Desining a flawless use-based system

[Callaxes]

We haven't had a discussion about this in a long time.

Now let's take the perfect CRPG for roleplaying. We have diplomatic paths, stealth, science, combat, you name it! The Main Quest can be finished in just about anyway. The game world is vast and has alot of interaction, you pursue any type of job (cartographer, farmer, smith, etc.) join a guild that uses your trade and get verry far into the game without touching the storyline.

Basicly it's a CRPG in which you can play absolutely, positively, indupadably any role you want. But here's the challenge -> How to make a well ballanced use-based system that supports all this interaction and adds the feeling of progresion and improovement to your PC. How can you prevent abusive using and unballancing scenarios in which you use one skill more then anything else? How to make it hard to choose which skill to be a primary one.


Before you mention Betrayal at Krondor keep in mind that game was limited only to combat, where the choice between fighting with an axe, bow or spell is pretty damn easy. But ballancing it in a game that offers you social skills, science skills, crafting skills, survival skills, etecetera is harder, alot harder.

 

[JarlFrank]

I dunno, really, but maybe you could let skills get lower when you don't use them for a long time, or at least if you don't use them in real [like, using lockpick in a dungeon on a hard chest rather than just buying a chest with an easy lock and opening this over and over again]. This could be even worse, then, though, with somebody getting enormous heights in one skill, while the others drop to zero. So, bad system...

But the idea of skills leveling up depending on *how* you use them is good. For example, if you just use training and never use the skill in real situations, the skill will only rise to a certain level - and maybe once your skill will stick on that low level, because your character is used to only fighting against straw men and thus isn't used to enemies retaliating - and thus sucks at fighting against real, moving targets who hit you back.

 

[Oarfish]

Rolemaster used a system of potential and actual stats that worked quite well. Potential was determined at character creation time. Characters started off with stats somewhere below their potentials, and could increase towards them with use. I can't quite remember the details, but skill increases were also biased by usage when gaining a level and were limited by stats.

The problem with pure use based systems (especially in TES) are their lack of limits.

 

[Callaxes]

Yeah, the leveling should be determined on the quality of using the skill and not on the quantaty. There should be certain level bariers, like not going above level 30 if you don't break the castity belt of Ducess Monroe. And after that you only need to break locks that a harder then the castity belt to go on.


But another good idea is to make non-symetrical skills. Why should each skill go up to 100, surely there's less knowledge to gain overall from swordfighting then from academia. Also devide the spectrum into different types of XP

Practical experience (XP gained from using the skill)
Teoretical experience (Xp gained from studying the skill, either through books or dialog or trainers)
A third type of experience don't one yet

A skill related to science would have less Practical experience and more Teoretical experience. Once you've split the bar, you have more freedom to ballance the metods to increase these skills.

 

[Balor]

Well, they HAD limits. Question is, their limits are same for all races and classes, therefore all chracters would looks alike in the end (provided you took time to pump up your skills).

Anyway, I've already written that, and I'll say again:
Skills should be separated into 'practical learning' and 'theoretical learning' (pardon my English).
Purely practical learning should be VERY slow. Possible, but too slow to be feasible. (Think learning “insert martial art here” all by yourself... figuring out all the blows, throws, stances, etc… that would take ages, literally.)
Theoretical learning - is where quests, guilds and classes come into play. If you complete quests, join the right guild, or your are simply natural born 'insert class here' - you will get ability to rise skills to a certain point MUCH faster.
As for stats - Mistmare (mediocre game with a few good ideas) had 'time' as resource for that. Stands to reason - you can only get better if you spend a lot of time training.
For instance, if Oblivion devs were NOT designing 'by idiots for idiots' and demonic invasion would indeed be invading (like, for instance, in Space Rangers!), 'flower-picking' and 'running into walls' would be a huge waste of VERY valuable time.
Unfortunately, 'pump all skills' and 'become masters of all guilds' morons would not be pleased by that, so we have what we have.

 

[Balor]

@ Callaxes
Great minds think alike, huh?
That's rather obvous, though. Strange nobody (save a few MUDs) tried it yet.

 

[stargelman]

I used to like this type of system a lot. After all I'm a rational person, and usage based systems are the most realistic of all. If you think about it (not always a good idea) an experience based system with skills is a bit silly: you keep shooting people and it makes you better at science? That doesn't really make sense.

Unfortunately as we have seen starting with games like Daggerfall, this system has some pretty massive weaknesses and is a real nightmare to balance. Of course it would probably be much easier if you didn't derive attribute increases from skill increases as done in MW - "Wait, I'm about to level up. Let me jump around like a happy bunny for a few hours so I can get some more strength!" but even without that there is just so much potential for abuse. That aside, the traditional idea of levelling and associated automatic increases in various stats (life energy most of all) doesn't fit into all of this very well. If I do nothing but cooking soup, getting my cooking skill up all the way to 100, should I automatically rise to level 10 because of that? This gets even more problematic if you have levelled enemies which of course you shouldn't.

I guess its theoretically possible to come up with a near perfect system that does not suffer from such drawbacks, but I've yet to see one. In its absence, I think a system like the one employed in Gothic 3 is really neat: you get practical experience like in any XP based game, but if you want to learn a new trick, you still have to consult with someone more knowledgable than you. So in other words, you need both the book and real world experience. Just like in real life! Gothics system has the additional advantage of reasonably solving the whole life energy increase problem. You want more life energy? Beg some god for it!

 

[Callaxes]

@ Callaxes
Great minds think alike, huh?

I'll take that as a compliment

Although we're definately not the first ones to do this, I haven't seen a use-based game to split the skills in hlaf, but I know that M&M and Arcanum did it, you hard a number defining how good you were, but also 4 ranks of knowledge or training to acheive.

 

[Jormungandr]

Really, though, skill/use based systems shouldn't need levels at all. Increasing a skill will slowly add to it's attribute directly (and make health directly related to endurance or constitution or whatever). Since leveling up usually only results in attribute increases and more health, it really isn't necessary to make it so tiered.

As for the abuse, perhaps there could be time limitations on skill increase. You cannot increase a skill more than once per day, 3 days, week, or whatever; the time limit can be different for different skills.


I do love use-based systems, though, and they work best when everything improves gradually. You slowly get better and better at casting certain spells, it isn't tiered; you don't suddenly gain access to new spells, you just suck less at them, gradually until you get quite good at the more difficult magics.

And I like this splitting the skills up idea.

 

[Claw]

Like I certainly said before, I'd do away with levels altogether for a use-based system, maybe have three or so levels, each imposing soft skill limits, and use the overall experience for level limits. That would be the only significance of those levels. No attribute points, no health increase, no improved AC or whatever nonsense other systems may have.

Also, players should have pretty good stats to begin with, and the system I imagine would favour increasing skills in the middle range.

I have nothing more to say that I haven't said before.

 

[Haraldur]

Has anyone else played Discworld MUD? I think the system there worked quite well. Most of your skill advances would be done through paying your guild (Assassins, Wizards etc.) to train you, with an XP cost as well (which I think is unnecessary), but you could also improve your skills with the 'Task-Mastering' system. In that system, whenever you did an action there was a small chance that you would advance a level in skill AND, most importantly, that chance was determined by how hard the task was for you.

For example, in the newbie area there are some training dummies that you can 'attack'. It is easy to gain 5 levels of skill with a particular weapon type by doing this, but it then gets harder and harder (and more tedious) to raise it further. However, later on when you use more difficult weapons and are up against more difficult opponents there is more possibility or gaining levels again, though it is still quite slow, so most people still use their guild to train them up as it is much faster.

I think a system like this, based on d20 system, might work:

If the DC is 15, then if you roll a 15 on it, the worst possible roll that succeeds, then there should be a chance to gain a skill level. This chance could be, say, DC*0.003, so DC 18 would, if you succeeded by rolling 18, provide a 5.4% chance of gaining a level. We could also say that the chance can never be above 5.4%, so you do not get ridiculous things like someone succeeding on a DC 100 or something.

That way, if you want to improve your skill you need to find harder challenges, and the system simulates 'a new insight allows you to perform the task in a way you did not know before' or some such.

In case it is unclear, rolling 16 on a DC of 15 would not give any chance of gaining a skill level.

 

[Squeek]

While I can see the value of a use-based system in terms of role play, it just doesn't seem like something that could ever be much fun. So despite its realism and logic, it falls down in terms of game play.

Maybe a use-based system could work in tandem with a regular system. Instead of improving his skills, daily practice might make the character more consistent with them. So you might be less likely to fumble a weapon or miscast a spell that you're in the habit of using. It might make more sense if it were a temporary effect too.

Some of my greatest RPG moments were times when I leveled. I wouldn't want that de-emphasized in any way.

 

[Human Shield]

I used to like this type of system a lot. After all I'm a rational person, and usage based systems are the most realistic of all. If you think about it (not always a good idea) an experience based system with skills is a bit silly: you keep shooting people and it makes you better at science? That doesn't really make sense.

And hitting rabbits with a sword all day makes you the best swordsman?

A progressive system is better, one where more challenging things are faced instead of grind. Why have the player play out boring practice and study? Reading science books mite make you better at science but does the player need to see it? XP systems provide progressive events but both could be done better.

 

[Callaxes]

We're not here to argue about which one is better. We're here to figure the best blueprint for a use-based system in a complex CRPG.

 

[DraQ]

Greetings all. I've been pondering joining this forum for some time.
At least one person frequenting this forum will probably recognize me as a member of TESF, nicknamed Bluescale. Now, the fact that I come here from TESF doesn't mean that I'm going to spam these boards with multiple threads where I would try to gain insight by asking such important questions as "IS Plenscaep Tornment as good as Oblibians!?! I herd graphx suxx!1 Dos it has any shadurz!? kthxbye." warranting righteous (and fully justified) wrath of codexers, so don't freak out.
Also, trying to coax me into clicking on obfuscated shock-porn links is pretty futile, so don't bother.


So much for the introduction.
Now, regarding use based systems:

Personally I'd start simple, then think about improving the system and expanding it by adding new features.
The obvious way to counteract grinding your skills by performing simple tasks over and over again (as it's possible in TES games) would be removing experience gain from performing simple tasks over and over again. One way to do that, in a system where all tasks covered by skills have certain, skill dependent chance of failure, would be scaling experience gains for successfully performing given task to be directly proportional to the chance of failure. It's easy to notice, that the maximum average xp gained per attempt would be at 50% chance of failure. Once the character can perform certain feat (like splattering bunnies with his zweihander) with no failures, no experience would be gained and to further develop given skill, character would have to seek some more difficult challenges. The best way to prevent mastering every skill that comes to my mind would be scaling the progress, so that mastering even one skill wouldn't happen so often on a single playthrough. Sure, this kind of measure is easy to bypass provided that player is determined enough, but if someone is really willing to invest countless hours into grinding his character's skills (putting something heavy on your keyboard won't work in this system) I don't see why should anyone stop him from being a loser.

I think that, while not necessarily flawless, this system would be decent enough while remaining very simple and thus easy to build on.

 

[gromit]

I'll go with the "brute force" option just to get it out there... quests usually giving xp, and this dream rpg having multiple solutions to all of them, we could instead assign skill-specific xp to each of these methods in different chunks, even breaking the distribution down across different actions or phases. Kick in the front door and kill everyone in the room, give such and such XP to firearms, throwing, melee, whatever, not real hard to track the ratio over the course of a known battle. Get in and out unseen by a set of guards, there's your sneak XP. Pick the lock on the safe or blow it up, resulting XP as appropriate. You get the idea.

Also note the opportunity to almost completely erradicate grind, if you're so inclined, by limiting this strictly to quest-related actions, or being very conservative in when and where you assign points. I haven't thought about it enough to know if I'd want to, but one could even go as far as to nullify related handouts as different phases of a quest are handled in other methods, e.g. making it a waste of ammo and health to shoot up a place you've already snuck out of, or, probably causing less outrage, not giving XP for picking the back-door lock to a hideout-cum-graveyard.

Not a very systematic solution, by definition and implementation completely unfit for pure sandbox play, but I like it a lot for something Arcanumy.

 

[Jasede]

snip

Just a question, off-topic- if that's really what you want in an RPG, why not play a LARP multiplayer game with a skill/stat system, like a MUD or some elitist NWN PW? There are quite a few good, high-standard ones where you can do all that, apart maybe from the use-based system.

Ok, carry on. Personally, I loathe use-based. If the game already gives me the freedom to create my character any way I like, I'd also prefer it if the game trusted me to allocate my skillpoints myself each level in a way that makes sense.

 

[Callaxes]

That's not what I realy want. I think use-based systems should be limited to just combat-heavy CRPGs like BaK. But I think a discussion on how to improove the system is very interesting. We had a bunch of these before I registered.

 

[cardtrick]

When I was playing MUDs, I liked use-based systems, and the more complex the better. I've never played an MMORPG, but I imagine I'd feel the same way about those.

For CRPGs, I've never played a straight use-based system that I've really liked. I believe it's impossible to have a pure use-based system that neither rewards tedium nor is unduly unbalanced.

However, I do think that hybrid systems can be very fun and workable in a CRPG. The best (only?) example of this is Prelude to Darkness. I loved that character development system, and I would definitely like to see it used again. (This is also similar to the taskmaster system used in Discworld MUD, which someone mentioned above.)

 

[Balor]

Well, about 'killing rabbits for 1000 hours makes your uber swordsman'.
Easily fixed - simply make 'swordmanship' not simply a skill '0% - don't know what a 'sword' is for, 100% - uber swordsman', but a set of moves and stances, both offencive and defencive.
You can train up your stances, and some offensive moves, but most defensive moves (and a good share of offensive) would require a partner, at the very least.
Stands to reason, right?
Some moves would require MULTIPLE partners... or enemies .
Therefore, by killing 1000 rabbits, it simply make you an uber rabbit-hunter.
Anyone who has wider array of trained moves will make a very short work of you.

 

[DraQ]

One way to do that, in a system where all tasks covered by skills have certain, skill dependent chance of failure, would be scaling experience gains for successfully performing given task to be directly proportional to the chance of failure. It's easy to notice, that the maximum average xp gained per attempt would be at 50% chance of failure. Once the character can perform certain feat (like splattering bunnies with his zweihander) with no failures, no experience would be gained and to further develop given skill, character would have to seek some more difficult challenges.

I thought it might be a good idea in a more compact form.
As I wrote on TESF:

The beauty of direct proportionality is that you'd automagically get most exp at 50%. Lower -

increased exp doesn't make up for low success rate, higher - leveling is slowed down by low exp gains.

We get:

y=-ax^2+ax

Where:

x=success rate (as fraction of 1)
y=average exp gain per attempt
a=XP multiplier (positive real number selected for optimal leveling speed)

It is simple square function with maximum at x=0.5 regardless of a.

I think that implementing various additional features, such as learning different techniques, as suggested by Balor, would be reasonably easy in such system.

Now, does anyone have an idea how to make specializing preferable to making jack of all trades?

snip

Just a question, off-topic- if that's really what you want in an RPG, why not play a LARP multiplayer game with a skill/stat system, like a MUD or some elitist NWN PW? There are quite a few good, high-standard ones where you can do all that, apart maybe from the use-based system.

Ok, carry on. Personally, I loathe use-based. If the game already gives me the freedom to create my character any way I like, I'd also prefer it if the game trusted me to allocate my skillpoints myself each level in a way that makes sense.

Wouldn't the next logical step be the game trusting you to imagine the consequences and roleplay accordingly though? It would certainly improve game's flexibility as only a finite number of possible player's actions can be predicted leading to finite number of possible paths determined by finite number of preprogrammed consequences.

Personally I prefer consequences of my actions to be handled by the game, including the consequences of the way I earn my XPs.

 

[Jormungandr]

Now, does anyone have an idea how to make specializing preferable to making jack of all trades?

Perhaps by implementing a sort of synergy bonus for having high values in related skills or, even better: giving bonuses for choosing multiple related skills as class skills (which would emphasis making a more specialized class rather than just working on the related skills on the side). Although, really, there's nothing wrong with making a more rounded character, especially in a non party based game.

Although that does bring up another point, and that is the Class itself. There needs to be something special about the skills you select for you class that sets them aside from the unchosen skills--especially if we did away with the whole leveling up bit. The miscellaneous skills need to be static and un-increasable, or penalized 25% or 50% of their value, or simply capped off early at say 50 or 75. Something to stop the class from being pointless later on in the game.

Personally I prefer consequences of my actions to be handled by the game, including the consequences of the way I earn my XPs.

Agreed.

 

[Cassidy]

A use-based skill development only works for FPSes with RPG elements like Deus Ex and System Shock 2, because it makes more sense in games where you have a partially direct control of your character(partially because your accuracy might still be influenced by skills and non-combat skills might be 100% PC-based instead) and might make gameplay more challenging if well implemented(as now you can't just spend points in rifle skill and get an uber weapon you can handle easily).

However, for "pure" RPGs it's just a blooming potential of failure and a beacon for twitch gamers, LARPers and console kiddies. Maybe if it was only a part of the skill development while most was still based on the way you allocated skills on your character. If a feasible way to evict grindfests is presented, maybe I'll change my opinion. But programming a munchkin-proof realistic algorithm for use-based development that wouldn't force the player to do the same thing repetitively either would be a daunting task.

 

[Azael]

I liked the hybrid system used in Prelude to Darkness where you'd get skill increases from using the skill, as well as skill points you could use to raise the skill of your choice from completing quests. This gives you a little more flexibility in shaping your character than a strict use-based system. As far as I remember, skill grinding wasn't required or encouraged.

 

[Claw]

Now, does anyone have an idea how to make specializing preferable to making jack of all trades?

That's easy. Limit overall skill. Jack of all trades only works well in systems where you can actually be very good at everything, and not if you can only hope to be mediocre at everything. Static or inert attributes may also be helpful, depending on implementation.

I envision a tree-like skill structure, where the higher tiers are inceasingly inert, while the lowest can be trained fairly quickly. The actual skill would be a combination of all tiers.
That way, a player would strongly determine his character's abilities during creation through the higher tier values which will change less during the game while leaving room for character development, mainly in the lower tiers.
For instance, a swordfighter could easily learn a new technique with a sword, and quickly be good at it because of his high values in the relevant higher skill tiers. However, lacking these "basic" skill would limit a character's ability to reasonable develop his swordfighting skills.