, three of whom were spearmen and two were archers. Why on earth would the AI send an incomplete stack? All by itself? So it died a horrible death as expected, and I waited and waited and eventually another stack showed up. 3 units, two of which were spearmen and one was an archer. That AI's military strength at the time was rated 2500. So I kept pressing next turn hoping that eventually the blitzkrieg would show up and I giant wave of enemies would descend upon me. Nope. They kept slowly streaming in one incomplete stack after another, single file, from all three AIs. EVENTUALLY the AI rolled out its giant gun: a stack of 1 champion and 5 pioneers. *facepalm* . This was followed up by 2 champions without any escorts who would then after coming back to life after 10 turns without even having full health charge and attack my main stack all by themselves. The strategy of confuse the player, brilliant!
So eventually I got bored and sent my two main stacks in opposite directions to conquer some enemy cities. I wouldnt raze them, I wouldnt build anything there, and I wouldnt station any units there to defend. Just wanted to see what the AI would do to try to take the cities back. To my surprise (and in hindsight I shouldnt have been surprised) the AI did next to nothing. It initially tried to take back one of the cities, which I let it a few times, only to retake it. It soon after that stopped trying to take them back, even though they were undefended. I went deep into the enemy territory and captured a few cities here and there and then retreated to my homeland. Surely the AI would retake undefended cities deep in its lands, yes? Sadly the answer is no, it went about its business completely ignoring the loss of the towns it once had, even though it had them fully surrounded with other towns it controlled.
So to recap:
AI does not send full stacks to attack, or at least very rarely
AI does not escort stacks with other stacks to create a big attack
AI does not escort champions enough and uses them as just a regular unit, probably treating them as just scouts
AI does not defend its border which it should do at all times, even during peace time
AI does not change attack patterns based on previous results and thus never learning from its mistake (aka always attacks same spot, same way)
AI does not setup multi pronged attacks (attack 2+ different places at the same time)
AI does not retake cities with any priority
I played Elemental 1.0 when it first came out for a few hours. Uninstalled it and waited till FE 1.1 to play again hoping to get the complete and fixed Elemental experience. Im still waiting.
No! I am a robot!
Just kidding. The AI could use some improvements. If you're expecting it to play exactly like you would, you might be in for some disappointment. I hope you keep on playing!
Unfortunately, after a few weeks of fairly hardcore playing, I have to agree with the OP. I can't play at this point after realizing exactly how bad the AI is performing.
You can always bump difficulty up to a higher setting. Or, if you already have it on insane, you can change the map to make things interesting (though anything that hurts you will probably also hurt the AI and anything that helps the AI will also help you... still, insane can be quite challenging).
My experience has been quite different. In my current game (on Normal difficulty) the AI declared war and immediately (i.e., next turn) attacked me on two fronts - literally the opposite ends of my realm. One attack was by a 300+ stack with a champion that had marched through my territory for several turns in preparation for a surprise attack on one end of my kingdom. The second attack was by three separate 250+ stacks (one being a 400+ stack led by the enemy sov) that all attacked my capital in waves on the same turn.
I've been very impressed so far.
Well its certainly possible to get lucky and have the AI do something impressive, but I think thats just circumstance rather than proper planning. Ive yet to see the AI do ANYTHING impressive even just once and im playing on difficulty settings where all the best routines are supposed to be used (or so it says).
I think this also depends on the sovereign and the resources available to that sovereign. They do take different approaches from each other...
I do agree, the AI needs work, but I think one problem is how differently people play the game, some tend to refer to the early game to be somewhere between turn 1 and turn 200... turn 200 as early .
Others like me usually finish up games or is in a winning condition around turn 150 or 200 unless the AI does some really impressive expanding and resource grabbing.
That said, the AI is extremely random at best, and it really does need some better algorithm's, especially around warfare and tactical combat. I did play the game extensively in the beta and suffered from this AI, but a lot worse.(But I kept playing as it was the beta, and my plays gave valuable feedback, I think).
The AI also needs to recognise patterns of abuse and try to counteract them, when I say abuse I mean a player spamming cavalry a lot, or spamming dodge, or archers.
Another thing that helps the AI is MORE UNIT DESIGNS, this have been said since early beta and was never really touched...More unit designs doesn't help the AI with its bad decisions though.
Edit: PS. I too have been leaving elemental alone, keeping track of the forums and updates hoping the devs fix all MY issues , I got too good at the game, and now only pointless difficulties which make the game into a whole other kind of game persists.When you play on insane, it is very much a different game, still possible, but not very fun.
Sincerely~ Kongdej
Saying the AI is "broken" implies, that it is incapable of doing anything. This is definitely not the case here. The AI is capable of training troops, grouping them together, sending them out to battle, fighting, coming up with some form of strategy, and more. Does it make mistakes? Yes, sometimes even pretty big ones. Can it be improved? Definitely. But saying it is "broken" is nothing but hyperbole.
Sorry for the rant, but the constant misuse of the work "broken" is a pet-peeve of mine. Still, I'd really like to know which land-based 4X game, with similar complexity as FE, has an AI capable of all the things you ask of this one.
I agree with everything you say Kongdej.
I keep bumping up the difficulty in the hope that something interesting might happen but all that does is that the AI gets better at churning out troops which should overwhelm you with their quality and numbers. However, it then implements attacks so poorly that you have half a chance of keeping going long enough to actually win.
Having ridiculous bonuses does not equate to good AI. Given that chess is a game where the computer AI will beat any human it should be possible to have an AI that beats all but the very best player without any bonuses whatsoever. After all the computer "knows" the research tree, combat system, terrain cost, spell effects etc.. It should be able to assess combats and threats with near perfect results. However even on the highest settings it plays like a idiot but with a ridiculously buffed army.
It is still pleasurable to play but it is not a real battle of wits.
I posted this a couple of month ago. Got flamed.
The problem is that the AI cannot play this game beyond training stuff in the cities. The AI does not "understand" the traits and other stuff, it just has algorithms that tell it what to produce.The most fun in this game usually comes from conquering the surroundings.
I dont know though why the Devs cannot design a less suicidal AI that is capable of producing stacks of varied units and is capable of abusing its own production bonuses in designing top units. Just making a lot of cav with some archers should be a good strategy.
Sometimes I think the AI is too "smart" for its own good - it has a relatively large brain but is incapable of using it because it gets confused within its different priorities.
of course the AI cannot use magic either but that's another topic.
Well, all I can do is hope that in the extremely near, like right around the corner, future, there will be patches that make the AI play better.
You ever notice that in just about ALL games now there are threads about how bad the AI is? Why do we keep buying these products when of course we can see they AI's aren't getting enough love when creating these games. Why are graphics so damn important? Why is the UI so damn important when the AI is crap and the game becomes not worth buying or playing anyways? We should boycott games until they finally make a decent AI. It's as bad as gamers buying games knowing that there are going to be bugs and crashes and unplayable issues when they buy a game anymore, it's been that way since really WIN95 and gamers just keep buying games when they are brand new instead of waiting for the bargain bin price which is sure to come not long after release.
Now I bough War of Magic because of Stardocks track record of making a GOOD AI with Galactic Civilization II, but, War of Magic was just horrible and Fallen Enchantress continues to be a pita with no challenge as well with the AI. Cmon now how hard is it to make an AI that will STOMP us humans? I want to read the screaming and gnawing and gnashing of teeth about how HARD the AI is. Let those that don't know strategy and tactics WHINE but BRING US THE MONSTER of an AI to this game NOW!.....PLEASE?
In terms of raw processing power, modern computers are dumber than insects. We rely on human insight to turn their capabilities into something useful, but ultimately that use is still a reflection of human understanding.
In other words: this is hard.
Have to disagree with you because back in the 80's and early 90's two games come to mind that only had 2mb to 5mb of programming and both could and would STOMP a human player on NORMAL settings. Those were "War of the Lance" by SSI and "Centurion:Defender of Rome" by Kellyn Beck and Bits of Magic . Then not too long ago Slitherine made a game called "Spartan" v1.013 that would STOMP humans on Normal settings and players wrote in even complaining about how HARD it was on EASY lmao. So, don't tell me it's "hard" it's LAZY AZZ PROGRAMMERS is what it is.
you would be very wrong about the insect thing. Most insects have fairly low processing power (like x100000... less than humans), even compared to modern computers
Actually in terms of certain "RAW" processing power simple computers today are faster than humans. There is a big difference in the way computers and humans compute.
Computers are linear and are far better than humans at computing linear things.
Humans do parallel computing and are INFINITLY better than computers when it comes to anything that can be computed parallely.
This is one of the reasons why computers CAN do certain things better than humans yet cant do 'simple' things such as recognizing faces and shapes nearly as fast as humans do (but do keep in mind that an immense part of the human brain is dedicated to recognizing faces and facial mimic)
There is a rather big difference here, and I understand how it can be difficult to produce an AI of THAT quality, Chess is a rather simple game, the difficulty in chess is guessing your opponents, move, but everyone can see everything all the time, and there is only a finite and limited amount of (proper) moves in that game. (I played some chess at a younger age, the need to foresee things intrigued me for a month or two).
In Fallen enchantress you have literally hundreds of choices, probably thousands over a few turns, and besides that, the game escalates into even more turns than chess becoming an ever more complicated puzzle for each extra unit, city hero put on the board, and each new technology learned, and spell made.Do not misunderstand this for flaming, or anything else than an explanation to help understand why it might be hard to get the AI right.
For the next few words, I am assuming a whole lot, deal with it .What I fear is that the people who learned, and put themselves into coding the AI you see in today's games are not similar to gamers, they might enjoy a game but they don't have the same level of intricate clockwork running around off how to play this game, as a gamer myself I try to find each combination of units, and heroes, and traits and whatever, and try it together and see what works in which way, doing this lets me end up with a few solid tactics at the end.Also those who get really good at programming AI don't have the time to get really good at playing games, and this of course affect the quality of the AI, it takes quite some time to become extremely proficient with a single set of gaming systems that is as complicated as what you see in fallen enchantress, and a developer would have to be extremely devoted to the game to get the AI to what nears that perfect state I so crave.
On a secondary note, do remember those people who get really good at writing AI's (for games, and other things) are probably (assuming again) hired by universities to work on the real AI deal.
As a gamer though, I feel obligated to crave and beg for the items I really want, moaning to some extent is the only way to really get your point through to game companies these days, so says the theory anyways: http://www.escapistmagazine.com/videos/view/jimquisition/6187-Why-Boycotts-Fail-Where-Whining-Tantrums-Win
We are not really disagreeing here -- computers are fine on specialized tasks. Then again, most humans cannot navigate to flowers like a bee can (for example).
Anyways, the fewer possibilities the computer needs to deal with, the easier it is to write a program to pick the best possibility. But a game with a lot of possibilities is much harder to program for.
I will agree that insects are incredibly dumb when compared to humans.
But most of these words...
anyways, it's not that computers cannot recognize faces (they can do a decent job of that - this is becoming a common application for law enforcement), it's that computers cannot figure out how to recognize faces. It's only when someone figures out how to make them recognize faces that that becomes a possibility. Also, the result of a program recognizing faces isn't something intrinsically useful to the computer - instead it's something that a human values.
Can someone name a land based game similar to this that has a better AI? I'd really like to hear it.
HOMM VI? Warlock? MOM? Civilization V?
Other than Galactic Civilizations, which doesn't have magic, I can't think of one.
I can't win above Hard on a typical map. If you really want the best challenge, IMO, you really need to play as each player and design lots of units. This game's AI will use the units you make and a lot of the challenge is that the AI only uses existing units.
AI is a pretty generic term for various rules sets. Let's consider one: how the AI deals diplomatically with the human and other AIs. I think it's fair to state that GC2 and Civ IV do a much better job of it. There are certainly more options in the former, all of which are well integrated, and handle situations without throwing you with a response that makes no sense. The latter is much clearer about what it likes and doesn't like, and these for the most part make better sense currently than in FE.
If you mean combat AI, then I've seen Warlock (which I think has some really bad AI in some respects, still) computer-driven opponents respond with strategic spells on players it's at war with, and using a sensible selection. This was even true of MoM, at least with the final patch. There's nothing like pissing off a death-oriented mage who starts cursing the land around your cities, one tile at a time. Similarly, the AI in the old Warlords games (not Warlords Battlecry) was exceptionally good. This applies to the units the AI creates, its plan for expansion, and the way it conducts battles.
I'm not suggesting anything by this. I think tactical combat AI is pretty decent in FE, though strategic handling of units leaves a lot to be desired, and diplomacy is a mess. But I still have great hopes for these being dealt with over time. Brad's a very good AI rules creator, but it's a tedious process that requires a great deal of testing for exceptions, and including rules to cover all the exceptions.
It might be interesting to build a list of "thoughts" we are having as players so that whoever is looking into the AI can check off whether the AI considers this.As a second step building lists of actions which achieve objectives would flesh out the AI's behaviors.
Without having an outline of what the AI looks like I can't begin to see whether it has a poor model of how to win the game, is just not considering things, whether it is poorly prioritizing options or poorly implementing its plan.
I used to use a chess program which more or less told you in plain English what it was considering and what it was going to do about it. It would be great fun to see such a log from the AI of an E:FE game and compare it against a humans thought list.
Humans are very good at getting rid of 99% of options as "just plain daft". (Now that causes problems in its own right but it does allow us to focus). Computers tend to be good at exhaustively doing things but my guess is that what is really lacking in the AI here is what we would call a plan.
But back to my original suggestion I reckon I only have 100 different "thoughts" when playing E:FE on the strategic map. (Obviously some are simple "defend that city with a unit of about this strength", and some are more nebulous "I should use hero X to control this part of the map".) Within these thoughts are action chains that might achieve them. I don't know if the AI thinks like this but it needs to. E.g. Build a unit-> move a unit or assess area->find appropriate "champion" -> evaluate strengths ->(Air magic)->grab air nodes->level up hero->chose air level ups in preference to alternatives->once obtained particular spell switch to eliminating local threats.
If we had lists of thoughts and actions then the developers could perhaps spot obvious issues. Where we might be thinking I need to "make a powerful stack" it may just have "make any stack". Or it might be trying to make a powerful stack just doing a very poor job of it. Where we might have "evade stack A because it outclasses anything we have locally but queue a task to round up sufficient strength to deal with it within X turns" there may no equivalent routines at all.
If there was interest I would play through a couple of games and write down every thought and associated action list I was considering and why I chose one when I was between conflicting options.
Thoughts?
CIV V's AI post G&K expansion is way better than the AI in FE. Not a bash, but it is. This is not just about tactical combat AI, but also the AI's ability to build buildings, improvements, move troops, research, settling and expansion abilities, spend cash, diplomacy, etc etc. The FE AI's use of custom troops, how they configure their armies, when they decide to attack, how they move their settlers when expanding, the lack of equipment on their heroes, and many other things are pretty darn poor right now. Civ 5's AI also is really good at focusing its armies and strategy to achieve a certain victory type. A war like AI will focus heavily on a specific social policy that helps its troop strength and then focus heavily on troops and war. I feel that the AI in FE is just kind of willy nilly and has no real goals when it comes to winning the game or even preventing the human player from winning the game. The AI needs to go all out and seek to prevent me from achieving victory, especially when it comes to the obvious ones like spell of making. When cast, all but the most highly loyal/friendly/ally should be declaring war and sending everything they got to stop you from winning the game. Anyway, I am sure they will keep improving the game, and I have hope, but I feel that we have to talk about this stuff so that it can be addresses by the community and developers and improvements can be made. If we all act like fan boys and say nothing is wrong, and say what other game has good AI, or make other excuses, this stuff will never be fixed.
My thought is that the game offers a variety of mechanisms for defeating and evading opponents, and each of these mechanism is worth exhaustive focus (ignoring all other issues in initial considerations).
Damage can come from attacks, from skills, and from spells. Damage from attacks can be avoided by armor, by evasion, by moving properly, by spells, and by defeating the opponent, and by treaty. Any viable strategy should make at least one defensive mechanism a priority and should also have a valid offensive mechanism. If you are using treaty as a defense, then either it's temporary and you are trying to achieve a late-game advantage (which means at some point you need to switch from inoffensive to acting rapidly) or you are trying for a diplomatic victory (in which case diplomacy also becomes your offensive priorityf ). Damage from skills and spells can sometimes be avoided by going with auto-resolution but that should be considered a bug, I think.
Armor can be bypassed using armor piercing weapons and/or poison and/or spells. Spells are limited by access to mana and shards, so they are fueled by essences and by altars/shrines/temples, so those targets should get some priority when you are fighting a spell using opponent. Some spells also have specialized defenses while other spells can only be avoided by ending the battle and/or strategically and/or by having so much health you can ignore them. High evasion needs either spells or high accuracy as a counter. These are examples of strategic priorities and need some kind of decent estimating mechanism on the part of the AI so it can assign the priorities proper weights and judge whether success or failure is likely.
So anyways, there's a complicated tree of options here and winning involves building up multipliers (including time in battle, if your defenses are sufficient) which exceed your opponent's multipliers. So a part of the AI design should include a representation of these multipliers and how they can combine with each other.
Priorities are tricky beasts -- they need to have associated with them some kind of expectation about progress so that you can recognize failure so that you can try shifting priorities when they are not valid. A priority shift might be a temporary thing to deal with a small set back or might be a permanent thing. Also, the game needs to implement multiple priorities, because it's not always the case that a priority is actionable. When you cannot make progress on your top priority you need to take action on a secondary priority. And this raises another issue: if priorities conflict (like maybe we have unit priorities and one priority involves going east and another involves going west) then a decision needs to be made -- sometimes it's better to just wait and not take action on the secondary priority and other times it's better to make the secondary priority the top priority and give up on the top priority (moving it to the lowest priority until "things change" -- I do not know how to properly describe "things change" but there needs to be some memory of why the decision was made...).
Of course another issue is that sometimes the AI is going to lose. Actually, when there are multiple AIs involved, some of the AI instances will always lose (except maybe with diplomacy). Still, ... it does not feel right that the AI is overwhelmingly strong at a certain stage of the game but if you can arrange to defeat it before that point, or survive past that point it presents little challenge.
Anyways, I think "priorities" are some kind of order, hierarchical collection of a data structure.
Well lets look at Warlock as an example. Again were discussing the AIs conduct during wartime only. Does it defend its borders? Yes it does. Stations plenty of units as well as building lots of towers on its frontier. Does it send units against you as soon as war is declared? Yes it does. There is no delay between hostilities being declared and the first attacks occurring. Does it send units in groups/waves to try to overwhelm you? Yes it does, Does it attack from multiple vectors? Yes it does. Now obviously its not perfect too and has a slew of faults, but when war is declared it will test your defenses almost immediately and across multiple places.
(snip)
I just can't understand how you can say that. G&K at Prince (its highest level without cheating) can't conquer a city. I don't think I'm even that good at the game and I've never lost a city to an AI unit. FE on challenging is challenging. I usually win but I feel pressured.
The only area Civ V's AI is better is in the diplomacy area which is something I think FE needs some work on. G&K may have improved the AI but it still does not know how to win.
Prince is not comparable to challenging. Challenging is like one or two steps above Prince. Prince is "normal". Also, the CIV5 AI on prince can take cities. In FE, the "normal" AI is a non-factor. The neutral monsters pose more of a threat than any of the AI, at least in the first bit of the game, and then it is just a grind to win the game without any challenges what-so-ever. I would go as far to say the same about challenging as well.
There are many great features available to you once you register, including:
Sign in or Create Account
