We’ve been blessed with a lot of fantasy 4X games in the past few years.
Besides the Elemental games we’ve gotten Warlock, Age of Wonders 3, Eador, and soon Endless Legends. If you like fantasy games, I highly recommend checking them all out (except War of Magic which is inferior to all of them).
Unfortunately, as game designers, we have an unusual challenge: Magic.
Magic isn’t fair
It is really really hard to write good AI in a game that literally allows players to conjure up all kinds of game changing stuff. The granddaddy of this genre, Master of Magic, didn’t really have an AI. In theory it did but the AI doesn’t really provide a challenge. The game’s amazing game design makes a strong case that it’s the journey that matters, not the destination. That is, in MOM, winning is a foregone conclusion. Of course you’re going to win. The question is how?
Revisiting Elemental: War of Magic
At the risk of being boastful, as problematic as War of Magic was at launch, it was highly innovative. The cloth map mode it introduced is kind of expected now.
Elemental: War of Magic introduced the cloth map zoom out concept, now obvious in hindsight
The city building features were pretty amazing
The problems of Elemental: War of Magic could be summed up as follows: It was a series of interesting game concepts that were not tied together versus computer opponents who weren’t sure what they are supposed to be doing.
If I had to do it over again Elemental: War of Magic would have been bulleted like this:
This design takes into account the basic problem in magic games: You can’t make a Civilization style 4X game and have powerful magic at the same time. If the object of the game is to conquer another empire, then you have to deal with balance and magic eliminates that balance. Conquering other cities should not be the goal in these games. It’s boring and tedious. It should be optional but not central to whether you win or lose.
By making magic both the tool AND the goal, you can eliminate mundane balance issues. Want to protect your capital by surrounding yourself with mountains or ocean? No problem. Go for it. But you can’t do that if the AI is required to actually conquer your cities in order win.
If city conquest is the goal, then magic has to be gimped and at that point why have it?
This was poorly put, as others have pointed out, but I think it really does get to the heart of the problem. Let's reword it to be less harsh:
The AI in 4x games is terrible because developers do not have enough time, appropriate resources, and a cohesive plan to properly program computer opponents.
From what I understand of game development (and I admittedly am looking from the outside in), AI is something that is developed to deal with the systems in a game after those systems are in place, almost as an afterthought, and never with enough time or money. Even the much lauded AI of GalCiv 2 was the result of Brad tweaking stuff in his spare time over the course of many patches, after the game was released. A better approach would be to develop systems and AI simultaneously, so that every step of the way you are building something that the AI can evaluate properly and use to effect. Of course this still needs to be built into a unified whole that becomes a computer opponents, but the result should be well built decision trees that mesh seemlessly with the various systems that make up the game.
not quite sure what this has to do with fantasy. you can have sci-fi terms like, i don't know.. asteroid strike, planet killer, cloning, that essentially does the same thing.
Comparing the games to chess is not entirely a good correlation.
While in a sense it's true that in a game you got more options (build 1 of 10 buildings on 1 of 10 colonize-able locations and move to 1 of 10 locations instead of which of the 8 peons to move, and 1 or 2 moves for the first turn), you can also mop them a lot better- out of the 10 buildings, only 3 actually worth building. from the 10 locations, 2 are different and better than the rest in all respects, and out of the 10 movement options you do not want to stay put, so it's actually 9 options instead of 10. The chess still has 16 options. 3*2*9=54, which is 3 times as many play options, but in order to decide which one to pick you will usually need to look at 3 turns ahead, not 10(as "the further i look the better" goes in chess).
This refinement of looking only in the near future with certain goals, and on the further future with a bit different goals means that overall many of your choices are predetermined as "best course of action", which means your option narrow down very fast.
Regarding the AI- I'd like to know how much time/developers are really put onto the issue. Brad said that the reason GalCiv has great AI (personally I didn't connect with GalCiv) is because he worked on it for years. On many of the games, you build the game, fix it up a bit and ship it. You got literally hundred of hours(if not more) built into things like the Graphic Engine for it to work, with extra time for people to use it. You got more time spent on building the game, combining it with the graphic engine etc. But considering that the game can be "played" as anything beyond meager testing only in its late stages, the AI development starts only near the end.
This also means that it doesn't get the love it deserves before shipping, and doesn't tend to be a priority.
As a "bonus feature", how much is the expected return of the AI? According to brad, investing to raise the AI from being challenging to 90% of the player-base to 95% of the player-base is not worth it.
Getting into hardware issues is not really an issue, I do believe you can free that much with good coding after seeing a video of a couple of guys talking thro a skype-like system at the 80's (which means they did on something like 56K bit connection at most), or the assembly 4K intro competitions. Both examples shows how wasteful our usage of our resources is, and that we can do better, much much better if we ever so wish. (with the obvious problem of investing time in it)
it's not insulting or derogatory. it's just the sad reality of the industry as it exists today.
true observations:
1) many 4X game developers admit that they don't prioritize good AI because people will buy the games anyways. game developers (as in the organizations who develop games) cut corners and are lazy. it's standard operating procedure for businesses.
2) developers don't have the budget to pay a lot of smart people to work on AI. look at the game credits to see an idea of how many people contributed in some way to a game's AI. 4X game development is almost always a low-budget operation, with AI being a tiny fraction of a tiny budget. they haven't even moved AI to the cloud yet, because that's an additional expense.
3) development organizations don't have enough skill to ship a decent AI within their time-frame/budget. part of that lack of skill is programmer talent. part of it is bad project management. part of it is game designers making the strategic decision-making too hard relative to the abilities of the AI.
the exact ratios of each limitation vary from developer to developer, but some combination exists pretty much everywhere, and it's why AI is terrible.
say what you want about AI being 'hard' (whatever that means), but don't pretend to use existing games as 'evidence' of what the limitations actually are.
I can go along with points 1-3, but not the last line. In fact, I would also add Tic-Tac-Toe as evidence; that's as easy an AI as you're going to get. At the end of the day, there is really only one limitation to artificial intelligence: human intelligence. One need only turn on the TV to see that clearly we have a dearth of resources on that front.
I think the only real impetus to change is to bring back a form of genuine character death.
It would focus designs sharply on consequences before they created "fantastic abilities" and it would turn the end game dilemma on its head again.
If a low level and a high level mage could each be killed by a single bullet when careless it would foster cooperative play, genuine alliances and betrayals and really put people in the story.
I don't think gimmicks and layered penalties are the whole solution to being overpowered--magic or otherwise.
AI is just one aspect of making a game permanent fun.
Exactly. When magic is taken to the extreme, like a lot of fantasy settings do, unmagical armies quickly become useless. Soon after that human civilization becomes pointless except as amusement for the wizards and that would tend to get destroyed when they fight.
The only problem with magic being taken to the extreme is that it's done badly.
Lets assume Earth had a very strong magic force around it. Life, when evolved, would utilize it. That means that a pack of Storm Wolves would run lightening-fast to catch the cow by surprise. The cow, if it manages to in time, will try to harden itself to be like a rock in order to survive.
When working the fields, the farmers would use their magic in attempts to repel bugs and ravens from eating the harvest, and so on.
A strong wizard being born in a city will be nothing but the local bully. The same way that certain men are physically stronger than others, if the villagers decide they don't like the wizard, they will kick his butt. They are spellcasters just as well, they know a few spells for fighting (what, do you think bar fights won't happen?) and the same way the strongest man will fall against numbers, so will the said wizard.
The problem comes with the restriction that magic is "inherent" and that only some people can use it. In a world where there is strong magic, if only a small part of the population can utilize it, they will quickly flourish, dominate their surrounding to extinction, and breed their magic using ability, so (assuming they don't drive themselves to extinction) they will, in turn, turn the world into the one I described above.
I honestly think the problem isn't that magic is over powered, it's the concept of mana that is. Sorry going to be rushed an unformatted here. I appologize.
I'm currently working on my own 4 X game. I've been currently looking at this puzzle as I design the combat system.
I want to see what it's like to approach it with mana being a scare but powerful resource. Each node will give 1 - 3 mana points. Those mana points can be used to cast a limited number of magic spells from units with the magic ability. So lets say you have 4 fire mana. Fireball takes 3 mana but it can do a ton of damage. If you expend the 3 mana you are limited to 1 mana for the remainder of the turn. Each turn you regenerate only .5 mana from each node up to your mana capacity (basically 1 bucket per node plus a few bonuses.)
If instead you want to buff your units, which should be more efficient then instant spells, you can imbue them which will halt regeneration of one of your mana buckets.
Brad, I am going to disagree with you. One of my favorite fantasy series is the Dominions series (currently up to number 4). You can only win by destroying your opponents, it has a large empire building aspect, detailed magic (and very powerful magic), and a great back story. It's AI while weak (compared to human players who have been playing the series in MP for the past decade) will give new players a run for their money and they will likely lost the first few games (unlike most of the titles you have listed).
The problem isn't magic. If the AI used magic effectively in one of these games you might have a different opinion that that is where the problem is.
My favorite 4x-ish (it doesn't really have exploration and extermination isn't your goal) game lately is Crusader Kings 2. The interesting thing is that there is no victory condition, only loss (if your entire family line dies out). You do what you feel like and set your own goals. It is all about the journey, not the destination. It is the stories that come out of the game that make the game interesting for me, not casting the uber-spell or taking out the last city.
Another game to look at for inspiration is King of Dragon Pass. While I haven't played this game in years, I still remember events from it.
In any case, you know where to find me if you want to discuss it more.
I've never played Dominions 4. The magic is very powerful? Are you allowed to change the world so that you destroy an entire continent or cities with volcanoes or eternal blizzards? Can you morph your landscape to add extra fertile land and summon entire armies of the undead whenever wish? Is the cost of doing these types of spells so prohibitive that you can only cast something of these spells once every 10 turns?
I tend to think that dominions does not have powerful magic in that sense. It probably has good magic, but I don't think there are any blue-bolts in the game. I find that a lot of the magic in D&D is rather weak compares what should be possible in a 4x game. Of course I've never played dominions and I do not know what you consider powerful magic in certain terms.
Wow, you know nothing about Dominions, yet you say what it does and doesn't have.
How about a mid level global spell called the Burden of Time?
"Causes greatly increased aging for units. Increases unrest across the worldeach turn. Also kills about 2% of the world's population each month.Population in provinces with a Death scale die a little more swiftly."
or maybe the high level Utterdark
"Plunges the world into utter darkness. Income and resources in all provincesexcept caves and deep seas are reduced by 90%. All units except undeadand blind units have their combat values decreased by 6. Units with Darkvisionare less affected according to the extent of their Darkvision attribute."
or how about creating a second sun to increase the heat (and death rate) of the world, or freezing over all water so that you can't sail across, or causing half the people in a country to be diseased and likely die within a few turns, or putting the entire world under a perpetual storm, or how about "Life after Death" where all friendly units get an extra life so when they die they instantly reawaken as a zombie, "Army of Rats" where all enemy soldiers are permanently shrunken to the size of rats and given the affliction Fright, or hundreds of different summonings (the spell list is roughly 100 pages of the manual (which is 400 pages)).
That is much better worded than your first post, still i am not really sure i buy into that.
Because, you know, if i dont judge AI behavior by comparing existing games, what do i judge them by? What do you compare them to? I for one dont know anything about AI coding, so i can compare only to other games.
Also your three points are useable on every gamepart in existence, ever. Graphics, design, sound, whatever.
"the exact ratios of each limitation vary from developer to developer, but some combination exists pretty much everywhere, and it's why AI / graphics / sound / design is terrible."
I did not say what it does or does not have, I only speculated on the possible things.
These are global enchantments? How much does it change, how you play the game? Do you change any of your strategy or tactics? Blind everyone seems like a huge buff to undead. I really don't know how the mechanics in the game work, so I can't say if those spells are powerful in the sense that they drastically change your approach to the game vs. they sound really powerful.
All those serious spells change game drastically. One of those can basically make a nation going from top to bottom within one turn ... But yet, all of them can be predicted and dealt with.
Dominions is a game that for me resemble the reality most of all games. The "balance" in normal life comes from millions of factors (luck, morale, food, history, heritage, tradition, military strength, economy, determinations, weather, politics, geography. science etc.). This makes "overpowered" nations so rare and usually they do not last for long ...
The complexity of dominions is so great that it becomes a truly balanced game, there is just so many ways of playing it, and to everything you attempt someone can find more than one counter. In MP this game is a beast from different league; in my opinion it is ahead of all competition by years.
Unfortunately it is also very time consuming
At first: Magic is not unfair, it is the great difference between the 4X games.
But let us talk about AI: We have a good AI in chess today. But it needs 50 years and tousands of developer and player for it. In my opinion, the best solution is an AI editor for the internet. Please give us the possibility to wrote a better AI. You can't know all strategies and a strategy games, you implement three or four or five, but if we know them, the game is dead.
If you want a game that will be played in five years like today, you need an interaction between players, a multiplayer! But there are so few good multiplayer-4X-games in the wild. As a player with over 10 years mp-experience in AoW I can say: Nobody want to see a 10-minutes-fight between two players and do nothing in the time. So it needs two things: 1. Play the game on the world map (with zoom for example) like in Warlock and 2. make the game more automatic. AoW has the problem with a shorter viewrange and movementrange - you need more microactions to manage your moves. MoM has a lot of passive actions, so the mass of fights are fast and exiting by magic: You can win 1-unit vs n-unit fights, but you can loose a whole army against a settler with very powerful spells.
In MoM we have great magic. Different schools with different focuses: destruction, summonings and fantastic creatures, speels for a better normal-unit-army and so on. Very cool, but the best are the spells to fight against other schools (i.e. to make the enemy spells more expensive or block it) and turn the game completly : I mean the spells like Zombie Mastery, Armageddon, Dead Wish and so on. Nobody will retreat a game after a few hours of gaming, because he had no chance to win anymore. This is the difference to games like Civ, MoO, AoW, Warlock or FE
In my opinion a 4X with a good magic system has great chances to be an AAA title.
This is a rude response. Parrot didn't seem to be making any judgments and was only asking questions.
You interpreted my response as rude, I interpreted Parrot's as rude and replied in kind. IRL, I may not have as I'd have been able to interpret his tone better, that's the problem with written communication. Brad can take my karma point from me if he wants.
I'm coming here late but I have to say, Heart Shaped Man's post just really struck me - that is exactly how I feel. I get bored once I've established my empire and the quests are done. I have been playing these fantasy games for hours every day (AOW3 most recently) and something has been bugging me that I just didn't quite get. I've noticed that I rarely finish a game but I love starting new ones.
I love the idea of just trying to create a kingdom, do quests, reach some goal, without having to conquer all of the enemies cities. Spending the time making the AI keep me off balance and challenged (rather than it trying to manage an empire) I think would make for a very interesting game. I think of WITP-AE as the Americans. I know I will eventually win but how will I deal with all of the scripted AI Japanese invasions? Where will I make my best bases? I also think of famous tactics I used in tower defense games (e.g. Orcs must die). The point being I don't have to conquer the map to have fun (and in fact it gets quite boring if I have to).
It is often inevitable that I will win, but it's the stories of the tribulations I faced getting there that makes it fun.
Mozo.
Written communication does make reading facial expressions difficult. In my field of work, I've been trained to be very blunt and straight forward and sometimes this comes off as rude. i took no offense from the post and I didn't intend my response to be rude. I found that I received more in-depth information from the reply than the original response as it gave examples of the spells considered to be powerful (which they do sound powerful)
Good magic can break a game quickly. it is fun to dynamically change the world to your will. Bending borders and creating things that should not be there in the first place. With proper magic system you will be seeing yourself in a RPG game lime MAGE from white-wolf. You can think it up, you can cast it... the AI would never be able to anticipate this nor be able to create such a game. This is the problem with magic is that there are no rules of why you shouldn't be able to do something... it's magic after all.
@Mozo293: I feel like you. The early game is great: A lot of exploration, weak troops and magic, but very few micromanagement.
But how about a beginning with a lot of units ("surviors") and a start in a destroyed city? With not so fast unit growth then? I think this setting has interesting and new possibilitys for a 4X-game.
I'm totally agree with The_Biz.
The problem is also in the lack of time. In order to start making AI and teach it playing, you need to have all game features implemented first. Modern developers rushes to release games in sales on beta stages (or even earlier). Due to financial reasons nobody wants to wait half year for AI coders to create and balance AI while all other game parts is generally complete. On the other side nobody wants to invest in AI development after the game was actually sold.
When you buynig the game on the release week you can only see box/pictures/trailers and cannot really evaluate AI. Assuming that developers are setting low priority for AI coding.
Those who trying to code AI while features are not done are facing risks at the project's final stages of doing useless work. This is actually happen when completed features are not the same as planned to be on early development.
This is very much true, in my case. I've had to rip out entire working AI systems and redo them because the design changed radically in the last month of development.
And it could be that's exactly what is meant by "magic is not fair." When you're a developer, coding magic is a lot of fun. You dream up all these spells that can do anything. And for the most part, balancing them is perceived to be a simple matter of changing the level or the mana of the spell. But then the AI guy along, and he has to abstract all this stuff out to a directed a cyclic graph. You explode the number of states the AI guy has to deal with exponentially.
I have some particular gripes about 4x Fantasy games:
1. LEVELS AND CLASSES: All these years of game development and we still use these Dungeon and Dragon mechanics? These systems impose linearity and rigidity in the game system. I would prefer to have units and champions gain skills more freely and naturally. Kill a monster? You get points to develop combat skills. Govern a city? Administrative skill points. Give characters freedom to develop naturally. Put more emphasis on non combat skills and roleplaying.
2. LACK OF ENTROPY: Many posters noted how they start a new game when the exploring is over because they have nothing to do. The problem here is that the cities run smoothly all the time; there is no "Wear and tear" so to speak. What 4X games need are mechanics to simulate things like creeping corruption and unrest that force players to deal with internal problems.
One of the most interesting mods developed for Civilization 4 was the Revolution mod. This mod added dynamics for new civs to rise in the middle of the game, overextended empires to crumble and colonies breaking away from their parent civ. This forced players to pay more attention to internal affairs. In FE:LH, the unrest statistic could be used to implement similar dynamics like rebellious cities, city riots and the like.
There are many great features available to you once you register, including:
Sign in or Create Account
