Beta

Slashdot: News for Nerds

×

Welcome to the Slashdot Beta site -- learn more here. Use the link in the footer or click here to return to the Classic version of Slashdot.

Thank you!

Before you choose to head back to the Classic look of the site, we'd appreciate it if you share your thoughts on the Beta; your feedback is what drives our ongoing development.

Beta is different and we value you taking the time to try it out. Please take a look at the changes we've made in Beta and  learn more about it. Thanks for reading, and for making the site better!

The State of Game AI

Soulskill posted more than 5 years ago | from the don't-let-it-get-the-railgun dept.

Games 88

Gamasutra has a summary written by Dan Kline of Crystal Dynamics for this year's Artificial Intelligence and Interactive Digital Entertainment (AIIDE) Conference held at Stanford University. They discussed why AI capabilities have not scaled with CPU speed, balancing MMO economies and game mechanics, procedural dialogue, and many other topics. Kline also wrote in more detail about the conference at his blog. "... Rabin put forth his own challenge for the future: Despite all this, why is AI still allowed to suck? Because, in his view, sharp AI is just not required for many games, and game designers frequently don't get what AI can do. That was his challenge for this AIIDE — to show others the potential, and necessity, of game AI, to find the problems that designers are trying to tackle, and solve them."

cancel ×

88 comments

GNAA FP (-1, Troll)

Anonymous Coward | more than 5 years ago | (#25622265)

All hail the Gay Nigger Association of America!

Re:GNAA FP (0, Offtopic)

shentino (1139071) | more than 5 years ago | (#25622763)

omg...

metamod funny.

Re:GNAA FP (1)

Tablizer (95088) | more than 5 years ago | (#25623081)

All hail the Gay Nigger Association of America!

Damn, this AI trollbot is growing pretty human-like. It could pass as a genuine web nut Scary tech.
         

Easier to just cheat! (2, Interesting)

Anonymous Coward | more than 5 years ago | (#25622397)

After all, publishers these days only care about churning out sequels quickly, so the so-called 'advanced' AI is basically just a computer versions of cheating player, instead of spending time on increasing the 'I' of the AI.

Most humans aren't that smart (5, Insightful)

TheLink (130905) | more than 5 years ago | (#25623377)

But does increasing the I of the AI actually make games fun?

The Problem that AI is supposed to solve in most Games is not "how to beat the human".
The Problem is "how to make it fun for the human".

Creating an AI that can consistently beat humans is not hard. Making it fun for most humans might not be so easy.

Fact is humans aren't that good at most games (amongst other things). You don't have to be very intelligent to be good at most games. How many of you can beat a computer at chess at high difficulty? How many people actually _lose_ in tic-tac-toe - I've seen more than a few :).

It's often not hard to make a computer extremely good at a game, at least good enough to beat most people. But does that make it fun?

In most FPS games, stupid humans want to be able to mow down _thousands_ of stupider computer controlled enemies - "against the odds". That's what makes it fun for them.

That's just not possible if the enemies start having a lot more brains. Then most players might have difficulty getting past the first 3 enemies :).

It's not that difficult to make an enemy FPS "bot" have superb tactics, coordination, timing etc. Especially if the map is pre-known (which is usually the case). You can code the tactics and heuristics in. If you hear the player in position X, group A enemies head to position Y and group B head to position Z, and bye bye player.

Imagine if enemies that are low in health kept running away and hiding, and then snipe at you from far away when they see that you are busy doing something else. While that might be more realistic, it might not be so fun eh? Who really wants realism in games?

At that rate the player can never pretend to be the hero he wants to be. He'll just be dead. And your game won't sell.

Same goes for RTS games, believe me, you don't have to make a computer cheat to beat humans - a computer can micromanage better than most humans.

Just ensure that basic stuff like navigation is better. Stuff doesn't have to be that smart, but at least they shouldn't be totally stupid - they should be able to walk around stuff without getting stuck - even a "dumb" animal can navigate open spaces better than many computer controlled stuff in games.

Re:Most humans aren't that smart (1)

Kenoli (934612) | more than 5 years ago | (#25623689)

A 'smart' enemy (or NPC or whatever) is not necessarily a 'hard' one. AI could be used much more than it is current to make characters look and act more natural and believable.
You've probably played first person shooters with enemies that act in very simple and predictable ways, and with non-enemy characters being fully scripted and just standing around with a blank expression once they've said their lines.

Re:Most humans aren't that smart (2, Interesting)

TheLink (130905) | more than 5 years ago | (#25623783)

Well I guess you could script some of the AI enemy soldiers to try to save their comrades - who have had their legs blown off by you or something.

The mortally injured ones might alternate from cursing you and begging you to put them out of their misery.

Some might try to surrender to you if they cornered, out of ammo and clearly out-classed.

Might make for a different sort of game. Smart but not hard opponents :).

At the end of the game when you have killed thousands of these, maybe you'd need a shrink ;).

I think the current level of game AI is tolerable.

I'd rather a better Aliens vs Predator game with Crysis sort of tech. With "predator" aliens that can climb trees or walls (unlike those in the previous AVP games - which couldn't - not even slowly).

Re:Most humans aren't that smart (5, Funny)

VShael (62735) | more than 5 years ago | (#25623825)

Imagine if enemies that are low in health kept running away and hiding, and then snipe at you from far away when they see that you are busy doing something else. While that might be more realistic, it might not be so fun eh? Who really wants realism in games?

Huh, you must be right. I guess that's why online gaming and LAN parties are so incredibly unpopular.

Personally, I want to see the day where two AI's argue on the battlefield...
"You camping b*stard!!"
"It's a legitimate strategy!"

Re:Most humans aren't that smart (1)

TheLink (130905) | more than 5 years ago | (#25623881)

OK I'm wrong.

But don't they already have pretty smart bots for those sort of games already?

If the Game AI is stupid it's usually a game where you're supposed to play against many enemies at a go, and that's why it's crappy.

If it's not stupid it's one where you play against very few enemies. Maybe even just one at a time.

Re:Most humans aren't that smart (0)

Anonymous Coward | more than 5 years ago | (#25623943)

Yeah, you're right. I come to LAN parties to play against AI.

LAN parties are popular because you get to kill people and even unskilled players get their share by getting that one lucky kill on someone superior. It's not the same thing as getting pwned all the time by a computer program.

Re:Most humans aren't that smart (2, Funny)

Genrou (600910) | more than 5 years ago | (#25624875)


Personally, I want to see the day where two AI's argue on the battlefield...

And then shouting "LeRoy Jenkins!!" and running away to kill enemies.

But that might not be that intelligent...

Re:Most humans aren't that smart (2, Interesting)

nick_davison (217681) | more than 5 years ago | (#25631639)

Huh, you must be right. I guess that's why online gaming and LAN parties are so incredibly unpopular.

Everyone still hates campers. Most LAN party games have rulesets that discourage it.

If you have rapid respawns, the guy who runs out, gets fifty kills vs. twenty deaths, is going to massively outscore the guy who nurses his last two percent of health and snipes five guys from hiding.

Similarly, with short round times, it's much more fun to try something a little bit crazy so you can tell your friends about how you got a last point of health kill. After all, you know there's another round starting in thirty seconds.

Ask yourself how many LAN party games really encourage a real world fear of pain, desperately trying to keep the one life you're given, etc.

The games don't encourage realistic play from players. Why on earth would you want it from AI?

And, for that reason, saying "AI could be more real because LAN parties are fun" doesn't really hold up - the humans aren't playing realistically, due to the rulesets imposed, who wants the AI to either?

Re:Most humans aren't that smart (1)

Allicorn (175921) | more than 5 years ago | (#25649269)

Already done - Quake 3 Arena's bots routinely discuss their opponents' performance during play. From a logfile, Doom and Major both being bots:

    0:32 Kill: 3 1 10: Doom killed Major by MOD_RAILGUN
    0:34 say: Major: Camping AGAIN doom? Didn't your therapist tell you to stop?

TBH, where this falls down on being convincing is probably only that it's spelled and punctuated so well.

Alli

Re:Most humans aren't that smart (3, Insightful)

kvezach (1199717) | more than 5 years ago | (#25623865)

Sure, it's easy to make hard enemies. Just look at the Duke 3D bot. Absolutely no brains, but it moved like a rabbit on crack, and therefore it would beat you every time (unless you had a great advantage).

However, bots like that aren't any fun. It's more fun if the bots have limitations that at least somehow resemble the limitations players have; can't turn quickly, nor move too fast, know the map by instinct, etc.. Then, within those constraints, use AI and use AI well.

Put a limit on how many commands the enemy can do in a certain time for FPSes - or a limit on the rapidity it can issue commands in RTSes.. and suddenly you have a much more interesting problem. Or for that matter, let the player decide how smart the enemy should be, and whether or not it can cheat (issue a thousand orders in a second?). If he likes playing against a cheating bastard, let him play against a cheating bastard; if he wants to play against a chessmaster with no mobility and all brains, let him do so, that he'll be surprised when it still beats him.

Yes and no (4, Insightful)

Moraelin (679338) | more than 5 years ago | (#25624007)

1. I'll somewhat disaggree about RTS. Sure, in theory it sounds good, but I've yet to see a strategy game where the AI doesn't plain old cheat to stay alive.

Yes, it can "micromanage" better, in the sense of cycling through all units every frame or couple of frames. Sure, it can do a "for" loop better than a human. But actually allocate resources intelligently, apply smart tactics against a human who built in the unexpected place, etc, is where computers are still as stupid as it gets.

Again, we could argue about theoretically being easier to make them smarter, but way I see it, that's the basic rule of thumb: does it need to cheat? Does advancing in the storyline to face a new enemy, described as more cunning and ruthless, actually just mean the same retarded AI with a bigger pre-built base and more silos full of spice/tiberium/whatever and more reinforcements out of nowhere? Does it just mean that the new "cunning and ruthless" enemy just gets better units from the start? Does upping the difficulty actually just means that the AI gets even more money and a damage bonus, as opposed to just un-hobbling that supposedly super-AI a little?

If any of those are true, no, you have _not_ coded teh uber-AI and then dumbed it down for the player. You can claim to have a too smart AI when it can start just with a town centre and 2 peons, just like the player, and put up a better fight. And, oh, make it run just as long a way to the mines/geisers/spice-fields as the players, at that. Starting with 3 resource nodes in an already built base doesn't quite qualify as equal difficulty.

2. Yes, chess makes a good poster child, but that has had decades of real AI research into just that speciffic game, and at that by real AI researchers. A brand new game, with brand new rules, within 3 years, and with the cheapest team possible... heh... sorry. I can't take that seriously.

Most of the games so far have even trouble pathfinding, or keeping their flamethrower guys from frying their own team mates in front of them, etc. Or look at bigger scale strategy games, like Paradox's, where it takes several years of patching just to get the AI to no longer waste its whole army attacking Switzerland. And even then often the "fix" isn't as much AI, as just making fortifications randomly disappear in combat, so eventually the mountains around Switzerland just stop giving a defense bonus. You know, 'cause apparently 100,000 soldiers with rifles and grenades can demolish a mountain. And then invariably it becomes vulnerable in some other way, to some new exploit created by the previous fix.

3. Ditto for FPS. What the computer has as an advantage isn't really better AI, it's unerring accuracy. It's trivial to make a bot that never misses, and has faster reactions than any human, because it simply needs to calculate the angle and pretend it aimed accurately that way. It simply doesn't have the whole issue of moving the whole arm with the mouse, or the finite resolution of the mouse, or the whole lag of the pipeline from mouse to seeing the cursor move (the TFT alone introduces another 1-2 frames lag) which is already known to ruin one's accuracy because it lets you overshoot before seeing any results, etc.

Some cheat even further, by basically having eyes in the back of their head, or being able to see through walls, or just not having the issue of "does that 3 pixel tall figure over there look like one of our guys or one of theirs? Is it even a human?" It just doesn't have to parse an array of pixels, it already knows where everyone is. Even when you say "if you hear the player at position X", you're already cheating. A player can at most judge "I hear some footsteps in that general direction" (and even that at best in 30 degree increments), but not an exact position, nor know if it's a friend or foe or neutral there.

Give the computer a spread comparable to a human for the selected difficulty level, give it a similar lag in reacting and turning, and limit it to the exact same field of view, and that supposed AI advantage vanishes into thin air right there and then. It wasn't great AI that slaughters the player, it was essentially just a dumb AI with an aimbot and a wall hack and an assortment of other cheats. It's just proof that even an unskilled idiot can own if given all those cheats.

4. That said, in the end, I don't even care that much about that, it was mostly to point out the holes in that line of reasoning. In the end, I don't care if the AI really _is_ smart, but I care if it looks like it has at least the basic training.

Make it actually look like it tries to take cover. Make it act like a squad, not like anarchistic yahoos who just stole some military uiforms. Make it lay suppression fire when that squad advances. Yes, give one of those squad members a Dragunov or a USMC Designated Marksman Rifle and have him hide behind a crate and snipe.

Choreograph it a bit, basically. Even if it has to be easy, reduce their damage or whatever, but make it at least look like they're giving their best effort.

Re:Yes and no (0)

Anonymous Coward | more than 5 years ago | (#25637193)

What I'm hoping for a couple generations down the line is an AI that learns from players. So, you want to find out where all the safe spots, shortcuts, and surprisingly advantageous rallying points are? Watch the players. To take the example of FPS games, the algorithm would start with a full recording of a set of human v. human matches on a map. It would then begin creating a node map for pathing through the map based on player movement. It could generate statistics for various nodes, which would note hesitation in the vicinity of a certain node (camping locations), traffic at particular nodes (identify bottlenecks, killing fields, and genuinely out-of-the-way paths), weapon selection patterns at certain nodes (to determine when to use pistols and grenades, and from where to fire long-range weapons?) inter-node transit time (to seek efficient paths between procedurally-generated or pre-ordained points of interest), survivability by location (find nodes where players consistently get fragged to identify dangerous spots, then weight other paths preferentially), and offensive advantage by location (find locations where players kill more than they die, and prefer said places). Special attention could be paid to players under certain easily-identified conditions, such as flag carriers, support roles, piloting, and so forth. Distribute the computational task to many, many gaming systems, and give e-points of some sort to big producers. Once you have a good amount of this data ready, with routes and nodes prioritized by a few indexes, you're essentially down to a traveling salesman problem, where you can scale the AI the same way chess systems are scaled: by increased depth of analysis (cpu time). Add in some randomness (options to take suboptimal paths, just as players do), and you've got an AI that should be doing something that looks more similar to what the players themselves are doing.

This could work in many different game types. To me, it looks pretty darn attractive in RTS games; you could have the algorithm study the players' scouting techniques in order to finally turn off AI maphacking, learn how players build on particular maps, determine what progressions seem to work best, or even adopt realistic micromanagement patterns (even master players can't make a hundred space marines continuously go man-to-man, after all; from what I've seen, "micro" is a matter of spreading fire evenly onto important secondary targets while trusting auto-fire to predictably hit a certain type of primaries). It could work on action platformers, piloting games, and rpgs, as well. Care would have to be taken to be sure that the AI would monitor its own progress and identify behaviors that seem to work well for players, but poorly for the AI (player maneuvers that exploit unforeseen play forces, non-obvious prerequisites, or bugs in game code in ways the computer can't reliably mimic).

Oh well, it's a pipe dream right now. Might it be doable sometime in the next decade?

Re:Most humans aren't that smart (2, Interesting)

Anonymous Coward | more than 5 years ago | (#25624519)

Who says AI has to be used to make opponents harder?

It could also be used to:
- Make opponents have conversations with each other.
- Exchange knowledge, gear, supplies, etc.
- Boost or ruin other opponents morale.
- Live as they should and not just stand still waiting for the player.
- Trade with each other.
- Play card games, chess with each other.
- Get recreational and make sand castles when bored
- Have sex with each other
- Go explore surroundings and remember them afterwards.
- Disobey orders
- Think new orders for themselves
- Be opportunist
- Get interested about mundane details like rocks if they happen to have geology training

Re:Most humans aren't that smart (2, Informative)

default luser (529332) | more than 5 years ago | (#25630865)

Wouldn't that be something?

Unfortunately, there's no way to produce an AI like this, because each one would be a work of art. The immense amount of time it would take the programmer to construct personalities like the above from the ground-up would be prohibitive, and no amount of tools could streamline this.

Really, this is the hardest part abouut AI design: classifying the entire human existence into easy-to-handle pieces. Unless you can successfully generalize human experiences and tendancies into neat little packages, there's no way you can create such an impressive AI as the above. You would spend too much time just doing each AI by hand.

Re:Most humans aren't that smart (1)

Velocir (851555) | more than 5 years ago | (#25635939)

They actually did something similar to the grandparent's suggestions with the Orc AI for the battle scenes in the 2nd and 3rd LotRs movies (yes, IAAKiwi). They just had a whole lot of characteristics, with acceptable levels of variation between each orc, and then they'd press play and observe how those characteristics played out. They found that some orcs would retreat once they started losing the battle.

Re:Most humans aren't that smart (2, Interesting)

default luser (529332) | more than 5 years ago | (#25647623)

Sure, that's fine and good if you limit your character to an "Orc," and limit their exposure to "field of battle." Limiting scope is the easiest way to get AI that "works."

A good example of limited-scope AI is the shock troopers from the original Half-Life: despite poor hardware specs, the troops reacts to player attacks, falling back and covering their movements with fire/grenades. This only "worked" because the troopers had a very limited list of actions to select from, and the paths the troopers could follow were already defined in stone. Just watch them attack or retreat: they will always launch attacks from or fall-back to the same strong points, using one of perhaps three pre-defined paths.

What I'm talking about is truly diversified AI - is your Orc male or female? Does it have an outgoing or introverted personality? Does it start fights, or run away from them? Does it hunger for something, or does it lack drive? Does it try to be well-liked by everyone, or does it care more about others' happiness? Does it pick it's nose? Does it dance and sing? What exactly makes your Orc engaging, interesting, pathetic, predictable and unexpected, all at the same time?

All the above factors and more can be described to create a more well-rounded AI, one that you can use in all sorts of situations. I mean, limited AI in limited situations has it's place (it makes a great background for a movie scene, or a well-scripted game encounter). But ultimately, what people want to see is AI that does something totally unexpected, like sontaneously get into an arguement with superiors, try to lie, etc. Unfortunately, indentifying all these various fragments that make-up the culmination of an AI's tendancies is difficult, and requires a good AI algorithm to put the data to good use.

Re:Most humans aren't that smart (2, Informative)

redscare2k4 (1178243) | more than 5 years ago | (#25624561)

I don't agree. In your example, you can make those groups of enemies flank the player but give them low accuracy, for example. So in low difficulty the AI tactics are smart but their competence is low. And those of us that like masochistic difficulty levels would enjoy havin to put some mines to cover our back from those flankers.

Re:Most humans aren't that smart (2, Interesting)

Paradigm_Complex (968558) | more than 5 years ago | (#25627003)

Yes, increasing the AI could very well make the games fun for many people. What you're talking about is not exactly the kind of AI that's needed. There is a very big difference between being able to beat the human player and being able to outwit the human player. It's not necessarily difficulty or realism that's needed, it's creativity and genuine surprise.

I want a game where the AI is smart enough to send a probe over and throw assimilator over the player's vespene geyser to keep the player behind in tech. I want an AI who will excessively use flash-bang grenades and then suddenly switch to a smoke grenade so the player will look away to avoid being blinded while the AI runs in and shoots him up. I want an AI smart enough to toss a veggie at his partner to give his partner back his up-b and help him recover back to the stage.

These things may be possible today by coding in the exact individual ideas, but human players will quickly learn all the very limited tricks the AI knows and become bored with the AI. Still, it's better than what the vast majority of games have and I'd really love to see such things. What we really need is AI that's creative. That would make games much more fun.

Re:Most humans aren't that smart (2, Insightful)

manekineko2 (1052430) | more than 5 years ago | (#25627493)

In most FPS games, stupid humans want to be able to mow down _thousands_ of stupider computer controlled enemies - "against the odds". That's what makes it fun for them.

That's just not possible if the enemies start having a lot more brains. Then most players might have difficulty getting past the first 3 enemies :).

It's not that difficult to make an enemy FPS "bot" have superb tactics, coordination, timing etc. Especially if the map is pre-known (which is usually the case). You can code the tactics and heuristics in. If you hear the player in position X, group A enemies head to position Y and group B head to position Z, and bye bye player.

Imagine if enemies that are low in health kept running away and hiding, and then snipe at you from far away when they see that you are busy doing something else. While that might be more realistic, it might not be so fun eh? Who really wants realism in games?

At that rate the player can never pretend to be the hero he wants to be. He'll just be dead. And your game won't sell.

I see your point, but disagree to a certain extent. Idiot AI in modern FPS really devalues the successes, at least for me, to a large extent. It should be an incredible accomplishment for a single man to kill a squad of 5 people, but it's just one encounter, and I feel nothing afterwards. If the AI could be juiced up in these games, then you wouldn't need to throw a 1000 enemies at the player to make them feel like a hero, because you'd feel more accomplishment from each encounter.

I recently just played through Deus Ex for the first time, and it is an outstanding game, especially for the time. However, the AI was really the Achilles Heel in terms of immersion in that game. At times, you can really get into games like that, but once the rules of the AI start to become apparent the whole experience starts to feel less like you're a hero, and more like a toy world. You shoot a guy in the chest, hide for a while, he decides it must have been the wind and goes back to his patrol without telling anyone. The whole thing becomes so exploitable that it just feels cheap.

Furthermore, in many FPS like Deus Ex or Halo, your character is so super-powered even with intelligent foes, he would still be a walking death-dealer. If he's not enough of one, you can always just make his shield that much stronger.

Lastly, when you add in intelligent AI for tactics, you'd also want to add it in for realism. Enemies that run low on health usually (though sometimes they might if their courage rating was high) wouldn't run away and snipe you. That's not realistic AI. They'd run away and cry that they'd been shot, and yell for help from their teammates who'd go and try and assist them.

Re:Most humans aren't that smart (1)

Ayavaron (971110) | more than 5 years ago | (#25632417)

I recently just played through Deus Ex for the first time, and it is an outstanding game, especially for the time. However, the AI was really the Achilles Heel in terms of immersion in that game. At times, you can really get into games like that, but once the rules of the AI start to become apparent the whole experience starts to feel less like you're a hero, and more like a toy world. You shoot a guy in the chest, hide for a while, he decides it must have been the wind and goes back to his patrol without telling anyone.

It was also pretty silly how people would just stand there if you threw a gas grenade at them. A group of seven or eight guards in a corridor could be dispatched with boring ease if you threw a gas grenade and took your time aiming so you could cap each one of them off with headshots.

Re:Most humans aren't that smart (2, Insightful)

Ringthane (415537) | more than 5 years ago | (#25627979)

"Just ensure that basic stuff like navigation is better. Stuff doesn't have to be that smart, but at least they shouldn't be totally stupid - they should be able to walk around stuff without getting stuck - even a "dumb" animal can navigate open spaces better than many computer controlled stuff in games."

A lot of games -- even high profile games like Halo 3 -- can't get this fundamental down. God forbid you have an AI character driving your Warthog anywhere. You're sure to ram into a tree or rock & get hopelessly stuck. Some older games have done it right. The original UT had a nice user-variable scale of skill for its bots & you could opt for bots which ramped up their skill to match your performance. And the original Half-life grunts seemed like NPC geniuses to me: retreating when under pressure, throwing grenades to cover their retreat, & then flanking... That game is 10 yrs old, but few games seem to have achieved that level of AI optimization.

Re:Most humans aren't that smart (3, Interesting)

maugle (1369813) | more than 5 years ago | (#25628417)

As far as RTS goes, I'd like to see the AI act more human. Not just in how "smart" it plays, but in how quickly it can act.

Remember in Warcraft 2 (I've started playing it again recently, so I'll use it as an example), how you had to send roughly double the number of troops as the enemy to get a fair fight? The AI would have all its ogre-magi cast bloodlust (or all its paladins cast heal) simultaneously, while you'd be struggling to get a single spell out.

And while that was going on, they'd be effortlessly churning out more units. During the time you're looking at your base and getting some more units ready, your units in battle would be getting slaughtered.

So, what I'm really looking for in an AI is human-like delays between commands.

Re:Most humans aren't that smart (1)

3.14159265 (644043) | more than 5 years ago | (#25628579)

"It's not that difficult to make an enemy FPS "bot" have superb tactics, coordination, timing etc. Especially if the map is pre-known (which is usually the case). You can code the tactics and heuristics in. If you hear the player in position X, group A enemies head to position Y and group B head to position Z, and bye bye player."

I'd say the holy grail is to have the AI without pre-knowing the maps, and without having to hand-code the small details.

Re:Most humans aren't that smart (0)

Anonymous Coward | more than 5 years ago | (#25631767)

"Creating an AI that can consistently beat humans is not hard."

Actually, yes it is - because the challenge is creating an AI that can consistently beat humans given the same information and resources as a human, and fitting within available CPU power.

In other words: it's not hard to make an AI win if you let the AI cheat. It's hard to make an AI play well while keeping the gameplay fair.

(And hard-coded tactics and heuristics aren't quite fair either, though the developers can use those to great effect for pre-scripting things, like specific missions in single player campaigns in an RTS).

Re:Easier to just cheat! (1)

sgt scrub (869860) | more than 5 years ago | (#25626233)

I've been playing Fallout. I think the "I" is pretty good. If designers could decrease the "A" I'd be happier. IMHO dedicated hardware for AI is something that desperately needs to be tapped. If someone could figure out a way to implement persistent neural nets the device could be used for an array of applications, not just games. ie. OCR, voice/face/fingerprint recognition, hearing/vision aid... I've always wanted to try hacking a graphics card to do it but they don't have any persistent memory. Cards that have memory don't have any good processing capability. I know I'm not alone though which is good. http://games.slashdot.org/article.pl?sid=06/09/06/1749252 [slashdot.org]

An example of great game A.I. (3, Informative)

B5_geek (638928) | more than 5 years ago | (#25622449)

The one game that has always stood out in my mind as having great A.I. was Comanche Maximum Overkill. The original (386DX-40 era) DOS game actually advertised in the manual that if you repeat the same attack pattern for 30 seconds then the game would adapt, AND IT DID!

Imagine this scenario; you are in a helicopter hiding behind a hill. Whenever a bad-guy gets close enough, you pop-up above the hill, get a missile lock, fire, then drop below the hill. If you repeat this pattern long-enough (30+ seconds) then enemy copters will sneak up behind you and blow you up. I was always impressed at this "Learning A.I." as opposed to what most computers games do.

RTS/TBS: build stuff quicker then you can and/or advance technology faster then should be possible.
FPS: Have 'super accurate' shots, higher health, bigger guns.

Re:An example of great game A.I. (3, Insightful)

Zephyrmation (1372025) | more than 5 years ago | (#25622657)

RTS/TBS: build stuff quicker then you can and/or advance technology faster then should be possible.
FPS: Have 'super accurate' shots, higher health, bigger guns.

This to me is a huge downfall of modern games - instead of making AI opponents "smarter", devs simply tweak the rules to give the AI more of an advantage.

That being said, it is incredibly hard to define an AI that doesn't have "unrealistic" skills when the players' skills are advancing in the same fashion. For example, your skill in Halo is to a large extent determined by how accurate you are, which is easily mimicked by AI. I can't count the number of times I've heard someone accused of using an "aimbot" because their skill (or luck) in an FPS seemed "too good" or "unrealistic". The same goes for RTS games - the top human players in the world are to a large degree measured by how many commands, or actions, they can perform in a minute - which is again easily transferred to an AI opponent.

In my humble opinion, what we need is some sort of standardized test for game AI - put one player and one computer opponent in the game with the exact same capabilities, and see who comes out on top after repeated rounds. After all, it's impossible to claim the AI is "cheating" if it can't do anything that you can't. And whoever learns from the other's actions best over the course of the game will come out on top. And if at the end of the day the two are evenly matched, I will happily put down money for the game.

Re:An example of great game A.I. (2, Interesting)

Rysith (647440) | more than 5 years ago | (#25622755)

The problem there is that the two factors you mentioned (accuracy and commands-per-minute) are both things that AI can far exceed humans at, especially if you aren't careful to limit it. I think that the real solution is to make a game where learning and adapting is more important than accuracy or speed, but then someone would have to write actual AI.

Re:An example of great game A.I. (2, Insightful)

Dun Malg (230075) | more than 5 years ago | (#25622803)

The problem there is that the two factors you mentioned (accuracy and commands-per-minute) are both things that AI can far exceed humans at, especially if you aren't careful to limit it. I think that the real solution is to make a game where learning and adapting is more important than accuracy or speed, but then someone would have to write actual AI.

Yeah, wouldn't it be nice to have games like that again? I blame the original "Nintendo generation" games for the proliferation of the abominations that are RTS games. In my experience, the contrived demands of the "Real Time" aspecs so completely dominate the game that what little "Strategy" remains might as well not exist.

Re:An example of great game A.I. (0)

Anonymous Coward | more than 5 years ago | (#25622965)

Adding to the parent: Unfortunately many RTS games have well-known "optimal" strategies for various situations, but since they're more focused on RT than S, the games still force you to execute the strategy in real time.

If you're not yet familiar with the optimal strategy, the challenge becomes roughly equivalent to a platform game where you're required to use the standard platformer N^2 learning progression. Basically you have to execute perfectly to step N and then try something that usually results in failure. Then once you get past step N, you have to repeat the learning process on step N+1.

However, even if you are familiar with the optimal strategy, you still have to execute perfectly. Personally I'd rather click "move zig, for great justice" and let the game AI execute the real-time aspects of my chosen strategy. I think the parent would agree.

Re:An example of great game A.I. (2, Informative)

KDR_11k (778916) | more than 5 years ago | (#25623141)

You know how many people try to automate seemingly trivial things with scripts in Spring [clan-sy.com] ? You quickly run into a barrier because while the strategy may be standard the specific execution must be adapted to the situation. E.g. a commonly requested feature is an auto-dgun widget for Total Annihilation clones/ripoffs but there's tons of factors to account for, the thing costs a lot to use, it can easily cause friendly fire that, depending on the situation, could be acceptable or unacceptable and it has to be aimed properly so the shot actually connects. If you automate that the enemy player will quickly learn the pattern of your automation and adapt to use it against you. That's why the tactics cannot be automated easily, a good player will employ tactics that beat your AI. Well, unless you make it impossible to override the AI's decisions but then you'll end up with units doing stupid things and players learning the situations in which the AI behaves well or badly and exploiting that.

Real time strategy is really more on a tactical scale most of the time, strategic advantages can be countered by good tactics and a player must handle both the macro and micro (pretty much strategy and tactics) at the same time to get the maximum efficiency out of his troops. Any predictable behaviour is going to be a weakness.

Re:An example of great game A.I. (4, Interesting)

SnowZero (92219) | more than 5 years ago | (#25622919)

Education can help; A lot of college CS programs don't force the breadth that an "AI game developer" would need. In my undergrad degree, graphics, AI, game programming, and distributed systems were "applications" classes, and you only needed one or two. Usually the game programmers would have to take graphics, because even if your the sound or networking guy, they'd expect you to know graphics like they did. If more game programmers had taken systems classes (such as operating systems), I don't think they would have had as rough a road with multicore either.

However, even given the current narrow classes, you can at least try to bleed through enough of the wider topics into game-oriented classes to get people footed. In the game programming course I was a teaching assistant for, I gave a couple lectures on AI (I was an AI/Robotics grad student). Then we gave them an AI-only assignment. It was a multiplayer tank game where we pitted their AIs against some test opponents, and then against one another in a tournament. It didn't allow much cheating in the AIs (only global visibility, where a player can see every unit at all times, which is nearly universal in game AIs). The assignment really seemed to be a hit, and hopefully for those students that went on to the games industry, it gave them the basis to branch out and learn other deeper AI techniques.

It would also help if the AI community wouldn't look down on things such as games as "lowly applications". In some sense I think game programmers would be best off talking to robotics people or even web machine learning people (spam filtering, web search ranking, etc). Those people are already doing applied AI, and particularly for robotics folks, working on many of the same problems that good game AIs would face.

Re:An example of great game A.I. (1)

Rockoon (1252108) | more than 5 years ago | (#25623057)

I think that you are wrong about the AI community looking down on games as "lowly applications."

I think that many posters here are confusing Artificial Intelligence with Machine Learning. The later is a subset of the former, but is often difficult to apply to games. Accepting adaptation is typically equivient to abandoning Game Theory.

Certainly there are entire fields of A.I. that are entirely unrelated, but there are many fields of A.I. whos core development is exclusively related to games.

Most games do not implement any heavy A.I. techniques because it is too difficult to provide skill gradients: Easy, Normal, Hard, Godlike. These skill gradients are pretty simple to implement as an escalation of "cheating," but not so simple as a tweak to AlphaBeta and pretty much futile with Machine Learning.

Re:An example of great game A.I. (4, Interesting)

SnowZero (92219) | more than 5 years ago | (#25623243)

I think that you are wrong about the AI community looking down on games as "lowly applications."

Well, at conferences such as AAAI, even robotics isn't treated as more than a "mere application". At least that's the treatment I felt going there and presenting work. Its ok I guess, because AI theory doesn't directly relate to where most work in applied AI happens (just like in applying machine learning -- most of the work is in the features, not the algorithms). However, I feel there is a real gap between what conferences such as AAAI are willing to embrace, and what happens at game development conferences or robotics conferences. Being someone who did RoboCup for many years, I really do know what it is like to span that gap -- some of the most important breakthroughs we made were not publishable in either type of conference. I really feel that the field lacks something like the AI equivalent of JGT (Journal of Graphics Tools) or Graphics Gems. If I were a more professorial type I might try to start something like that, but instead I just wait and hope someone else will, and then I could contribute to it.

I think that many posters here are confusing Artificial Intelligence with Machine Learning. The later is a subset of the former, but is often difficult to apply to games. Accepting adaptation is typically equivalent to abandoning Game Theory.

Other than the fact that applying machine learning successfully is difficult in general, I don't know if I really agree with that statement. Game theory is a subset of AI just like machine learning. Game theory is very important for abstract games such as chess, but I'd argue that most "physical simulation" games need things more like potential fields, advanced motion planning, high quality hand coded policies, and geometric stuff of that ilk. Old stuff like rule base AIs really has a place in games too -- the work done on scaling up expert systems is really like "software engineering for AI", and sadly a lot of that work didn't get published either.

A simple form of learning we used in our robotics work was weighted experts given a set of hand-coded policies (~= "a set of AI strategies" for the non AI people out there). We used that to learn during 30-minute autonomous robotics games against robotic opponents, and it worked, even during a relatively short game. All you need is a way to define successful subgoals (Ex: scoring in a team game, kills in an FPS, or areas won in an RTS) and you can get convergence to the optimal strategy in a logarithmic number of rounds. In a turn-based game, if those experts each applied game theory, you could have game theory and learning combined in a pretty reasonable way. Yes I know that would not be optimal, since you normally would just pick the strongest AI, but humans rarely play optimally, nor is even fun to always play against the same strong strategy.

Certainly there are entire fields of A.I. that are entirely unrelated, but there are many fields of A.I. whos core development is exclusively related to games.

True. I guess I just haven't really seen the dots connected. Then again, I haven't been to any AI conferences in the last year or two, so maybe that has changed already. I hope so.

Most games do not implement any heavy A.I. techniques because it is too difficult to provide skill gradients: Easy, Normal, Hard, Godlike. These skill gradients are pretty simple to implement as an escalation of "cheating," but not so simple as a tweak to AlphaBeta and pretty much futile with Machine Learning.

Well, here's another place where I think the communities need to work more closely. Coming up with search strategies that are more human-like when measured statistically would be fascinating work. Limiting search depth and random mistakes works in some games, but we could do a lot better. I think some of the commercial chess games already do a pretty good job of various skill levels -- though I fear in that case it's just that chess has been studied so much, rather than some general theory that other games could draw from.

It's true that most of the initiative has to come from the game development community, since they are the ones who should want to improve their AIs. My point was mainly that it wasn't entirely their fault, as many seem to think. Graphics researchers did an incredible job at making their advanced work eventually accessible to normal software engineers, and I think those of us in the AI community (however fractured) could do a better job of that.

Re:An example of great game A.I. (5, Interesting)

Dutch Gun (899105) | more than 5 years ago | (#25623369)

The problem there is that the two factors you mentioned (accuracy and commands-per-minute) are both things that AI can far exceed humans at, especially if you aren't careful to limit it.

You're correct. I've written AI for a number of commercial games. Some of the most challenging AI is for games in which the players are competing with the AI on what are supposed to be equal terms. An AI can home in on a player's forehead with a sniper rifle with little difficulty. It's a simple mathematical equation. How do you simulate the aiming a player has to do?

The solution I came up with was to put the target's aim point on a set of springs attached to the player. By jumping around and changing direction quickly, the player would tend to throw the bot's aim off (imaging the target bouncing around, attached by the springs). But, stand still or move in the same direction for too long, and the bot would home in on the player. And, of course, just like a human player, the AI would get in a lucky shot every once in a while as the target crossed in front of the player.

You have to come up with creative solutions to make the game "feel" fair. That's not the kind of stuff that's typically taught in college courses. Naturally, formal training doesn't hurt, but there are a lot of challenges unique to game development.

Re:An example of great game A.I. (4, Insightful)

TheLink (130905) | more than 5 years ago | (#25623443)

Yeah, most people don't seem to get it. They show how stupid/ignorant they are by asking for "smarter/stronger" opponents.

It's trivial to program a computer to beat humans in most games.

But the problem AI is supposed to help solve in games is not "How to beat the human".

The problem AI is supposed to help solve is "How to make it _fun_ for the human, so that lots of humans will pay $$$ to play".

Most people would have near zero odds against a top notch computer opponent - FPS, RTS, whatever.

Does anyone really think that Starcraft, Doom etc would have sold so much if playing them was like playing against a World Champion or two?

I've seen the top humans play in FPS and they can aim pretty well. You'd never be the Hero winning against the odds if the _thousands_ of enemies you fought were even only half as good as a world champion.

Re:An example of great game A.I. (0)

Anonymous Coward | more than 5 years ago | (#25624613)

I disagree. I want an intelligent opponent that has the same accuracy as I do. In FPSs, über bots pwn me with accuracy weapons(railgun,etc.) and their inhuman accuracy, but if I do an all rocket game, where they need strategy and tactics, they don't know what hit 'em. Oh, and the inhuman accuracy is not the AI, it is an aimbot. If you were doing actual AI(optical recognition) you wouldn't hit a snail.

Re:An example of great game A.I. (1, Interesting)

Anonymous Coward | more than 5 years ago | (#25624615)

Again, AI don't have to be used to make opponents harder or easier. Just to make them more human.

For starters, all computer opponents should have as accurate FoV as what player sees. Yeah that's right, calculate the 3D FoV for all opponents too. In realtime as they are also moving around.

Without real FoV, there's no point trying to do opponents that explore their surroundings, don't know beforehand what's behind new corners, can remember where they've been, etc.

In any case, AI should not know where player is unless AI: either sees where player is, other AI tells it where player is or AI has tracking skills and player actually left tracks.

Re:An example of great game A.I. (1)

delt0r (999393) | more than 5 years ago | (#25626421)

Not in RTS I'm afraid. Its too open for good AI, the state space is too big, since things the computer can do well is already done for the human player too. Its like good chess AI verse good Go AI.

Re:An example of great game A.I. (1)

SupremoMan (912191) | more than 5 years ago | (#25622775)

This is the reason I only play strategy games where I can play other people. The game is totally different when everyone must follow the same rules. I can't tell you the amount of times some jackass claimed to be a god of civ4 or some other nonsense, and got his shit pushed in in a multiplayer game like he didn't know a scout from a tank.

Re:An example of great game A.I. (1)

antifoidulus (807088) | more than 5 years ago | (#25623031)

And that is part of the reason you probably won't see a massive investment in game AI. Its much easier to add networking to the game than it is to try design, implement, and test a good AI.

Re:An example of great game A.I. (3, Informative)

SnowZero (92219) | more than 5 years ago | (#25622827)

This to me is a huge downfall of modern games - instead of making AI opponents "smarter", devs simply tweak the rules to give the AI more of an advantage.

Indeed. Cheating AIs make me cringe. I'd really rather see a dumber AI that doesn't know where every unit on the level is, than play a more "skilled" one that's just using the fact that it's not playing the same game to gain its advantage.

That being said, it is incredibly hard to define an AI that doesn't have "unrealistic" skills when the players' skills are advancing in the same fashion. For example, your skill in Halo is to a large extent determined by how accurate you are, which is easily mimicked by AI. I can't count the number of times I've heard someone accused of using an "aimbot" because their skill (or luck) in an FPS seemed "too good" or "unrealistic". The same goes for RTS games - the top human players in the world are to a large degree measured by how many commands, or actions, they can perform in a minute - which is again easily transferred to an AI opponent.

It's hard to define, but not necessarily hard to measure. Record a bunch of humans playing, look at plots of where they aim based on location, distance, velocity, etc, and build a statistical model. Or, if you've got something more algorithm based, measure it the same way and make sure its distribution on plots looks fairly human. Of course, game companies will have to be willing to hire statistics/AI type people (or train their devs in those areas), and devote the money to gather player data and time to make it happen. So far few companies have gone that route, but I think more will in the future.

Several years ago I went to an AI conference (AAAI), and they had a quake bot that was coded by some rule-based AI experts (SOAR bot). I was into quake quite a bit at the time, and I'd have to say that was the most fun human-like AI I've played in a game where the player and AI had equal footing (same unit(s) and capabilities). I don't think the game companies have been willing to hire those kind of people yet, but as I said before, hopefully that will change, especially once customers realize that screenshots don't always mean good gameplay.

P.S. I did AI/Robotics for my degrees, and work on machine learning for a living. Haven't worked in the game industry, but I've worked with a bunch of people in school who have gone into that.

Re:An example of great game A.I. (2, Interesting)

hardburn (141468) | more than 5 years ago | (#25622897)

I suspect that all the devs say they want a great AI in their game, but when deadlines start to come up, AI is one of the first things to get cut. That's why every RTS in history that got a preview in a magazine a year before release promised a "groundbreaking AI", and yet the same game when released still has ore trucks driving around a hill, across three bridges, and through the enemy base, just because that particular piece of ore was the closest in a straight line.

I noticed devs getting slightly clever of late. In C&C: Generals, the initial resource piles are right next to your base, and harvester-type units don't go to resource piles that aren't in a certain range. Plus, the late-game economy depends on things unrelated to those resource piles. So the underlieing problem still exists, but is rarely noticed with the way the game works. Which is some kind of progress, I guess.

Re:An example of great game A.I. (0)

Anonymous Coward | more than 5 years ago | (#25624291)

Indeed. Cheating AIs make me cringe. I'd really rather see a dumber AI that doesn't know where every unit on the level is, than play a more "skilled" one that's just using the fact that it's not playing the same game to gain its advantage.

The trouble is, that the player may not see the difference between a good AI playing the same game as the player, and a poor AI that cheats.

I once chatted to the AI programmer for one of the C&C clones that came out several years ago. He put a lot of work into making the AI try to make intelligent decisions, mostly by using a fairly large rule set. The AI got it's information through the same calls that the player used, so it was subject to the same fog of war and vision limitations as the player. However, because the programmer put in a number of rules that let the AI set ambushes in likely spots, players 'proved' that the AI 'must be' cheating, and the game was loudly condemned for it's poor AI on the net. Of course with the fog of war effect, the players never saw all the unsuccessful ambushes that they never triggered, so they just assumed the AI knew where to look for them.

It's hard to define, but not necessarily hard to measure. Record a bunch of humans playing, look at plots of where they aim based on location, distance, velocity, etc, and build a statistical model. Or, if you've got something more algorithm based, measure it the same way and make sure its distribution on plots looks fairly human. Of course, game companies will have to be willing to hire statistics/AI type people (or train their devs in those areas), and devote the money to gather player data and time to make it happen. So far few companies have gone that route, but I think more will in the future.

There was a Formula One racing game that did something along those lines. One of it's selling points was to have each AI driver race like the actual F1 driver it was named for. They ran the neural net in a training mode where rather than aiming for the best race time, it was aimed at producing the average split times for that particular driver on that track over the preceding few years. It worked surprisingly well, in that the bigger name drivers ended up with much more aggressive race lines through the course. Unfortunately, although several reviewers commented on the realistic feeling AI, none of them were willing to add points to the games final score just because the AI was really good. As far as I know, the next racing game done by that studio had a much more standard AI, that took far less time to code.

Re:An example of great game A.I. (3, Interesting)

tibman (623933) | more than 5 years ago | (#25623403)

I've been in love with NeuralNets since first sight. After a few projects i really wanted to try my hand at a game. I came up with an idea for an RTS that was something like this: The majority of your units were ai, FF nets that evolved genetically. So you don't have to micro manage your economy or patrols or anything. You say, i want a factory here and your builders will find the resources needed and build that factory. They will have to learn the optimal way of doing this. The most successful individuals are the blueprints for the next generation. If that building is destroyed your command still stands, you want a freaking factory in that spot.. so they will toil away and continue building it. The battle side was supposed to be similar but i haven't gotten around to it. But the idea is an RTS that frees you up from the micro & details and let's you focus on the strategy aspect. A side effect is your units become unique over time, so even though your untis may be very successful on one map, a different type of terrain really makes life difficult for them for a few generations.

Right now they learn how to gather materials and move it to construction sites. Actually a lot of fun to watch. But it is still a long long way from anything resembling a game.

Back on topic though. An AI that adapts to it's opponent's weaknesses is well within reach of any game developer. The tough part with genetic and nnet AI is lack of control. You can't easily script a sequence and you can't guarentee what the user/customer will experience. I have had degree's of success over "difficulty" by restricting how many ticks of "thought" a net gets per second. But even then, the net will eventually adapt to this slow thinking.

Re:An example of great game A.I. (0)

Anonymous Coward | more than 5 years ago | (#25623491)

Me and a friend ran back in 2004 an project to improve Warcraft 3 AI, and at the end we ended up with just this.
It was one of my goals making the AI and actualy "AI" and not some kind of build sequence.

Example, with our AI you only have to type, Build 3 gryphon riders(Human end tier high tech unit) and the AI will find its way there on its own, expand, build workers, gain resources(expans i the algorithm decides it can reach the goal faster that way) and build its techtree in the most optimal way possible to get to these 3 gryphon riders as soon as possible.

I always been a big fan of RTS game AI and try to create improvement mods for most that I like, mainly looking forward to Blizzards Starcraft 2 right now that most likely will have good potential for great AI modding.

Sadly almost all(if not every) RTS game out there has AI inferior to any avrage human player, at least I havent found any game with challenging AI.

Re:An example of great game A.I. (0)

Anonymous Coward | more than 5 years ago | (#25625045)

Let me suggest you try a Real-Time Tactical game or something where the number of units is restricted. You're going to need a lot of neurons to be able to store a useful amount of patterns in a game of any significant depth, so if each unit gets its own net you're looking at a massive number of calculations in a game of any real size.

Re:An example of great game A.I. (1)

tibman (623933) | more than 5 years ago | (#25636625)

Very true, i've run into this ceiling many times.

The best solution i've come up with is keep the units as simple as possible. In many cases a FeedForward net with 4inputs, 8 neurons in a hidden layer, and 3 outputs scales pretty well and behaves like an intelligent insect. This was good for builders and other economy based units but seriously lacking for combat units. If the nets are simple enough you can even build hash tables for input & output and skip the all the computation completely for known values.

I really appreciate the suggestion. I'm having a hard time figuring out how to keep the player busy since most of the busy work is out of their hands. I'm going with an idea for a commander unit that you directly control (and gives bonuses) so you feel more involved in the battles & exploration. There is still something missing though : /

Globulation -- Re:An example of great game A.I. (1)

Tzarius (688342) | more than 5 years ago | (#25637967)

I've played a little OSS RTS game called Globulation 2, it sounds just like the game you describe. Even though it's unfinished, it's a lot of fun to play and, and to give these commands (such as "Make a building here", "Attackers stand here" etc). It suffers in larger/longer games when units trip over each other and starve to death, or cannot survive the trip from the battlefront to the inn, but I think these issues can be resolved.
Try it out:
http://globulation2.org/ [globulation2.org]

Re:Globulation -- Re:An example of great game A.I. (1)

tibman (623933) | more than 5 years ago | (#25645981)

Thanks a lot, checking it out now.

Re:An example of great game A.I. (4, Insightful)

cgenman (325138) | more than 5 years ago | (#25623025)

When working on the RTS Empires: Dawn of the Modern World, we had a difficulty setting that was truly insane. The enemy wasn't a computer pretending to be a human, it was an impenetrable black wall of impending dismemberment. If you really wanted to defeat it, you had to pile-on 7 vs 1, with at least 3 top-notch players overseeing the operation. Those were long, intense, brutal battles... Helping to bringing down that beast was a real badge of achievement. Unfortunately we had to cut it back a bit before release for technical reasons, but the ability is still there.

A realistic FPS would have the enemy sneak up behind you and stab you before knew they were there. A realistic racing game would end in firey death the moment you accidentally rode up on the curb. A realistic tactical squad shooter would have your men pinned down by heavy opposition fire until they called in an airstrike on you. A realistic war game would involve lots, and lots, of digging.

This is a long way of saying that AI isn't about promoting hyperrealism, but rather is in service of making the game fun.

That having been said, I'd kill for 3D pathing that doesn't suck. If I need to do another escort mission with some idiot who can't walk around a boulder in the middle of the road, I'm going back to Tetris and I'm never leaving.

Re:An example of great game A.I. (3, Insightful)

elrous0 (869638) | more than 5 years ago | (#25626909)

If I need to do another escort mission with some idiot who can't walk around a boulder in the middle of the road

That is why I hate, hate hate, HATE escort missions and missions where you must keep some AI character alive in games. It's one thing to deal with a stupid AI enemy (perhaps a little laughable, but not frustrating). It's quite another to deal with a stupid AI ally. Basically, in most of the games I play on higher difficulty settings, the ally AI's are just cannon fodder. Anything else is just damned annoying (especially when you can't even command them to get down or take cover).

There are a few games that do this right: Halo 3 (which makes vital character invulnerable and everyone else dispensable), Half-Life 2 (which periodically replenishes your cannon fodder AI allies and makes the one non-cannon fodder character near-invulnerable), and Mass Effect (where you can command your allies directly). And there are also games that do it wrong, like Oblivion (where your allies on escort missions don't scale as well as enemies, can't repair their weapons and armour, and do stupid shit like walk off cliffs).

Funny you mention Tetris (1)

tepples (727027) | more than 5 years ago | (#25628303)

That having been said, I'd kill for 3D pathing that doesn't suck. If I need to do another escort mission with some idiot who can't walk around a boulder in the middle of the road, I'm going back to Tetris and I'm never leaving.

Funny you mention Tetris. Since 2001, most Tetris games have had pathing bugs. Look at how this T piece jumps into an F-shaped hole [ytmnd.com] , which works in Tetris Worlds, Tetris DS, Tetris Zone, Tetris Evolution, Tetris Splash, and Tetris Party.

Re:An example of great game A.I. (2, Interesting)

VShael (62735) | more than 5 years ago | (#25623817)

Agreed.
Another great game to do this was Descent.

If you flew into a room, started shooting madly, then reversed out so you pick off the enemies as they moved into the doorway one at a time, they quickly learned.

Soon you'd find that the enemies wouldn't chase you. And would in fact, surround the doorway, as if knowing you would have to enter sooner or later... and then they'd all shoot you.

It made the game very replayable.

People man! (1, Interesting)

Anonymous Coward | more than 5 years ago | (#25622583)

Granted we haven't seen great AI in some time, I still rather play people any day of the week.

RTS - People will do things unexpected and make mistakes.

FPS - People will "get lucky" and there is always fun in that.

the list goes on.

Until your A.I. can call me an asshat and actually MEAN it, I'd rather play people.

On a side note, once your A.I. can call me an asshat and mean it I want him unplugged...

Re:People man! (0)

Anonymous Coward | more than 5 years ago | (#25623771)

Bidi bidi bidi. You are an asshat.

main problems..... (1, Interesting)

Anonymous Coward | more than 5 years ago | (#25622807)

the main problem is that AI is HARD. like NP HARD. and its difficult to program so it goes in last. what would change it is a predefined AI library which can elarn and be plugged into multiple games like the havok physics engine. something easy, can be shoved in last and can learn.

Re:main problems..... (1)

jbarlow (35149) | more than 5 years ago | (#25625583)

Mod parent up. This is exactly what's needed - especially for the smaller programming teams. More games period (and therefore, statistically, more good games) can be made if the team can focus on the game itself, and just plug in graphics, physics, AI, etc APIs.

Granted, it's sortta been done. Things like DMM, Euphoria... they're getting closer for sure, but wow, TFU was disappointing.

Re:main problems..... (2, Informative)

WuphonsReach (684551) | more than 5 years ago | (#25631511)

The secondary problem, as you alluded to, is that nearly every game uses a completely different system for representing the world. Different combat types, different terrain, different environments.

Which makes it extremely difficult to build upon previous work.

Fallout 3 (1, Funny)

Anonymous Coward | more than 5 years ago | (#25622961)

Can take a hint on this. As much as I find standing on a vehicle with an NPC perpetually running at me entertaining it just shows sloppy work.

Game buyer doesn't care (2, Insightful)

Saffaya (702234) | more than 5 years ago | (#25623059)

As someone who wanted to develop better AI for games, I'll say this : the state of AI didn't change because there is no customer need for it.

When AI becomes a selling feature, then it will be given more consideration by developpers AND allowed more resources by managers.
Which may be never, as it faces a tough adversary : the 'ooooh Shiny' whizz effect of graphics.

Re:Game buyer doesn't care (1, Insightful)

Anonymous Coward | more than 5 years ago | (#25623405)

It's mostly the fault of mainstream game producers There are not many [widely distributed] games where the player needs to use their brain to succeed....consider any games for which gold farm sweat shops are plausible.

Most games are not designed around awarding intelligence ( to the general chagrin of the /. community) but it is the reality...people who can think critically are the extreme minority, and game companies generally cater to the majority (lowest common denominator)

I could ramble on, but my target audience will surely be able to fill in the blanks for themselves

Re:Game buyer doesn't care (1)

cgenman (325138) | more than 5 years ago | (#25629881)

The intelligence-rewarding sections of the game all need to be tightly scripted. Otherwise you run the risk of having impossible situations. Similarly, the AI in most of these situations needs to be excessively predictable, so that the player can put together the component levers, bells, and whistles in their mind and see how the rue-goldberg machine would function.

Also, are we talking about the same types of games here? Most games are about simple escapism and the sorts of Human adrenaline rushes that one doesn't get sitting behind a desk all day. They're not about outthinking your opponent, so much as feeling a sense of control and accomplishment over the virtual world.

Pure escapist games do well because a lot of people want pure escapism. Someday, Puzzle Quest will be as big as Super Smash Brothers Brawl, but that is not today.

Re:Game buyer doesn't care (0)

Anonymous Coward | more than 5 years ago | (#25624323)

I'd like to see computer angry and pissed off (showed by some avatar or similar).
maybe that doesn't count as AI but it would make gaming more fun.

Re:Game buyer doesn't care (1)

p0tat03 (985078) | more than 5 years ago | (#25625225)

That's ridiculously easy, and is seen in a lot of FPS games. Guards will start pissed, then run away in fear when you clearly outclass them. Even Halo's AI had weaker units run away when confronted by you, and return in greater numbers.

Fear of Good AI (0, Redundant)

Tempest451 (791438) | more than 5 years ago | (#25623171)

Do we really want and AI that could hunt us down like dogs and kill us? We secretly like out AI to be dumb as a bag of rocks. Most games with "good" AI simply give the NPC a pre-programmed knowledge of the terrain, perfect accuracy, or the ability to soak up damage. If I were to make all things equal, I would give a game AI the same visual range, weapons, accuracy, and resistances as the player and then we'd see how smart it really is.

Most sophisticated AI so far (0)

Anonymous Coward | more than 5 years ago | (#25623315)

I have had extensive experience evaluating AI over the years. The best AI I've ever observed was the one where Carrie Underwood beat Bo Bice.

OpenTTD (1)

TheThiefMaster (992038) | more than 5 years ago | (#25623729)

Check out the OpenTTD NoAI [tt-forums.net] branch. The AI for the original Transport Tycoon reacted quite badly to having its cheating turned off, and in OpenTTD generally sucks, even to the point of bankrupting itself sometimes. The NoAI branch is an attempt to make AI that don't cheat, and are incredibly good. An AI can always be made slower or stupider.
There have even been some experiments into building an interconnected rail network [student.tue.nl] instead of sticking to point-to-point lines.

If there is anyone here who thinks they can program a good AI, I recommend you get involved.

Developers need predictable behaviour (3, Insightful)

setien (559766) | more than 5 years ago | (#25623757)

As a game developer myself, I can tell you one of the reasons why game developers often use finite state machines for AI instead of advanced neural networks that employ clever learning machine learning algorithms: It's orders of magnitude easier to analyze and understand (and thus debug and fix) how and why a FSM does what it does than a complicated neural network.

When you're making a game, you want results that are easy to predict and easy to schedule - if you decide to make advanced AI and train the NPC behaviors, it's hard to schedule and very hard to pinpoint and definitively fix a problem where one or more NPCs suddenly start acting extremely strange and un-human. And it's hard to fix if they become to clever.

It's one of those cases where simple models can get you most of the way, and it's more reliable and it's much cheaper to develop (in terms of processing time and implementation time).

Re:Developers need predictable behaviour (2, Informative)

skelterjohn (1389343) | more than 5 years ago | (#25625089)

Yeah... you don't want to use neural networks for game AI. Reinforcement Learning, on the other hand, is. At its heart the RL problem is the same as the sequential decision making problem. An agent acts in the world, receiving observations and numerical reward signals that it tries to maximize. The RL community is young (compared to the AI community as a whole) and is building up the theory and experience needed to approach these sorts of problems. All of my work focuses on agents learning to play video games (FPS and platformers, as more or less two separate threads). It's coming, and we'll be ready to help the AI soon...just leave a couple cycles free from all those fancy graphics so we can do some thinking in the background, ok?

Controller Problems (1)

javakah (932230) | more than 5 years ago | (#25624643)

I think part of the issue is that methods of interactions with games are still fairly clunky.

Keyboards with lots of different buttons can take some time to navigate. Moving appropriately with mice is also not super efficient. Even special gaming devices for the PC are not that great. Game controllers for the Wii, Xbox 360, and PS3 are not super great either.

With controls for interacting with games that are not all that great, designing great computer AI that can potentially react as well as or better than a person would react if they were really in that environment is unfair to a human player.

Ironically, I blame my bad grade in AI on (2, Funny)

bugeaterr (836984) | more than 5 years ago | (#25625027)

my teacher who, seriously, spoke with a lisp.

AI vs. Human (0)

Anonymous Coward | more than 5 years ago | (#25625063)

I usually play FPS games ...

When I play single player, I usually finish the game in normal settings. I struggle finish it with anything higher.

When I play the same game online, I am usually one of the top players.

I would prefer to have friendly AI improved first and then go for the enemy AI

Precedural dialogue (0)

Anonymous Coward | more than 5 years ago | (#25625197)

I believe the reason procedural dialogue is so behind all other forms of AI (also relating to turing tests) is that it is so unpopular among game designers. It has been neglected for deades with everyone preferring pre-scripted dialogue because they are too lazy to build a real language engine into their game.

Evil will always win because AI is dumb! (1)

jollyreaper (513215) | more than 5 years ago | (#25625203)

I pop in the latest RTS and I have that initial moment of "ZOMG! BEST GRAPHICS EVER!!!" but once I scrape my jaw off the floor I see that the units are as dumb as ever. Same pathfinding problems that were around in Warcraft 1 and we're how many years later? Ultimately it just means I'm playing the same game as before with prettier graphics. YAWN.

Better AI in Survival/Horror (1, Interesting)

Anonymous Coward | more than 5 years ago | (#25625691)

There are some good points about having AI enhance the fun for the player instead of greater difficulty or realism but there are still some genera of games that more difficulty and realism would be better. I use horror and survival games for an example. I've never found great interest in scripted monsters jumping out at you and then running mindlessly in your direction. I'd love to see a survival game were you have a sleek intelligent killing machine using opportunity and time to take you out, that backs off when in danger, and uses the environment to its advantage rather then yours. A game for example could be something like the first Alien movie. Lets see a horror/survival/puzzle game were you have to find a way off your ship in space, make crude probably ineffective weapons while being hunted down by a lone creature. The suspense alone could be huge if built with the right environment and right monster while just exploring. Why does a game need a large numbers of bad guys to mo down when with a great AI, level design, sound design and so on you should be able to have a good game with one?

Tell that to GNU Chess (0)

Anonymous Coward | more than 5 years ago | (#25626193)

Man, GNU Chess is one hard bastard.

Arsehole!!!

OMG Really? (0)

Anonymous Coward | more than 5 years ago | (#25626249)

"AI capabilities have not scaled with CPU speed,"

SHIT! And here we all thought AI was based on CPU speed. How wrong we were...

The AI in FEAR (2)

Cowmonaut (989226) | more than 5 years ago | (#25626305)

I wish more FPS games had an AI that behaved like it did in FEAR. That single player game had a lot of replay ability for me just because the AI responded with logical tactics.

They hear you make a noise, they'll move in to flank you while a few of them keep you distracted.

IF they notice the grenade they'll take off and run, they don't just run when there is a grenade nearby. A few times I've seen them just stand there still like they didn't notice the 'nade land near them.

If you go hand to hand, they wouldn't always shoot you because they might hit their team mate (or were all dead/dying and couldnt shoot me possibly).

The list really goes on. For a tactical shooter the AI was phenominal. Fast forward to COD4, probably the best modern warfare shooter out (haven't touched FarCry yet). Some spiffy tactics are there but if you throw a nade, they run, and they basically don't miss your head.

It almost seems a step back in a lot of that game.

Re:The AI in FEAR (1)

WuphonsReach (684551) | more than 5 years ago | (#25631551)

I greatly enjoy the AI in FEAR - I think they did an outstanding job.

Re:The AI in FEAR (0)

Anonymous Coward | more than 5 years ago | (#25632983)

Um, no. The list doesn't go on. You've hit all of the AIs capabilities. It's all rendered pointless when the NPCs screams out everything they are going to do.

I liked FEAR as a game, but the AI wasn't as hot shit as some people like to tout.

Why hasn't game AI been developed? (1)

Chris Mattern (191822) | more than 5 years ago | (#25628065)

Because game AI developments don't give you pretty pictures, bullet point features or big numbers to put in the ads and on the box.

Author is a Jackass (1)

g-san (93038) | more than 5 years ago | (#25630753)

I liked the article. Some good bits in there, lacking in detail or a good algorithm like any AI article.

But I am going to euphemistically call the author a Jackass for the following:

For example, inside a classroom there would be one specific set of social norms if it's full, a different set if it's empty, and wholly unrelated reactions when being shot at.

Thanks. I've got this idea in my head about always think about your characters animating with adverbs. I'm feeling mildly inspired. Then you give me a visual of a classroom being shot at? Find another example quickly, you insensitive clod.

AI as a driver for new kinds of games (0)

Anonymous Coward | more than 5 years ago | (#25681517)

From a cost-benefit standpoint, there's little incentive to make major improvements to the AI systems of the kinds of games that are popular now (first-person shooters, RTS's, sports games, etc.). The methods currently in use are well-understood and produce a reasonably fun playing experience. To build an AI in some more sophisticated paradigm (such as neural nets) would require a huge up-front investment of money and effort and would be a difficult proposition to sell to most game publishers when existing methods are good enough.

I think that game AI will blossom when AI itself makes leaps in other areas. I imagine a research team somewhere will come up with a fuzzy logic-based system that deals with some problem in science, industry, defense, or whatever, and someone on that team will think to him or herself, "hey, this has a fun side application!"

Check for New Comments
Slashdot Account

Need an Account?

Forgot your password?

Don't worry, we never post anything without your permission.

Submission Text Formatting Tips

We support a small subset of HTML, namely these tags:

  • b
  • i
  • p
  • br
  • a
  • ol
  • ul
  • li
  • dl
  • dt
  • dd
  • em
  • strong
  • tt
  • blockquote
  • div
  • quote
  • ecode

"ecode" can be used for code snippets, for example:

<ecode>    while(1) { do_something(); } </ecode>
Create a Slashdot Account

Loading...