Slashdot Mirror
Should Computer Games Adapt To the Way You Play?
jtogel writes "Many games use 'rubberbanding' to adapt to your skill level, making the game harder if you're a better player and easier if you're not. Just think of Mario Kart and the obvious ways it punishes you for driving too well by giving the people who are hopelessly behind you super-weapons to smack you with. It's also very common to just increase the skill of the NPCs as you get better — see Oblivion. In my research group, we are working on slightly more sophisticated ways to adapt the game to you, including generating new level elements (PDF) based on your playing style (PDF). Now, the question becomes: is this a good thing at all? Some people would claim that adapting the game to you just rewards mediocrity (i.e. you don't get rewarded for playing well). Others would say that it restricts the freedom of expression for the game designer. But still, game players have very different skill levels and skill sets when they come to a game, and we would like to cater to them all. And if you don't see playing skill as one-dimensional, maybe it's possible to do meaningful adaptation. What sort of game adaptation would you like to see?"
I'd like to see it configurable. Check box that allows adaptation, with sub-items that define what type of adaptation will occur.
Don't blame me, I voted for Kodos
Rubber-banding is no different than a golf handicap, tennis ladder, or beginner/expert/pro leagues in most sports. It's simply not fun to play too far out of your skill range. The talk about "rewarding mediocrity" is misplaced in an activity that exists only for fun - it should be rewarding for everybody, otherwise players would (and should) quit.
What if the game taught you to be a better player? For example, it could slant the gameplay to teach you one strategy, then once you'd mastered that, move on to teach you a different one. If you do well enough, it starts to require combined strategies, etc.
PHEM - party like it's 1997-2003!
Most games already have a option to choose how hard or easy you want your game. This works better than autoleveling, because If I set the game to be hard, and I die too much, maybe thats exactly what I want, and If set game too easy and I kill everything, maybe thats what I want.
Good games, like World of Goo, have options to skip night imposible levels (since is a puzzle game, you could be stoped totally to experience the whole level). This is like these ols space games with "megabombs" that clear the screen. But that "megabomb" is limited.
Challenge is good wen you want challenge, havin games that kill challenge would be fatal. And this one of the reasons Oblivium was a bad game, and Morrowind was a much better game.
postdata:
Also, dificulty is not that all important. Fun is important. Games sould be fun. The dificulty is not the reason. But since we are talking here about dificulty, I have talked about it, and what it means.
-Woof woof woof!
If I just keep sending Kasumi to the pool to watch her rock back and forth on an inflatable dolphin in the pool, you might as well remove her swimming suit completely.
I don't know what you'd call that, I would call it a reward for not exhausting my character in volleyball matches under extremely hot temperatures.
for a game to achieve a given level of revenue at a given price then you can compute the number of items you need to sell. if you make it too hard, your demographic won't support it. if you make it too easy then you bore the hard core and also may lose the demographic size you need.
the question is does medium hard work?
if not then you need to have variable difficulty to capture the area under the demand curve.
Also if lets freinds and guests compete on the turf of an expert. the expert may enjoy having more freinds than the person at his level.
Configurable is nice but i'd probably not be an expert enough to know what i needed until I had played it for a while and gotten frustrated.
Some drink at the fountain of knowledge. Others just gargle.
One problem, potentially, if you 'adapt to players skill level' *too* well, is that as they get better (or as their character gets more powerful in an RPG type system), they might feel like they never get to enjoy the increase in either their skill, or power. It can feel like treading water, if as you get better, the game gets so much harder that you never get any feeling of accomplishment, no sense that you are any better or stronger than you started out, even though you *know* you've gotten better, or have more powerful abilities.
However, at some point, you do want more challenge. The trick will be, adapting to the players, while still giving them some opportunity to experience their increase in skill or strength.
This could be applied to almost any game genre, btw. I mean, consider an FPS. If you've gotten better at managing your economy, strategizing attack tactics, etc, but the computer remains in lockstep with your real skill increase as a player, then it can be very frustrating. At some point, you want the satisfaction of just slaughtering the AI player that used to beat you on the same 'skill level', because your skill has actually increased.
Basically WoW has it right. Oblivion was annoying as as soon as i level those "bandits" suddenly had very very good gear. I don't like that it's no fun, sometimes it is nice to walk to an area you have been before with your gear and butcher the low level stuff for fun.
Bestheda also fucked up Fallout 3 with this, you can pretty much complete the game in under 3hours (iirc) with hardly any leveling as the monsters are pretty much all scaled to the player.
I do like rubber-banding as long as it is managed (eg a lvl 4 monster, depending on my skill, can have the stats of say a lvl 5 monster but never any higher) this allows for a small degree of rubber-banding so good players will have a harder time but can still return to low level places.
Players enjoy certain aspects of particular genres:
1) In an RTS like Battle for Middle Earth, the draw is general defending large armies with large armies, the thrill of out-strategizing the enemy (AI), and the final devastating blow to your opponent's base. If you're playing well, and dominating the enemy, then make the game last a little longer: send out a large "backup" force from the enemy that really makes your main force struggle...but once your main force is weakened (or not), you're given time to rebuild. You may be prepared for these reinforcements to hit you and split your main force to flank them when they do arrive, etc.
2) In an FPS like Quake or Doom, you might reward run'n'gun playstyles with simply more enemies to slaughter, or be slaughtered by. More strategic FPS players may actually get the same reward, or perhaps have enemies begin to spawn behind them to make them start watching their backs, heightening the tension that comes from playing an FPS slowly.
3) World of Warcraft players might get the Amazing Sword of Brilliance if they actually attack two mobs at once instead of ganging up on one.
It has a lot to do with what people decide is fun in a game, and one reward system won't work for each genre -- but it may work for the majority of players in that genre. Find what the players are looking for in that game, and give them more.
Cartman! Stop saying the F word!
I think there's room for choosing a difficulty level and having the game adapt as well. Didn't RE5 do that? You chose how hard you wanted it to be, but within that the game also decreased enemy health if you died over and over, and increased it if you survived fights without dying. So it was self adapting but within constraints you could choose yourself.
There's also a clear difference between games in which you compete against other people which try to provide an enjoyable experience, and games in which you are trying to win by having more skill than the other players, and single player games that are intended to be enjoyable and what people enjoy varies from person to person.
The enemies did not "Increase in skill", as if they matured and became better fighters, they simply leveled up as you did. :/
That's not adaptive AI
There are 2 things that need work in games-- AI and facial animations. It's been 10 years since UT99 and in UT3 the computer basically rolls a dice that determines if it's going to kill you. If it's going to kill you, it usually kills you on the first shot. Which never happens in real life. Something as simple as this, which would be so easy to get around, makes the game feel so cheap. Yes, I play with people online, but when there's only 3 and we need a 4th for iCTF, having a bot ruins the fun.
Facial animations-- see Half Life 2, in my opinion. Even though the character animations themselves are a little stiff, the lipsyncing is top notch, and the Gman can display emotions such as confusion, malice, irritation, etc. Combined these all work together for a great suspension of disbelief.
I was intrigued by the concept of adaptable games until I played Oblivion. Granted, Oblivion made the worst possible decisions when it came to adapting Mobs to your level: it had an uneven leveling "curve" to the point where gaining a level could make previously easy monsters into a nightmare. It used obscure leveling mechanisms where you could gimp your character to an unplayable point if you didn't happen to pick the right class or jump often enough between leveling.
Since then, I don't care about adaptive leveling, because it is a much harder problem than it appears to be on the surface. Part of the fun for me is to go from getting stomped by the computer to stomping the computer, just because I got better at the game. Sometimes I want the challenge, but then I select it, not the game. Judging from the amount of Starcraft games that are labeled "7v1 stomp the comp", I'm pretty sure I'm not alone in this.
Adaptive difficulty should really come only in two flavors: select an overall game difficulty, so that you know what to expect; or enter some dungeon or bonus level/path that you know is much harder than what you've done so far. Don't force me into a harder game just because I've been doing so well so far. It could have been just a lucky streak, in which case I'll get really frustrated with the sudden ramp-up in difficulty.
Those who can, do. Those who can't, sue.
When we developed Tracers back in the '80s we tuned the reward system so that the game would just run at a higher speed (voltage, in the circuit-board language of the game)... every time you won a level, the voltage would ramp up, when you lost a life it would ramp down. Most people found themselves in a cycle where the game would get harder until they started losing lives, and then it slowed down again until they started winning levels again.
The higher the voltage, the more points you got for blocking off and killing an opponent... but we found that the best players quit paying attention to the score. The challenge in the game was pushing the voltage higher and higher. That number was the thing to beat.
I don't like games that try and hide the mechanics of the process from people, but when it's exposed like this it can be extremely effective.
Although these advanced systems can be done in single player, stand alone experiences, I predict we will see a lot of progress made in the MMO space where it is easier introduce dynamic content. One thing sorely missing from MMOs is custom built challenges. The game has access to all of that information on the character and how to play...why not start using it to change the things prsented to them?
- Using general terms for an example: If you enter an instance with a Warrior, a Thief, Wizard, and a Cleric but you kill the dragon and get some Ranger bow everyone goes "BOOO!". The game knows what classes came in so instead of just tossing out static loot from a static table, start considering who walked in and what improvements they need. Instead of forcing players to grind content for drops they know a monster has, they should come back for a chance on loot they know will be useful to someone.
- Since the game knows what classes came in, why not start seeding the instance with challenges configured for them? Each of the classes in the example are strong and weak to attacks and monsters, like for instance this group is a little weak on "ranged attacks" but stronger on defense. This group would avoid any static content they know would have a preponderance of stuff that flies or run around them. How about have them go into an instance that configures it to have less fliers, less stand back but features stuff that hits a little harder than normal?
- If the group is working well together and is stomping everything, why not up the difficulty a little till they aren't stomping everything? If the group isn't doing well, why not ease the difficulty so they aren't wiping every turn?
The basic idea is that the game should be smart enough to see at least the game/character data and evaluate what should be easy and hard for them to beat. This isn't so much "hand holding" but crafting a more interesting experience. If you swap the Thief for a Ranger and go into the same area you get a different mix of monsters and a guarantee that someone is going be rewarded. If you come in with a weak group you get a challenging experience. If you come in with a strong, expert group you get a very different but still challenging experience. The game designers should want you get through the quest handed to the players, to experience the story of the content, but still provide enough of challenge to feel accomplishment. Right now this is done with carefully crafted static content that involves a bit of statistical analysis that can be easily memorized or grow out of.
Heh, how about game difficulty set via Breathalyzer!
This issue is a bit more complicated than you think.
Except that's yet another case of talking out the arse without knowing what the real problem is.
The problem is: in many of those games with rubberbanding, there is already another mechanic for those tiers you describe. And the rubberbanding is nullifying the other mechanic. _That_ is what some of us complain about.
E.g., in the Gran Turismo series (and many similar games), the focus isn't on just jumping into a random race and having your 15 minutes of fun. You have to earn the car and the upgrades to qualify for the next league, and then even more upgrades to win in it. There is already a mechanic to simulate those leagues, and to justify why you should spend several days grinding your way through them. (Read: why you should play each of the few race tracks more than once.) Throwing in rubberbanding is nullifying all that, and turning it right back into a kiddie kart game. Suddenly it's hard not to notice that the whole tuning and upgrading aspect is bogus, since the opponents really are just tied to your car with rubberbands. What's the point in grinding to upgrade your engine HP by 50% when, effectively, every single opponent just got the same upgrade?
E.g., in Oblivion and generally an RPG, there's already a mechanic for simulating those leagues and tiers. It's called xp and levels. (Or skills, if it's skill-based a la Oblivion.) If your skill is too low to beat this opponent, you're supposed to go raise it somewhere else, and if it's too low, well, then just go fight something higher level instead. Do you understand that crucial aspect? There is no need to simulate those leagues and tiers in a game which already has another mechanic for just that. And adding some form of rubber-banding just makes the other mechanic a pointless waste of time. Why bother grinding your character to level 50, when effectively it gave you no advantage at all?
And it doesn't help that all too often it's done _badly_ too. E.g., since we're talking about Oblivion, the end opponent is actually a lot easier to beat if you somehow manage to get there as a level 3 character, than if you did all the quests and have a level 30 character. Effectively, you're better off if you skip 90% of the game and just do the absolute minimum that gets you through the short main quest arc. It's not that all that grinding and exploring and getting equipment doesn't give any advantage, it's that it actually becomes a disadvantage.
A polar bear is a cartesian bear after a coordinate transform.
I seem to remember in the promotional materials for the NES game Zanac (by FCI) that the game was supposed to get dynamically harder the better you played. When I was playing, I specifically remember this being the case, and that I enjoyed the game more as a result. I used to be able to play straight through to the 10th (out of 13) levels without dying once, and then I would die multiple times in a row. As if sensing my desparation the game would scale back the number of baddies it was throwing at me, and then I could regain my footing, collect some powerups and move on. Then the game would throw more and more at me until I got to the unholy nightmare 13th level.
Time to go dust this game off on the Wii...
Having to struggle against myself does NOT sound relaxing to me.
www.Migrainesoft.com - Computer giving you a headache? We can fix that!
I'd specify that "rewarding mediocrity" is a misleading term in a single-player game.
Single-player games are not as single-player as you might think, with online high score boards and achievements and the like.
I just had a terrifying thought of playing a game and seeing an animated paperclip pop up and say "It looks like you're trying to destroy the mothership, would you like some tips?"
Support Right To Repair Legislation.
There is a big difference between man-vs-machine and multiplayer games. In multiplayer games, there is certainly the need of a certain handicap to make the game fun to everybody. In man-vs-machine, I'd say that yes, the game can get more difficult, but also that the rewards must increase. So if the enemies get stronger, you have to at least have the option (if you are skilled enough) of getting better weapons or whatever. Also if the measure of the game is the score, for example, then the score should reflect that you have walked a more difficult route.
Rome taught me patience and assiduous application to detail. Virtues which temper the boldness of great, general views.
I thought I was the only person who had the issue of wanting to explore a new area, but getting slaughtered by the first thing [popping out of the ground|falling from the sky|warping in out of nowhere] was a major irritant.
I'd love to see an MMO that allows one to explore, with some logical limits. Like a real person, you can't just run all over the world in 5 minutes. You actually need to build up your endurance (hooray for stat gaining without a level, preferably -- someone who explores a lot and carries lots of stuff would likely have more endurance than a flabby something or other that's just wandering around the outskirts of town), buy equipment for exploring some areas (mountain climbing means you need pitons, rope, carabiners, etc; safari exploration means you might need some type of insect repellent, a machete, and a prayer to protect you from [insert random creature here]) and make money by bringing things back from your explorations to sell.
Of course, this kind of an idea would be hard to apply adaptive skill levels to, honestly.
One of these days, I am going to flip out. When I flip out, I'll be back in five minutes.
"Some people would claim that adapting the game to you just rewards mediocrity (i.e. you don't get rewarded for playing well). Others would say that it restricts the freedom of expression for the game designer."
What, their freedom to guess wrong and alienate a large chunk of their playerbase? Player skill is going to be on a bell curve, and the best you can do without some dynamic adjustment is to hope to hell you've nailed the difficulty perfectly at the top of the curve; that way you're the least wrong for the fewest number of players... but even then, you're still going to be unplayably wrong for 10% and irritating to another 20%. And this will only reward skill for that narrow slice of players for which the game was initially slightly too hard (and then becomes pefect as the player improves).
The flaw in rubberbanding is only that it still can't read your mind. The developer's idea of "normal" may actually still be too easy or too hard, and then the game guarantees that it stays too easy or too hard throughout, no matter what you the player do. Really what we need is a hybrid between the old "easy/normal/hard" choice and dynamic adjustment. That puts enough wiggle room back in that the developer can be wrong yet the player can still fix it and have fun. And the holy grail here is to have it require minimal interaction - if you implement this right, it's correct by default for the largest reasonably attainable number of players, and for the rest it's correctable through the simple and well-understood easy/normal/hard mode choice.
My first thought on reading the summary was that it sounded a lot like the local chess and go competitions when I was in high school (a few decades ago, before computer games were common). I was one of the top players. I didn't much get that way by reading a lot of chess strategy books or by beating a lot of novices. I did it by consciously deciding that I liked losing better. That is, I challenged players who were better than I was. They usually learned to try each trick on me just once, because the second time I'd have worked out a reply. Also, from then on, they had to look out for the same trick from me.
Nowadays, I don't play many computer games. But if I decide to take it up, it'll be because of access to slowly-increasing challenges. If a game doesn't behave as described here, I'll get bored with it fast and go looking for something that's more interesting.
Actually, part of the reasons for getting out of games is that I realized that software development is a kind of game that you can get paid well for. The basic setup is: When you get the recalcitrant little beastie to do what you want, you get points (and possibly a raise for the next project). When the designers of the system (OS, runtime libs, compilers, data designers, whatever) trick you and the machine interprets your code differently than you expected, the people responsible for the system code get points (and possibly a good position building the next release of the system ;-). A good programmer is one who can win at this game against the system designers.
So as a programmer, you're constantly challenged by the new challenges that are hiding out in the latest releases of the systems that you're programming for. You really are playing against some of the brightest human opponents on the planet. It's a much more interesting and challenging computer game than anything actually advertised as a game.
I've described this theory to a number of bosses in the past. One of them chuckled, and explained that this was probably why I hadn't ever "graduated" into management. He'd seen my code, and it was too clear and well-documented to ever be a good player on the "system" team in the game. The other programmers wouldn't face the challenges they expected from my code, so it was obvious that I wouldn't be welcome on the other team. So I chuckled to, and told him that I was happy playing for my current team. I got to build things that users actually use, which was a nice bennie. Sometimes they've even paid me for copies of my code, while people only pay for "systems" code because they have to for the machine to be usable. We both thought it was all pretty funny. But maybe this was partly because we were both paid pretty well to play.
For some reason many "system" programmers don't seem to appreciate this characterization of the software industry ...
Those who do study history are doomed to stand helplessly by while everyone else repeats it.
Let me explain: what I want is a game where not only the world doesn't adapt to how I play but, also, that it's not even designed to how the developers *think* I'm gonna play. Things like Final Fantasy VII, for instance, where even the strongest bosses in the first few areas would be killed in a single hit from the random encounters you get at the final dungeon. I want a game where I feel I was just thrown in a different world, that I'm merely a participant in something bigger, rather than The One True Hero around whom the whole world is built.
STALKER did this, to a degree, where in the beginning with your trusty pistol and simple jacket you're forced to run from mere bandits, while in the end-game you can hunt military soldiers for fun and profit with your customized AK-74 and bulletproof suit. It did have an "NPC difficulty curve" (mostly due to quests leading you to more dangerous areas as the game progressed), but it was flatter than most and that worked to the game's favor, IMHO.
No problem is insoluble in all conceivable circumstances.
Call me old-fashioned but I've always believed that one of the pre-requisites of calling something a "game" is that it should challenge you. Give you something to learn and get better at. There was no adaptive difficulty on Mario, Zelda, or Metroid. If you wanted to advance in the game (or even beat it), your only choice was to practice, explore, learn from your mistakes, and hopefully get better. A game that automatically makes itself easier when you do poorer isn't a game, it's just a time-waster. In the same class as the click-on-the-pretty-pictures web games and every board game that boils down to sheer chance.