Jump to content
  • Advertisement

Archived

This topic is now archived and is closed to further replies.

SpittingTrashcan

Skills as Movements, or Use the Processor

This topic is 5768 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Recommended Posts

My dissatisfaction with many computer RPGs arises from the paradigm they have inherited without question from tabletop roleplaying. Using pencil and paper, difficult statistical questions and issues of exact location, position, and velocity are abstracted and simplified to streamline gameplay. CRPGs accept without question the fudge factors introduced to accomodate puny human brains, not considering that the computer is capable of much more mathematical detail and precision. As a result, much more of the game engine is needlessly transparent to the player in a CRPG as opposed to say an FPS. Quake doesn''t tell me how it''s calculating the trajectory of my rocket... why does Neverwinter Nights have to tell me about the results of every virtual roll? Additionally, clumsy mechanics that could be eliminated or improved on persist due to inertia or just blindness to alternate approaches. To me the most blatant violation in this regard is the abstract skill system. This system, inherited from the radically different tabletop play environment, relies on a human arbitrator to listen to a desired action and determine what category it falls into. Then the effectiveness of the action is calculated using a numerical value applying to that entire category of actions - the "skill level". Unfortunately, the computer cannot categorize arbitrary actions as a human GM can, and so the range of actions must be limited to a finite (and usually quite small) set. Thus, in order to keep an abstraction designed for tabletop, the computer version discards features the tabletop system had without gaining anything. What if, instead, the computer more closely modeled the way skills are learned in real life, particularly in the case of physical skills which use the body? In real life, we observe another person carrying out a movement. We then attempt to copy the movement we observed. As we practice we get closer to the ideal movement and our effectiveness naturally grows in proportion. We can also invent and refine new movements for our own bodies which others can then learn, and so on. Translate this into game terms. A new movement is constructed as a skeletal animation. The physics model determines how the character''s moving body interacts with the world and displays the results. If the result is what was intended, the movement can be further refined by the player creating the movement, or by evolutionary algorithms which create multiple variations on the movement and test each for improvement on specific criteria. Eventually the results approach optimal, and a new "skill" is created. Other players whose characters watch the character using the skill receive the skeletal animation data, until they can attempt the movement themselves. The initial movement will be a degraded copy, but as the characters use the movement they will also adjust with evolutionary algorithms until they too approach optimal results. I have seen experiments conducted in which simple motile models whose shape and motion were controlled by variables were refined through evolutionary criteria based on ability to move forward... the results after 100 or so generations were artificial creatures capable of swimming quite nicely, and this was starting with totally random base creatures. So I know I''m not totally beyond the pale of technological or algorithmic possibility here, given that the characters are given a "head start" by receiving much of the movement information through observation. I have but two questions: Can it be done? And, should it? I believe it would truly add a whole new dimension to gameplay if it could be made. Imagine being able to construct your own custom moves, or learn any move you see. It boggles the mind. --------------------------------------------------- -SpittingTrashcan You can''t have "civilization" without "civil".

Share this post


Link to post
Share on other sites
Advertisement
My only concern is that it might not be that fun.

I can think of a few games where skills were "learned" by the player. The sierra Quest for Glory games worked that way. The more you threw things, the better at throwing you got. The more you fought, the better at fighting you got.

I''m sure there are more examples, but actually learning something is a tedious task that I would rather simply give to the computer to do. I don''t mind the "tiered" skill system, or developing skills by points, at least for now.

devinmaxwell

Share this post


Link to post
Share on other sites
Stats are actually something that players like.
Games are not supposed to be realistic at all. They are just meant to be fun to play. And it''s fun for most CRPG players to deal with stats, that''s all that counts.

Share this post


Link to post
Share on other sites
Your post seems to diverge a little so it''s hard for me to work out what you''re really saying. Are you saying that character-stats are a poor paradigm compared to the player himself or herself learning how to play? Or are you saying that there should be a system that better resembles the real ''learning'' progression of a skill better than a statistical one? Personally I don''t agree with either point...

[ MSVC Fixes | STL | SDL | Game AI | Sockets | C++ Faq Lite | Boost | Asking Questions | Organising code files | My stuff ]

Share this post


Link to post
Share on other sites
Ah, some feedback.

loftyideals: The fun factor is something of a concern for me too. I agree that practice has the potential to become tedious. The advantage of a point-based, abstract skill system are in getting rewards through gameplay: the more you play, the higher your experience and the more points you get to spend on skills, which you can then use immediately to full effect.

My original post was lacking in specific examples in order to convey the overall concepts in an abstract way rather than tying the discussion down to a specific instance. That was a mistake, as it makes the mechanic seem drier and more academic than it is, and doesn''t illustrate the advantages it holds. Therefore, consider the following example: the roundhouse kick.

A player wishes to design a new move - a roundhouse kick, where the character steps forward, twists his body around his forward leg, and brings the rear leg around to kick in a horizontal snap. Using a skeletal animation utility that comes with the game, he creates such a series of movements. The game physics engine then determines, based on variables such as the character''s body shape, weight, center of gravity, and muscular power, what the result of that series of movements will be. Let''s say the first version is unsuccessful: the character is unbalanced and will fall down. The player can further manually refine the movement sequence, or he can have the computer generate random variations in the movement and determine, automatically or by player choice, which is the most successful. Eventually, the movement is ready to go. The player now has a move completely unique to his character: the roundhouse kick. When he uses it on an object, damage is determined by the velocity and relative hardness of his foot and the target. Rather than have the character practice this move, which would be boring, the computer can run the refining process on its own as the player does other things in game; then when the character has gained enough "experience" the game unlocks the move for in-game use.

Other players see the roundhouse and want to learn it too. To do so, they watch the creator using the roundhouse. Once they have seen it enough, they can put it on their "to learn" queue. As their characters gain experience through play, new moves are refined and unlocked from the "to learn" queue.

Now take a moment to reflect on the sheer, staggering number of different moves that could be created and used this way. Every single fighter in the game could have their own personal style and repertoire of favored moves.

Moves could also be chained together in an intuitive and natural fashion. Each movement has a start and end state. When a single movement is executed, the move starts with the character moving from neutral stance into the start state, then executing the movement, then moving from the end state back to neutral stance. However, if multiple moves are enqueued or assembled into a single macromove, and if their start and end states match or are close, a significant increase in speed could be effected as the moves naturally flow from one into the other.

Of course I''m not suggesting that every single movement in the game must be learned this way; that would be ridiculous. Basic movements such as walking, running, jumping, picking up and grabbing, etcetera should be part of every character''s repertoire automatically. However, more flashy stuff like backflips, baton twirling, and fencing parries should have to be learned.

cahaan: I disagree heartily. I''m a player, of both pencil and paper RPGs and computer RPGs. On the computer, I hate stats, I hate having to see stats, I hate having to manipulate stats, and I hate the fact that CRPGs are to far too great an extent about manipulating stats rather than having fun. But your mileage may vary. Lord knows I''m hardly the average.

Kylotan: Sorry for the mixing of what probably should have been two separate posts. I am actually saying neither of those things. My two points are as follows.

1. CRPGs are held back by their adherence to a paper-and-pencil paradigm which doesn''t translate well to computers.
2. Example: skill systems, which as they are limit the number of actions and simplify their use; while using the system I propose you could have an unlimited number of actions, each modelled accurately in terms of results.

Does this clarify things any? I really think I''m onto something here with point 1, of which point 2 is just one of many examples.

---------------------------------------------------
-SpittingTrashcan

You can''t have "civilization" without "civil".

Share this post


Link to post
Share on other sites
If most players were like you, there wouldn't be visible stats anymore in RPG. But as you can see, it's NOT the case.
That's not because you don't like it that other players will not like it.
Everything is relative and you can disagree. But you will not be right neither will anyone. It depends of your preferences, but it seems from a market point of view that most players enjoy stats like I do.
BTW, I think it's not a good idea to be realistic in game design. Let the realistic stuff for the technical/3D side. How would you implement your "movements" stuff in a playable game ?

That's the implementation of a good idea that counts. And I think the exemple of "learning movement" concept you mentionned would better suit to a Creatures-like sim game rather than CRPGs. But if you come with a good implementation of that idea and a good final game, I will be the first one to try it

That's my 2c.



[edited by - Cahaan on October 1, 2002 8:10:50 PM]

Share this post


Link to post
Share on other sites
Don''t forget...one of the most prohibitive things to a system like you say are the input devices for computers. We can only input information through the keyboard and mouse, so how to you plan on creating these "skeletons" or "user-defined moves", etc. It seems like it would just be a 3d modelling app where it would take weeks for someone to create their own move.

And how do you represent your "mastery" of these moves? Obviously you cannot just start out with the ultimate fighting techniques simply by watching your friend for five minutes. So how do you represent it? A percent? That brings us back to what we had before - the stats showing openly.

Another thing. How do you plan on keeping track of all these skills you have learned, and how would you use them in combat? A list showing everything you have learned throughout your character''s career? That would be absurd, because even though you progress, you would still use the basic skills. Think of all the boxing moves there are. Straight jab, hook, cross, uppercut, body shot...and all the defensive counters, duck, roll, block high, block low, etc. How would you use all these in a combat situation?

Another thing- non-combat skills. How would you represent things like magic and thievery? Or even harder - passive skills like meditation or how about something like Musicianship? It would be very hard to represent those as skeletal animations.

Generally, I don''t really think your idea would work - it would still boil down to numbers. A player wouldn''t spend time examining the skeletal animation for his quick stab and try to improve his mechanics... no, he would just look at the damage done, and if it wasn''t satisfactory, go watch his friend bash on a monster and practice it to get more power.

Share this post


Link to post
Share on other sites
NeoSmyle,

I''ll try and address your points.

Yes, you would need something like a 3D modelling app to create these skeletal animations. Yes, they would take considerable effort to build. Experience with other games indicates to me that there exist a large subset of players who are willing to go to great lengths to create custom mods which change the game''s appearance and behavior - sometimes rivaling the original game developers in creativity and effort. Creating new moves won''t be for everybody, but I suspect many people would be willing to take the time to create their own special tricks, just for the cool factor.

Yes, these techniques could not be mastered instantly. They would have to be taken a piece at a time. A long, flashy macromove would consist of several component moves, each to be learned over time. They would be placed on a "learning queue" and learned in succession. Your progress toward mastery of these components would most likely be represented numerically inside the game engine , but there''s no need to give the player such detailed information. You could use a graphical indicator such as a filling bar, or just inform him when he has gained a new move.

Skills could be tracked via a list - perhaps a little animation demonstrates how each works. They could then be grouped by the player into sets, where each set maps a move to a keyboard button. The sets themselves could then be pulled up with other hotkeys. Say Q, W, E, and R are the four "use move" buttons, and 1-8 pull up different move sets. That''s 32 skills accessible with two keypresses each, and with the custom grouping of related skills you wouldn''t need to switch between sets all that often. Another option would be to have a fighting-game style system where key combos as well as single taps are mapped to different moves. Think "Oni" - how many moves were in that game? Now imagine multiplying that by 8 button map sets. Sound intriguing?

My idea isn''t meant to handle more abstract skills - I''m still working on that problem. It''s just for physical movements, for now.

With a good physics engine, raw damage isn''t the only factor in moves. Ever play or see Bushido Blade? In that game, damage was largely irrelevant as almost all attacks were lethal. Instead, the dynamics of movement and counterswing became critical. A player chose character and weapon, and different characters handled different weapons in different ways: the small, lightly built character was lethal with the light rapier, but clumsy with the hammer. There were no all-stopping blocks, only parries which would knock aside attacks. Shifting stances gave access to different move sets. Now consider what would happen if every character in the game could pick, choose, and create their own move set. To consider only numerical damage would be to ignore the wide range of other factors in combat, and is another flaw in the simplified paradigm of P&P RPGs (where combat is abstracted into chance to hit and numerical damage).

Perhaps it would clarify things if I called this new type of game I am attempting to describe by a new name. I am not trying to make another RPG such as you are used to. I am trying to make an RES: Reality Enhanced Simulator. Reality, for the basis in realism; Enhanced, for the fudge factor introduced to make things more fun than reality; Simulator, for the physics-based model rather than a more abstract rules-based "game".

So, who wants to play an RES?

---------------------------------------------------
-SpittingTrashcan

You can''t have "civilization" without "civil".

Share this post


Link to post
Share on other sites
I see some of your points. I was actually thinking of some of your answers would be as I wrote my first post, but I was just playing devil''s advocate and wanted to make sure you saw both sides of it.

Now, are we talking about an MMORPG here? It makes more sense than a single player RPG. The concept also seems to have a steep learning curve; it just has a lot of components, and it might be hard to grasp as a beginner. I''d say to lessen it, I wouldn''t put so much emphasis on the new move creation. It would definitely be an advanced topic, so I''d keep it away from the beginners. OASN (on a side note), I think that players should be able to name moves they create, so when other players learn the custom moves, it will still be known by its creator, giving them some fame and extra incentive.

I think it would be a good change though, as long as it was kept mainly towards the fighting aspect of the game, and less from the overall gameplay.

Share this post


Link to post
Share on other sites
Neosmyle,

- By default I am talking about MMORPGs... I favor them over single-player RPGs because I like interacting with others.

- Making custom moves won''t be for newbies, and won''t be necessary to have fun playing the game. It''ll be more like modding or creating custom skins; few will be able to do it but all will benefit from their efforts. Certainly creators should be given credit for the moves they make. People who make new moves in real life martial arts, sports, or gymnastics are legendary...

- The applications to combat are immediate and obvious... but have you considered the apps in other athletics? This system could also be used to make new slam dunks, better passes, cool skateboard stunts, gymnastic vaults and tumbles... once you start thinking about it, the applications are wide in scope.

By the way, excessive focus on combat is just one more of the many problems which arise from not thinking critically about the P&P paradigm. Not every RPG need be all about cutting a bloody swath through the minions of evil... if it''s all right with you, I''d rather take the role of a rock star, painter, or b-ball player. If nothing like these options are open, it''s not an RPG; it''s a combat game by any other name.

---------------------------------------------------
-SpittingTrashcan

You can''t have "civilization" without "civil".

Share this post


Link to post
Share on other sites

  • Advertisement
×

Important Information

By using GameDev.net, you agree to our community Guidelines, Terms of Use, and Privacy Policy.

We are the game development community.

Whether you are an indie, hobbyist, AAA developer, or just trying to learn, GameDev.net is the place for you to learn, share, and connect with the games industry. Learn more About Us or sign up!

Sign me up!