First, an apology to everyone for inadvertently transgressing by directly resurrecting the previous thread on this topic. I was directed to the thread from an outside link, and never thought to check the dates on previous posts. For those who didn't follow the previous thread (or don't want to have to reread it), the discussion was about possible alternatives to the traditional "murder-based" experience system where characters get abstract experience points from killing monsters and whenever enough points accumulate, all the character's skills improve. Ideas seemed to divide into three types: those that kept the basic abstract experience points, but tried to award exp for actions other than kills; those that advocated only improving skills used rather than having arbitrary choice of skills to improve; and finally those that just suggested a feature that could be added to either system. Experience related: A system where exp gained for actions varies by character class and action type (warriors gain lots for fighting, little for lock-picking, rogues vice versa). primarily story-based exp exp handed out based on assessment by other party members exp for reaching checkpoints (like in Deus Ex) skill advancement by putting exp into linking base skills to generate new skills General/mixed: Skill values being hidden from the player, and having to be deduced by trying to use the skill. higher exp awarded for failed actions than successes genocide discouragement by increasing maintenance costs with level. suggestion that failure-based exp/skill improvement follows real life more closely than success based awards. genocide discouragement by only awarding exp for actions with significant risk of failure diminishing returns on repetition of tasks (large amounts of logging overhead) Skill related: A system where learning on each skill can be switched on and off, with gained experience divided by the number of skills being trained, and only relevant skills to the action performed actually gaining. Potential abuse pointed out: switch skills learning on/off as you use them so only the skills that gain exp are on at any given time. skill atrophy through disuse skill improvement with any use (succeed or fail) but only if success/failure threshhold falls close enough to current skill level skill improvement system entirely hidden from player so character seemingly adapts automatically to player's play style Call of Cthulhu system - potential to increase a skill comes with significant success (using it when it affects character survival/success rather than in training). Actual improvement requires a skill check with failure causing improvement. Finally, my errant post suggested a system where every skill has a current base skill level, which is the average check result, and a randomness level, that sets the range of the random element in a skill check - for example, with a base skill of 5 and randomness level of 2, the skill check would produce an unmodified result from 3 to 7 (5 plus or minus up to 2). The player can choose one of two states for the skill to be in - in the first, which represents the character experimenting with different ways of using the skill, ignores the randomness level, instead using the default (maximum) randomness level for that skill. Any successful skill check causes the base skill to change by some fixed proportion of the random result (say half for the sake of illustration, probably rather lower in practice) so, for example, if the skill above (base 5, randomness 2) were in this mode, and the default randomness level were 5, then a check which succeeded with an unmodified result of 9 (random number of +4) then the base skill would increase (to 7) If, however a check then succeeded with an unmodified result of 3 (new base skill of 7 and random number of -4) then the base skill would decrease (back to 5). In the second mode, representing a consistent approach, the base skill level of a skill is fixed, all checks using the skill get a fixed bonus (say +2 for my example) and any use of the skill, whether success or failure, reduces the randomness level for that skill by a fixed amount (say 1 for my example). So the unlucky player in my example above, having decided to change mode for the skill before using it on some really easy checks and losing all his progress, now tries to use his skill again, getting a result in the range 3 to 7, and the randomness level decreases to 1 whether his result was a success or failure. On his next use of the skill, the result would be between 4 and 6, and his randomness with the skill would reduce to 0. Until he changes back to the experimental mode, all skill checks with this skill will give a result of 5. Some quick (pre-emptive) comments on my system: having the base skill decrease sometimes when in experimental mode represents learning bad habits. From a game-balance perspective, it discourages genocide and non-combat equivalents by making them (on average) totally non-productive in terms of raising base skill. In practice, you'd probably want to introduce a downwards drift by subtracting a small amount of the new base skill value after each success (otherwise the player, by bashing goblins for long enough, would sooner or later get his combat skill to max out). Alternatively, keep the actual skill level hidden so the player has no idea when to stop goblin-bashing. If you actually read my previous posting of this system, I gave experimenting a fixed penalty rather than conservatism a fixed bonus. I thought this might make the explanation a little clearer, there's no other difference. When switching between conservative and experimental modes, you may want to increase the randomness factor slightly (or over time while in experimental mode). Also, you may want to stop it decreasing before it hits zero to preserve some randomness. The decrease in randomness level from failed checks could be different from that for successes, and in whichever way your biases run. I suspect emphasising failure is probably more realistic in conservative mode since since you are attempting to hone an existing technique. Critical success/failure in conservative mode (if you have them) should probably result in no/double change in randomness level - a perfect result teaches you nothing about consistency, while fumbling highlights your mistakes. In any serious implementation, you probably want to take advantage of the computer's ability to handle large numbers or floats - having an initial base skill of, say, 10 out of 255 and randomness level of 100 with fixed decrease of 1 and bonus of +5 to checks in conservative mode and change of (5% of random number)-1 in experimental mode sounds plausible. Worth playtesting to get right. Might want to consider a modification where the bonus for conservative mode starts at 0 and increases gradually to a set value whilst being conservative to reflect slight improvement due to honing of known technique(s) in addition to the reduced randomness. Anyway, any other comments on alternative means of character improvement (ie not the standard hack-and-slash for experience points) are welcome. [edit] put in link to older thread - my skills having improved over the past few months [/edit] [edited by - rmsgrey on March 30, 2003 3:28:13 AM]

I''m a follower of the "stat and skill improvement with practice" camp. But I also think that advancement should be organized in a different way than it is now. Specifically, I think that the player should not have to spend play time advancing his character, nor do I think that stat and skill improvement should be unbounded. Here''s how I''d like to do things.

During gameplay, the player goes about accomplishing tasks, exploring, interacting with other characters, and generally just playing the game. While in gameplay, he earns NO EXPERIENCE. Shock, shock! He doesn''t need it though; the challenges he faces can be overcome in many ways, and his character is almost guaranteed to be competent enough to use at least one of them. If a problem can''t be solved at his current level of ability and skill, that''s OK; he doesn''t need to solve it right now because there are many other things to do.

However, as the character goes about his business, as the PC observes other characters using skills at a rank greater than his own, he is learning from observation. If he''s never seen the skill before, he now has the ability to learn it through imitation. Furthermore, he may make contact with characters willing to teach him new skills and techniques.

Now, it''s the end of a gameplay cycle. The current adventures are over, and it''s time for the PC to get back to his routine. After all, "adventures" are breaks from the routine, or they wouldn''t be very adventurous would they? But the routine isn''t as fun as gameplay, so we the designer don''t make the player actually play it out. Instead, we present the player with a list of options for his character''s offtime actions. This list is based on the opportunities the player has uncovered during gameplay. He can work at jobs to earn money and improve job-related skills. He can train with tutors to learn their skills at an accelerated rate (but at a monetary price). He can study books or manuals he''s found, to try to learn things on his own. If he has the skills to make things, he can spend his time making them.

All this takes a fair timespan in game (in fact, it takes until the next opportunity for adventure), but for the player the process is a matter of a few minutes. Then the next gameplay cycle begins, and the player can now use his newly improved skills to try and tackle new challenges or old ones he couldn''t previously handle.

A side note: In terms of how success and failure is handled, the ideal for me would be a complete simulation of character actions rather than an abstract success/failure model. But barring that, I rather like the Silhouette System''s method for handling skill checks. When you make a "success roll", your skill is the number of d6 you roll. You choose the highest of those rolls, and add your relevant statistic for your final check number. In this way, high skill improves your average roll but high stats improve your maximum roll. Good skill means more consistent results and fewer screwups, but only natural talent lets you pull off really astounding feats.

-STC

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

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

How about we make skills grow in adaptation cycle. What we do is assign each skill a basic success rate, and then record an adaptation meter which acts as a percentage of success added on. To visualize this, lets say we''re enabled in SKILLJ, be it someting incandescent or swiping someone''s shoes. SKILLJ has basic 25% success rate, leaving 75% up for grabs. Each time the player uses the skill, the meter is increase by a number selected at random inclusively between 1 and n where n is derived elsewhere, by skill, by arbitration, whatever. What we then do is modify the effective success rate as so:

Success = BaseSuccess + ( (100-BaseSuccess) * AdaptationRate / 100 ) * (Enemy''s percentage ability to dodge / 100)

What we have now is a way for players to become more adept at a particular skill based on usage, without having the player stumble around in the dark for an hour.

I don''t know if this already suggested in another form or not, but my logic is something simple for the player to work on. If you want to introduce attrition, then you could have another random degree used for subtraction, say a 10% chance -1 if you don''t use a skill that battle.

OK, since the original thread here has been re-necroed by someone else, I thought I''d *bump* this one directly since it''s a little easier to catch the main points of the old one from.

Skill by practice is the best way to go about things, IMHO.

It has a nice effect, that being that it doesn''t directly require murder (increases in combat skills are more due to the use of the skills rather than the kill itself), so implementing proper fleeing/surrender systems are possible.

I don''t have so much of a problem with characters going after weaker monsters for experience...BUT there should be the chance that the player can still die.

In other words, a level 1 goblin SHOULD have a chance (however miniscule) of killing that level 50 hero. And a horde of 50 level 1 goblins should represent approximately the same challenge as a single level 50 monster.

Most RPGs don''t represent this...in DnD, when you''re level 20, no level 1 monster can hit you, much less kill you. In console RPGs, where levelling is encouraged, its a low risk practice with nearly no attrition.

Attrition is a big part of the challenge of those 50 goblins.

Sure, the hero can easily dispose of them as individuals, but over time, he starts to accumulate wounds...his movements begin to slow and he starts to tire. There should be a real danger from prolongued battle that blurs the line between low and high levels.

Its not EXP itself, or the fact that murdering creatures generates a reward, but rather that most XP hunts involve no risk, and low level creatures become meaningless.

ElAntonous has a good idea I think ... and I would like to temper it with another ...

Most players of RPGS are power players, meaning there are not really role players, but people who enjoy succeeding, being better than others, and impressing others and themselves (through neat feets they acomplish) ... they would rather play an AD&D character with stats 18, 16, 15, 17, 14, 17 then one like 12, 17, 11, 14, 7, 15.

Even still though ... they still get sickened of the sheer tedium of having to slaughter hundreds of meaningless, automatic NPCs to make levels once they have a few ...

So what to do ... what I propose, is that people do implement some sort of system where the danger increases noticably as you become outclassed and outnumbered ... so that you can walk safely alone and not paying much attention as a level 30 fighter, AS LONG AS, there is no powerfull (high level) or organized (large numbers) group set against you. I don''t mind that a goblin individually has no chance to kill a half health, half stamina, half everything level 30 fighter ... but if he can just do SOMETHING to reduce him ... then eventually it would matter.

One thing I''ve never seen used in RPGs is zones of control. Some games have stamina, some games have backstabing, some games do not allow retreat, or have fast monsters. But where are the games that make is harder to get away from 10 weak enemies who surround you, than just 1 who is in your path (you know, the ability for them to close, swipe, or sprint at you, while you dodge the one ...). Where are the games that allow running speed to flucuate, so that enemies may gain on you while you flee, or fall behind (but not ever falling farther behind) ... where is an element of morale or determination, that allows someone fighting for his life to run just a few moments longer ...

It all comes down to one fact. In order to improve a character in a stat based game, you must spend time doing repetitious tasks. If you could do it without time waste, then you could max your character instantly. If the tasks were not repetitious then you would be able to advance the character too quickly (non-reptitious tasks are limited in number).

Its all a matter of making those tasks that the player must repeat, interesting. Ultima Online, for example.. failed in this. Instead of killing hundreds of monsters, I spend hours chopping wood, or building armor. In concept, its great. But in reality, it sucks. Killing mobs is more exciting, random and a more interesting repetition.

Hiding stats from players is an even worse idea. Many, if not most, players enjoy "maxing" their characters. If you remove this possibility then you remove a large majority of the fun from games for many players. This is even more relevant in MMORPG type games, where statistics are all there are to differentiate thousands of characters (Yah, like I really care if my character has a small goatee, and somebody else''s has a medium goatee).

Hence- my opinion. Alternative XP/ character development is relatively impossible to implement in a manner which is fun for the player at least for now.

In real life there is motivation to be lazy (its fun). In a game, if you are lazy you''re left staring at a static image.. so there is no motivation for laziness. This causes the problem that "alternative xp" tries to eliminate: Half of the game''s players are nearly identical....

I''ve always been under the impression that instead of trying to get rid of battle as the main advancement, people should try to CHANGE battle. Developing a system of battle with elements that are similiar to the NES game Mike Tyson''s Punch Out, would be ideal. For any who don''t know, you are presented with a target. This target has specific weaknesses, and specific strengths. The target attacks using a pattern, and within this pattern there are moments of vulnerability, where the attacker (player) can damage the target, and rarely other times.

Monsters/mobs would conform to this. It becomes impossible to simply kill hundreds of mob''s nonstop, since it becomes largely a matter of skill which will give a different differentiating factor between players, other than time. There are some people who can play "Punch Out" all the way through without taking any damage, and there are others who can play for months and never beat the game.

Of course this is an entirely new concept, and would take massive tweaking.. not to mention determining how different classes would interact with the same target, or multiple player characters attacking a single target, etc... but I think its the direction that games will head towards..

Another problem is simply feasability. A system such as that would require calculation of responses with millisecond delays completely changing the result. In the scenario of an MMORPG, this is unacceptable due to lag... and allowing mobs/monsters to compensate for player''s lag by "freezing" themselves, is obvious invitation for exploitation... so there seems to be no solution, at least for now..

quote: Original post by haro
Monsters/mobs would conform to this. It becomes impossible to simply kill hundreds of mob''s nonstop, since it becomes largely a matter of skill which will give a different differentiating factor between players, other than time. There are some people who can play "Punch Out" all the way through without taking any damage, and there are others who can play for months and never beat the game.

I have to agree with this. and I believe, Zelda implemented this system (the bosses). Monsters must be beaten in certain ways, not just the brute hack and slash technique.

I think damage and armor system need attentions also. For example, in the case of 1 lvl50 character vs 50 lvl1 goblins, these goblins barely do any damage, maybe because the lvl50 guy''s armor was too thick. And the damage dealt by the character overwhelms the goblins HP. If any of you have played Dragonlance (an old tactical RPG game for Amiga/DOS), I think they implemented this very well. Even if your characters are lvl20 with 150HP, you still find some troubles fighting a horde of minotaurs whose HP is only 70 points, because the damage you dealt has a very little difference than the damage dealt by the minotaurs. In some cases, minotaurs may inflict more damage than you do. Even if you fight some pretty low level thugs, they still do damage, although not much but can still be considered threats.

The worse XP/skill system I have seen is in AC2 (couldn''t resist to mention it, I hate it, completely stats based, very little skill involved). In PvP, a lvl35 character fights a lvl50 character. the lvl50 guy deals 500 damage, but the lvl35 guy deals only (at most) 100 damage, that''s a HUGE difference (not to mention that the chance of hitting the lvl50 guy is about 1:5). With two strikes, the lvl50 wins with full HP.

Current project: 2D in Direct3D engine.
% completed: ~35%
Status: Active.

quote: Original post by alnite
The worse XP/skill system I have seen is in AC2 (couldn''t resist to mention it, I hate it, completely stats based, very little skill involved). In PvP, a lvl35 character fights a lvl50 character. the lvl50 guy deals 500 damage, but the lvl35 guy deals only (at most) 100 damage, that''s a HUGE difference (not to mention that the chance of hitting the lvl50 guy is about 1:5). With two strikes, the lvl50 wins with full HP.

Current project: 2D in Direct3D engine.
% completed: ~35%
Status: Active.

But, another little problem arises without giving very large gains for levels. Where''s the motivation to level? If a level 35 can potentially do just about as much damage as a level 50 (With some luck), then why gain the extra 15 levels? In MMORPG''s 15 levels can mean dozens, if not hundreds, of hours playing time... And with clients having less motivation to play, then the average subscription time would likely decrease substantially.

Too bad a system where skills and stats are earned could never work. Ie- to learn a certain spell, you had to explore an ancient cave, read the glyphs from the wall, and from then on you could cast the spell..

Sounds nice, but in reality just turns into a boring journey for every person who makes a spell caster. They load up their spell-list faq, load up their maps, and waste hours running through caves each level to get their spells... No way to really make it any fun...

Then there''s always the idea of "the more people using a certain spell/skill, the more difficult it is for new people to learn that skill." but that turned out to be one of the worst mistakes ultima online made, imo. It just creates a boring atmosphere since there is no chance for advancement to the very top.

I think a very interesing idea (though, somewhat of a different genre) would be a tradewars meets medievial era type game. The game is dynamically created. Each player that joins the game is alloted a piece of land and a hut to begin with. You then have the ability to develop your hut over time, into a castle. Spread out your lands, connect your lands with roads for quick travel, protect your lands with everything from walls, to archers, to proximity mines (not chemically detonated/operated, magically detonated/operated!). Of course the idea is to extend your empire, forge alliances, and eventually take over the world. When the world was taken over, or a total of 80% of the land voted for a restart, the game would begin again (you gain more votes, the more land you control, so if one person managed to control 80% of the world, he could force armageddon if he wanted). Your provinces and buildings would develop using real life time (with the ability for queuing a list of operations to be done), but you could develop yourself and your armies in real time.

For example, there would be wilderness in the game. The wilderness would have to be cleared before you could settle that land (or you could leave it as wilderness, and use it as cover for an attack against another province on the other side of the wilderness). There could be caves/dungeons which would require entire armies to conquer, and when conquered you could gain relics for your "avatar", or advances for your entire civilization (ie- all advances now require 10% less time). These relics/advances could be traded to other nations for anything from land to soldiers to other relics...

Ooo.. I get excited just talking about it, but I''m rambling and its 4am here.

A few quick ideas (feel free to develop them further)

Mini-games for skill improvement - rather than just clicking "make armour" and waiting 5 minutes while a progress bar fills in, instead play the "armour making mini-game"

Item based improvement - rather than having to perform set repetitious tasks to enhance a given skill, instead perform any task you feel like, and get a skill enhancement item at the end - which can be traded for the skill you wish you had but don''t want to bother training in...

Quest-based improvement - rather than improving for every random act of genocide, make it so that you only improve when you do something plot related

And, as an aside, my rule of thumb as a DM for AD&D 2nd edition (though I''m sure it could be generalised) is that each level higher doubles the number of opponents of a given level in an encounter of appropriate difficulty. Which means that a mob of Kobold (roughly level 0.5) want to outnumber 1st level PCs about 3:2, 2nd level 3:1, 3rd level 6:1, etc. Of course, this still doesn''t provide a fair fight - the PCs should win handily, but take noticeable damage. Half a dozen or so of these encounters should leave the PCs in serious need of R&R. So for a level 20 character, you''d be looking at about a million goblins to offer an interesting fight... though, at the rate of one goblin killed every second, you''re looking at about 11 days of constant combat to wade through them all...