Archived

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

rmsgrey

Learning by not doing

Recommended Posts

rmsgrey    153
Just a belated afterthought to the threads of several months ago on learning through failure, etc. How about if, whenever you improve in a skill, you get some sort of serendipity points for all your unrelated skills - and when you get enough serendipity points for a skill, it spontaneously improves. The idea is that, in real life, doing something apparently completely unrelated can suddenly give you a new insight into an area where you''d hit a block.

Share this post


Link to post
Share on other sites
Guest Anonymous Poster   
Guest Anonymous Poster
someone ought to reply, I think it makes sense, though it may be hard to figure out what skills relate, truly random increases might look weird to the player

Share this post


Link to post
Share on other sites
rmsgrey    153
I''m still not 100% sure what I''m aiming for... sort of nibbling round the edges of an idea. Part of it is to encourage diversity in a character rather than extreme specialisation - rather than having a master of all swords that wouldn''t know how to get a sound out of a recorder if you stuck the pointy end in his mouth, have a slightly more rounded character that has pretty much mastered sword-fighting, but can also play a musical instrument or writes poetry or knows how to behave in court (royal or legal depending on setting). The sort of individuality that a good p/p GM encourages.

I intend to pretty much assume a range of 0-100 for skills any time I throw numbers around.

Some ideas:

Have some sort of plateau effect, whether static or dynamic - once a character reaches a certain point in a skill, further training just doesn''t help much any more. There''s some concept that he just doesn''t get which holds him back. This could happen at fixed levels (25,50,75,85,90...) or every time the character misses 3 chances to improve the skill in a row. Once you hit that plateau, the only way to improve further (at the normal rate?) is by doing something unrelated until the key idea suddenly drops into place. For example, Marcus the Legionary can perform any individual sword move or sequence of strokes perfectly by the book, but he has difficulty moving from one sequence to another, changing mid-stream or improvising in an actual fight - as long as he fights in formation with the legions he''s fine, but he''s tying to go freelance. When he takes up Pottery (say) one day he''s struck by a chance resemblance between the shape of his latest failed pot and a sequence of strokes. Playing with the clay a little, his pots don''t get much better, but he finds forms that, in his mind, resemble the various sword strokes and sequences. Seeing the clay forms flowing into one another under his hands, he suddenly intuitively grasps the idea of improvising in combat in a way that eluded him previously. Next time he picks up a sword, his new insight enables him to surpass his previous performance and learn freely once more (until his next block...)

I think serendipity point gain should probably be proportional to: gain in new skill; 100-new skill level; and level in old skill. There should probably also be something in there to reduce the effect of studying something closely related to an existing skill at higher level.

Under this iteration of my idea, accumulated serendipity points just give you a percentage chance of unblocking a skill, and maybe a one point increase. I like the idea of reaching the stage where the only way to improve in a skill is by generally broadening your mind. I''m not sure whether hitting a block should stop all progress in a skill, or just slow it to a crawl (say 100 times slower)

Share this post


Link to post
Share on other sites
Hamdoon    122
That''s a pretty cool idea, I like the goal of encouraging non-specialized character development.

Another way to go about it might be to make a secondary skill affect the way the character uses their primary skill (rather than acting as a "key" to further advancement ). For example, sufficient skill in a musical instrument ability might give a swordsman a certain kind of special attack, like a "Rhythm strike" or something like that. You could either build those skill combo bonuses in explicitly for skills that the designer identifies as working well together, or come up with some kind of a system for generating the skill combos automatically from the attributes of the skills themselves.

Share this post


Link to post
Share on other sites
walkingcarcass    116
The key to making something like that work in the general case, is to abstract objects in the style of some puzzle games (classic example: the scythe in Grim fandango, it has about a hundred sensible, realistic uses) if this kind of system is skillfully designed, diverse exploration of the game world and it''s objects is easy and as narrowly or broadly designed as the player makes it.

********


A Problem Worthy of Attack
Proves It''s Worth by Fighting Back

Share this post


Link to post
Share on other sites
rmsgrey    153
Hamdoon - interesting idea, but two immediate problems spring to mind: firstly, doing the skill combos by hand squares the complexity of the design work on the skill system, while automation requires a lot of work to come up with appropriate synergistic features for each skill (a list of primary and secondary or enhanced and useful subskills might do it). Secondly, and more seriously, barring superhumanly good game balancing, there will be at least one ''broken'' skill combination that is significantly more powerful than most - again discouraging diversity.

walkingcarcass - I''m not convinced abstract objects would provide any kind of answer to skill-maxing. OK, you''d avoid the problem of the PUSD (Platinum Uber Sword of Doom), but you''d still have the problem that everyone ends up as a standard fighter or mage or fighter/mage with damage dealing skills maxed out, while any skill that doesn''t directly help progress in the game is left at or near default level. I agree something should be done about the PUSD, but it probably belongs in a different thread...

Share this post


Link to post
Share on other sites
Your example of Marcus the Legionary is a neat illustration of the idea, but I think it would be impractical to develop an engine capable of modelling epiphanies. I''d rather see a system in which the character changes with skills, rather than simply acquiring them.

For instance, a man trained all his life with hammer and tongs might be a fine blacksmith. He''ll be industrius, physically powerful and able to do whatever he wants with metal. However, he''ll be a simple man with little imagination and almost no knowledge of the world outside his smithy. If this man takes up a hobby, like botany for instance, that has him walking through forests and fields, and examining things that grow there, then he''ll gain a certain insight into nature, and an appreciation for the symmetry and balance that can be found there. When he returns to his forge, he will be a different man, and his work will change to reflect this transformation.

So perhaps it would be best to have some kind of aptitude system. Someone with a rich artistic background will be more apt to do imaginative, creative things, but will perhaps lack the ability to solve problems through sheer perseverence. Someone trained in combat will be strong and brave, but might lack the diplomatic skills that would make him a great leader.

So the optimal character wouldn''t be the invincible warrior or the undetectable thief or the powerful mage, but would be a true renaissance man, able to apply his understanding of many disciplines in any given scenario. Technical skill would be there, and physical capabilities would of course be essential, but true greatness could never be achieved through single-track training.

The fighter who butchers hordes of goblins would never be as terrific as the fighter who butchers goblins, trains with other people, practices archery, plays chess, reads philosophy and plays the bassoon while holding a good job as a cobbler.

Of course, how you''d implement this is beyond me. Maybe some kind of sum-of-parts system for each skill.

Share this post


Link to post
Share on other sites
3dcgi    127
Dungeon Siege doesn''t do anything as sophisticated as your description, but it takes a step in that direction. For example, using a sword increases strength, but it also slightly increases intelligence. And the opposite also applies. Using magic increases intelligence, but also slightly increases strength. The helps keep the characters from becoming too much of a specialist.


--
http://www.3dcgi.com/

Share this post


Link to post
Share on other sites
Shrap    122
Have you guys played UO? That game required combinations of skills for you to be well off. With skill combinations it gives the players many more doors in which they can develop their characters. Sure there will be a few combinations that are superior to others but that is what patches are for, that tone down these combinations so that the game is balanced. I feel that for a player to be a good swordsman, it shouldn''t just be based on their swords skill, base it on how well they know their opponents weaknesses (anatomy skill), where to hit the player(tactics), and basic swords skill. After playing many MMOG''s, I still feel that UO''s skill system is way more dynamic than all others, and gives the player the most freedom. That is what made the game interesting, that every aspect of your character was molded to your liking.

Share this post


Link to post
Share on other sites
rmsgrey    153
quote:
Original post by Iron Chef Carnage
Your example of Marcus the Legionary is a neat illustration of the idea, but I think it would be impractical to develop an engine capable of modelling epiphanies. I''d rather see a system in which the character changes with skills, rather than simply acquiring them.


The example with Marcus the Legionary goes into rather more detail than the game engine would - in playing the game, the player (Mark) trains in sword fighting (and possibly regimented fighting depending on details of skill set - probably not an option). After a while, Mark''s character stops improving. In order to unblock the character, Mark knows he has to try working on an unrelated skill for a while - he picks pottery, and after a couple of improvements in pottery skill, Mark goes back to sword fighting and can advance some more - until next time his character stops improving.
quote:

The fighter who butchers hordes of goblins would never be as terrific as the fighter who butchers goblins, trains with other people, practices archery, plays chess, reads philosophy and plays the bassoon while holding a good job as a cobbler.

Of course, how you''d implement this is beyond me. Maybe some kind of sum-of-parts system for each skill.

My proposed system has the advantage that the renaissance man who is always ready to turn his hand to something new will tend to develop more overall than the highly focused specialist (with only a handful of side skills) - at the expense of not really mastering any one skill completely. The obvious disadvantage is that it doesn''t ever make the epiphany explicit - you know your character has had some sort of insight, but the details aren''t reflected in the game... to implement that, you''d probably need some sort of categorising of skills to help reduce the quadratic complexity - and for game balance, you probably want to keep the differences reasonably cosmetic - the nature loving blacksmith who makes works of art and the science loving blacksmith who develops better steel both improve the value of their work by about the same amount, but the artist''s work is more ornamented and the scientist''s more durable (which raises some of those game-balance issues again... bother)

Share this post


Link to post
Share on other sites
Dauntless    314
There''s a saying in martil arts circles..."if you master one thing, you master everything". I think the saying means that to become a true master, you have to focus your attention on even the most minute detail in order to understand and make it a part of your skill. And this attention to detail and absolute focus is what eventually helps your other skills as well. A corollary saying is, "how can you do something diffucult if you can''t master the details?". SO this actually would corroborate what you suggest...to increase other skills slightly as your main skill advances.

In physiological terms there is a not well understood effect that if you train one muscle group very hard, ALL muscles will benefit to a small degree. The larger the muscle you train, the greater the bonus side effect. This is why Sprint runners also tend to be pretty well built, even though many of them don''t do any upper body exercises. So there''s also good reason to do this even for attributes as well as skills

There is something to be said for being a "jack of all trades, master of none" as well. Without having been exposed to many different perspectives, skills and knowledge, you have a one dimensional way of thinking. Perhaps you could have a "creativity" attribute. This attribute gets increased with the quantity of skills that you possess (as well as the level of the skill so as not to have a character load up on a million level 1 skills). This creativity attribute is a secondary attribute in many skills (though not all) in which imagination is a strong factor in the skill (many arts, sciences, programming,performance based skills, etc).

Share this post


Link to post
Share on other sites
Shylock    122
I like the idea dauntless, and the creativity skill could also be used as a basis for speed at increasing skills. ie, a creative person would develop their own sword menouvres, or come up with their own ways to effectivly crush herbs for a potion.

Share this post


Link to post
Share on other sites
rmsgrey    153
The creativity sounds like a good idea. Not sure how it combines with my "mental blocks" idea though...

Anyway, *bump*

(I''m too tired right now to write anything really coherent, so I''m bringing this thread where I can find it again)

Share this post


Link to post
Share on other sites
rmsgrey    153
I think doing creativity as (minimum(number_of_skills, average_skill_level)) should work.

My inclination would be to have creativity affect everything rather than just selected skills - either by having an unfamiliarity penalty for new or unusual uses of a skill which is reduced by creativity, or by having a flat bonus to all skills, or by having a bonus to attempts to improve in a skill (or bonus points to assign at level up if you really need to use a level based system...)

Share this post


Link to post
Share on other sites
Argus    118
On the other hand, many players actually want their character to be the absolute strongest in a particular skill.

The game mechanics should probably encourage diversity, but is it better to enforce it, or to create a game where diverse characters are simply more effective? ie. instead of forcing a character to pick up court etiquette (for example), create a game such that having the skill is useful enough to put points into anyway.

Also you can force exactly the same thing by making skill costs proportional to skill level already developed.

As for realism, as a counter-point to Marcus the Legionary, consider Tiger Woods. Much of his development is centred around golf, and AFAIK he didn''t need to take up pottery to get better.

Share this post


Link to post
Share on other sites
Saluk    127
You guys are too darn smart

I always feel like I learn a lot from these silly forums.

I''ve always liked skill systems with specialization, where they generally use a top-down system - you improve in the top skill, and you get to improve some in the sub skills.

But now listening from this, it should almost work the opposite way - you improve on the teeny-tiny skill and the overall skill improves.

It also gives me more ideas about having truly MEANINGFUL outcomes of skill progression. I''m absolutely SICK of each skill level making things a little bit better. Really, every time a skill gets a point, there should be extremely noticeable differences to the way you play.

Like, say you have swordfighting. At level 1 you can do a simple slash. It''s not much, but it gets you by, especially if you time it well. Then you get level 2 and it teaches you block. Suddenly, the game became a DIFFERENT GAME. Going from simple slashes to slash/block changes everything.

Ok, sorry for being a bit off-topic
I''m not totally sure about the inspiration thing. It sounds good in your example, but I don''t think it should happen more than 1 or 2 times in a character''s life. I can see how some things that have related knowledge may be mutually beneficial (learning calculus helps you understand physics, learning physics helps in calculus) but the totally unrelated increase should be rare. Lets say the skill levels go from 1 to 30. A skill should only be blocked at around level 25, and throughout the game, you should only ever be able to raise 1 or two skills past that level. This could keep the effect to something cool rather than an odd quirky gimmick.

Also, some disciplines might be more easily blocked in this way than others. Some skills aren''t as difficult to master as others.

Share this post


Link to post
Share on other sites
rmsgrey    153
quote:
Original post by Argus
On the other hand, many players actually want their character to be the absolute strongest in a particular skill.

The game mechanics should probably encourage diversity, but is it better to enforce it, or to create a game where diverse characters are simply more effective? ie. instead of forcing a character to pick up court etiquette (for example), create a game such that having the skill is useful enough to put points into anyway.


Throughout this thread, I have been thinking of a system which doesn''t force you to pick a specific skill to advance, rather making you learn any other skill... So there is no need for a character ever to learn court etiquette, but they will need to learn something other than pure hack''n''slash in order to reach the highest levels of swordcraft.

quote:

Also you can force exactly the same thing by making skill costs proportional to skill level already developed.

As for realism, as a counter-point to Marcus the Legionary, consider Tiger Woods. Much of his development is centred around golf, and AFAIK he didn''t need to take up pottery to get better.


I don''t know Tiger Wood''s life story, but I expect he has some skills beside golfing. Also, with the version of my "mental blocks" where they trigger on failing to advance three times in a row, some characters will be able to max out their swordfighting skill without ever doing anything else.

Share this post


Link to post
Share on other sites
rmsgrey    153
quote:
Original post by Saluk
You guys are too darn smart


*blushes* thanks
quote:

It also gives me more ideas about having truly MEANINGFUL outcomes of skill progression. I''m absolutely SICK of each skill level making things a little bit better. Really, every time a skill gets a point, there should be extremely noticeable differences to the way you play.

Like, say you have swordfighting. At level 1 you can do a simple slash. It''s not much, but it gets you by, especially if you time it well. Then you get level 2 and it teaches you block. Suddenly, the game became a DIFFERENT GAME. Going from simple slashes to slash/block changes everything.


In practice, in order to maintain a sense of progress, you''d probably want to have some sort of sub-levels in that system so that the character could get better at slashing while he waits to get the block, etc.
quote:

Ok, sorry for being a bit off-topic
I''m not totally sure about the inspiration thing. It sounds good in your example, but I don''t think it should happen more than 1 or 2 times in a character''s life. I can see how some things that have related knowledge may be mutually beneficial (learning calculus helps you understand physics, learning physics helps in calculus) but the totally unrelated increase should be rare. Lets say the skill levels go from 1 to 30. A skill should only be blocked at around level 25, and throughout the game, you should only ever be able to raise 1 or two skills past that level. This could keep the effect to something cool rather than an odd quirky gimmick.

Also, some disciplines might be more easily blocked in this way than others. Some skills aren''t as difficult to master as others.

When you get blocked in a given skill would be heavily implementation dependent - and also probably rely a little on luck. The totally unrelated increase (or in my later versions unblocking) isn''t about needing calculus to solve planetary orbits in physics, it''s about not being able to figure out integration, going away and doing something completely unrelated (kicking a football or playing Half Life or writing some code) and when you come back, being able to breeze past the point you were hung up on before.

As far as possible, I like to keep my suggestions for RPG mechanics pretty modular and independent of any specific implementation - on the other hand, I also try to keep them mutually compatible, so I''m actually thinking of a system with an asymptotic limit to skill level so no-one ever actually maxes out, skill improvement based on my experimenting/practicing system (in the learning through failure thread) and I''m still trying to figure out the hit point equivalent...

Share this post


Link to post
Share on other sites
tyrecius    122
I''ve just had a thought. Different disciplines assisting each other is rarely a matter of serendipity. Rather, when one reaches a certain expertise in both of the disciplines, one has a good enough grasp of the common elements that they begin to seem to be the same thing. When one first learns about algebra and geometry they are completely seperate things. But later with analytical geometry and calculus, they are united and learning more about geometry gives insight into algebra and vice versa.

So as a model of a skill system imagine a sandbox. At the beginning of the game it is completely empty, just bare wood. But on the floor of the sandbox are various points, one per possible skill.

Every time a skill is practiced, a little bit of sand is sprinkled over the point that represents that skill. And to determine how good somebody is at a skill, look at the amount of sand nearby. There might be a radius determined by some attribute (intelligence, creativity, player level, or perhaps skill by skill). So the computer calculates how much sand is within that radius and that is the skill rating.

Now, as time goes on there are many different piles, one per skill. But like any sand pile, the more you put on top, the more it spreads out. So at some point, especially if the radii increase as time goes on, piles will start to overlap. In the beginning only nearby sand piles will affect things. But at the end of the game, many different piles will affect a skill.

Now to actually implement this, many things can be dumbed down. The sandpiles could be quantized and one could make the system such that the different sandpiles don''t pile together. They could all be in alternate grids and all the alternate grids are added together when summing a skill.

Anyhow, this seems like it would model things well. Nearby skills would always effect a skill more than faraway skills. But as the player became more advanced (and perhaps the radii increased), each skill would effect the others more and more.

-D

Share this post


Link to post
Share on other sites
A few days ago in World Classics class, I was thinking about a Highlander-esque system, and found myself integrating some elements from this discussion. I should be writing a paper, so I''ll try to be brief:

My idea involved one hundred characters, each of which had one "attribute" from a list of 100 attributes, which are organized into ten categories. For instance, physical capabilities would be a category, and so ten random people would be imbued with one superhuman physical trait (strength, speed, agility, endurance, etc. I didn''t get into details) In that skill they would have a 55% capability, and they would have 5% in all the other skills in that category, so the strength guy would have 55% strength, 5% speed, etc. This is of course an abstract, and how it would interact with an actual stat system is up for grabs.

Anyway, any time you manage to kill another character, you get their skill power(s) (no doubt with some fancy pyrotechnics). So strength kills speed, so he has 60%, 60%, and eight 10% superpowers in the "physical power" category. As the game narrows down, you''ll have characters with 20 or more skills, and if you get all ten of a category, then they''re all up to 100%.

There''s a little bit more to this system, but if I get it to where it''s worth discussing I''ll start a thread for it.

The point is that grouping the skill into complimentary divisions can lead to a more harmonius stat development. Maybe subdivision could help organize the system.

Just a thought.

Share this post


Link to post
Share on other sites
Krysole    140
just slightly off topic perhaps a fundamental aspect of the MMORPG design when reworked might provide the solution to most all of the common probs. 'Age' plays an important part of life. If the game describes the players as asthetically immortal then the immortal players or Platinum Sword of Doom paradigm will inevitabally occur. Instead someone could make a HARD LIMIT on a players age makeing the game 'skill' related instead of playtime enforced. Major reworking required on the design part but it would mean players have to consider their characters 'complete' or entire life. Not just start->glory->more glory!

edit: i must watch what i'm doing...i must watch what i'm doin... i ...
-----< krysole >-------------------------------
//TODO: create a sig....no wait?

[edited by - krysole on March 25, 2003 10:57:53 PM]

Share this post


Link to post
Share on other sites