08 Jul, 2011, Oliver wrote in the 1st comment:
Votes: 0
For a little while, I've been pondering the nature of the way that MUDs generally (though not always) handle the method of players' skills gaining points. For reference, this applies to any game in which players can increase their skills or stats by using a skill; generally speaking, there is a random chance to gain points on use. This is primarily the domain of DIKU-rivatives, but not solely.

I don't particularly like the way it's handled because it can lead to a lot of player frustration at times. You can use Super Deathray a thousand times before gaining one skill point, but using Grab Object once might net you one on your second try. When the skill that refuses to gain is more expensive than another, this becomes more frustrating still. Some games try to handle this by weighting the chance to gain according to arbitrary scales, but it still doesn't remove the possibility of terrible random streaks.

Has anyone ever attempted a system wherein abilities have a constant (or pseudorandom) value, and once the ability has been used that many times, you automatically gain a point? This idea seemed to be the most logical to me, as it removes that random component and risk of frustration. Obviously, the value would have to scale with the skill's ease of use, and the player's current incremented value would (or could) be kept invisible.

Has anyone found another solution that they like better?
08 Jul, 2011, quixadhal wrote in the 2nd comment:
Votes: 0
In my particular DikuMUD, skill gains are a pair of curves. When you first acquire a skill, you know it at 1% and are very clumsy with it (spells, skills, whatever). Each time you cast, you have a small chance to improve. As your skill level goes up, the odds of you doing it right improve, and so you have a better chance to gain another point. This continues up until around 60%. At that point, you now start learning from the times you fail, rather than the times you do it right, and so your chances to improve start diminishing again until you hit the cap at 90%.

You can always spend a "practice session" to improve as well, so folks who don't have the old-school grind patience can avoid some of the frustration…. but then if you don't like grinds, my old Diku isn't the game for you. :)
08 Jul, 2011, Oliver wrote in the 3rd comment:
Votes: 0
quixadhal said:
In my particular DikuMUD, skill gains are a pair of curves. When you first acquire a skill, you know it at 1% and are very clumsy with it (spells, skills, whatever). Each time you cast, you have a small chance to improve. As your skill level goes up, the odds of you doing it right improve, and so you have a better chance to gain another point. This continues up until around 60%. At that point, you now start learning from the times you fail, rather than the times you do it right, and so your chances to improve start diminishing again until you hit the cap at 90%.

You can always spend a "practice session" to improve as well, so folks who don't have the old-school grind patience can avoid some of the frustration…. but then if you don't like grinds, my old Diku isn't the game for you. :)


Oh yes. I'm familiar with that system; sounds pretty stock?

I'm mostly interested in alternative methods that remove the probability of long streaks of randomly-induced failures (though decreasing chances / higher usage caps as skills approach their maximum is certainly reasonable).
08 Jul, 2011, Vatiken wrote in the 4th comment:
Votes: 0
I personally use a "Elder Scrolls" style of skill management in which each skill has its own level and experience. Codebase hasn't gone public yet so can't say if it will work great in action, but in testing, it seems like a solid improvement over the diku random method.
08 Jul, 2011, Littlehorn wrote in the 5th comment:
Votes: 0
Oliver said:
Has anyone ever attempted a system wherein abilities have a constant (or pseudorandom) value, and once the ability has been used that many times, you automatically gain a point? This idea seemed to be the most logical to me, as it removes that random component and risk of frustration. Obviously, the value would have to scale with the skill's ease of use, and the player's current incremented value would (or could) be kept invisible.

Has anyone found another solution that they like better?


It would be abused to the point where everyone would just spam it until it's maxed. It also removes the excitement from random gains. That reduces your player experience some. On top of that, it reduces the amount of hours your players play your game. Random means on the first try or the one-hundredth try. Thus, increasing the amount of hours spent in game in order to progress your abilities. Anything else could mean the lifespan of a character is reduced to the point where they find end-game faster to where they either A) quit the game or B) roll a alt. Which if you ask me, is something you do not want to do.
08 Jul, 2011, Tyche wrote in the 6th comment:
Votes: 0
I personally prefer systems where the player chooses which skill they want to increase when they develop their character.

However, Chaosium's BRP system has characteristics that might provide some interesting ideas. It's based on game sessions, which isn't very useful for muds. However, one could base development on level advancement (if you use levels) or perhaps time-played. If a character succeeds in using a particular skill and the action has a high difficultly, during the course of a level or time period, it gets flagged. During development a percentile roll is made against the skill, adding in the characters intelligence bonus. If they roll higher than their skill rating they gain 1d6 percentage points in that skill.

The system has the affect of making skills harder and harder to increase and doesn't give any benefit to spam learning a skill. It does require that you track skill success and assign a difficulty rating to skill rolls. For example, if one was to use combat skills or spells, one would have them flagged if successful if they are used them during the course of battling a difficult to defeat foe. If one was increasing a swimming, climbing or jumping skill, the action would have to be flagged as difficult by measure of current, temperature, encumbrance, distance, height and/or danger.
08 Jul, 2011, Rarva.Riendf wrote in the 7th comment:
Votes: 0
Littlehorn said:
It would be abused to the point where everyone would just spam it until it's maxed. It also removes the excitement from random gains.

Random or not whatever the system it will be abused any way. It may or may not be difficult to abuse but it will.
If you need to cast a spell, people will cast the spell on some easy mobs, or on themselves and heal when they are too hurt etc. (may cost a lot of gold but that it is still a trade off)
And dont even start thinking: I will only make it so you can gain in a skill by not using it a lot in a row, people will jsut make so they practice many spell at once instead of one. With multiple aliases so the order is never the same, so you cannot even detect it.
You can not 'win'. Just make it so it increase naturally while players level. When they get to their max level, most of their usual skill shoudl be at the max already, unless they had a 3 int all the way….

Hell I was using autofire in the old Dungeon Master to practice fighting skill……(I learned this way that you could become even more than master…=))
08 Jul, 2011, quixadhal wrote in the 8th comment:
Votes: 0
Oliver said:
Oh yes. I'm familiar with that system; sounds pretty stock?


Perhaps it became stock in later derivatives, but since I coded the changes myself, it wasn't stock at the time (original DikuMUD). I believe the stock system only gave skill gains for success.

Littlehorn said:
It would be abused to the point where everyone would just spam it until it's maxed. It also removes the excitement from random gains. That reduces your player experience some. On top of that, it reduces the amount of hours your players play your game.


I hope you're not suggesting that people only play games because of the time it takes to hit some mythical "level cap"? Maybe certain games, but I hardly think that's a healthy attitude to take when designing your game system.



One idea that might be appealing if you want to remove randomness from the equation is to only give a skillup the first time you successfully use a skill in a given situation, or against a given opponent. So, if you're casting fireball, the first time you hit an orc with it, you'd gain a skill point improving fireball. You'd only gain another one when you hit something else, like a goblin. With skills like pick-lock, you may need to track individual locks (doors, chests, whatever).

That way, there's no random dice rolls involved, but the players still have to travel and play through content in order to advance their skills.
08 Jul, 2011, KaVir wrote in the 9th comment:
Votes: 0
Oliver said:
Has anyone ever attempted a system wherein abilities have a constant (or pseudorandom) value, and once the ability has been used that many times, you automatically gain a point?

You're basically talking about each skill storing both a "level" and "experience points". Muds that use skill decay would probably work that way, but other than that it's not the sort of change most people would notice.
08 Jul, 2011, Scandum wrote in the 10th comment:
Votes: 0
I use a cooldown after a skill gain of 5-15 minutes during which there is no skill gaining, with a relatively high chance of skill gain outside of cooldown so rare skills can be increased without too much trouble.

This quite effectively stops players from spamming for skill gains.
08 Jul, 2011, Runter wrote in the 11th comment:
Votes: 0
I plan to use a composite class system where you have 4 base classes: stealthy, strong, magical, and supportive. Names to be decided later. For the purpose of my examples warrior, thief, mage, cleric works. You get to pick 2 out of the 4 in addition to your specialization class. For a total of 3 classes that accumulate to build your full toolbox of abilities and skills. Then I have cross over skills potentially available in all of the base classes. The convention I'm using is 60-40-20-0.

A basic example might be second attack. The chance to swing your weapon twice. warrior-60%, thief-40%, cleric-20%, mage-0%. Then the skill is added together based on the bonus each class gave. warrior+thief = 100% chance to swing twice. warrior+mage = 80% chance and so forth.

A second example may be sneak. thief-60%, mage-40%, cleric-20%, war-0%. thief-mage(100% sneak) being the sneakiest yet potentially a loss to melee vs thief-war(60% sneak).

I also insist in my system of skill percentages representing a known mechanic. Like a base chance of activation. For example, I don't like 100% in a skill representing an unknown percentage of activation. Such, to me, reeks of hiding mechanics.

In addition to this I have skills which may be completely unique. In those cases only the class(es) the skill or spell is unique to contributes to its proficiency. Although, I'm big on cross-over so I'm more likely when designing skills to add stuff that may fall into the 60-40-20-0 system.

The skill percentage gains come at each level but it follows a formula so that the percentage gained is linearly interpolated with the number of levels you have left and the total value this base class offers. No practice sessions or anything of the sort.
08 Jul, 2011, Rarva.Riendf wrote in the 12th comment:
Votes: 0
Runter said:
I also insist in my system of skill percentages representing a known mechanic. Like a base chance of activation. For example, I don't like 100% in a skill representing an unknown percentage of activation. Such, to me, reeks of hiding mechanics.

Well having 100% change to swing twice does mean you will hit 100% of the time as well ? If not that reeeeks of hiding mechanics.

(I consider % are not a % of succeed, but more like a % of attempting it safely. If you have 100% dodge means you will attempt to dodge. Not that you will succeed at it.)
08 Jul, 2011, Runter wrote in the 13th comment:
Votes: 0
Rarva.Riendf said:
Runter said:
I also insist in my system of skill percentages representing a known mechanic. Like a base chance of activation. For example, I don't like 100% in a skill representing an unknown percentage of activation. Such, to me, reeks of hiding mechanics.

Well having 100% change to swing twice does mean you will hit 100% of the time as well ? If not that reeeeks of hiding mechanics.

(I consider % are not a % of succeed, but more like a % of attempting it safely. If you have 100% dodge means you will attempt to dodge. Not that you will succeed at it.)


It does not reek of hidden mechanics. If second attack is "swinging twice" then why would that mean you HIT twice? I use a specific mechanic that is not hidden to determine land/miss of swings. And there's a difference between *attempting to dodge and failing* and *not even attempting*. The same thing goes for swinging. There may be skills that activate on miss. It's a distinction with important difference.
08 Jul, 2011, Rarva.Riendf wrote in the 14th comment:
Votes: 0
Quote
It's a distinction with important difference.

You are saying you do not use any randomess in fight ? At all ?
08 Jul, 2011, Runter wrote in the 15th comment:
Votes: 0
Rarva.Riendf said:
Quote
It's a distinction with important difference.

You are saying you do not use any randomess in fight ? At all ?


I didn't say that at all, but players cannot make informed decisions about their character if they don't even know the weight of those decisions. I believe in telling players the exact correlation between statistics (like skill percent) and what they do.
08 Jul, 2011, Rarva.Riendf wrote in the 16th comment:
Votes: 0
Runter said:
I believe in telling players the exact correlation between statistics (like skill percent) and what they do.

Unless you give them the code, you just do like everyone else: 100% in a skill just means your chances are the highest possible.
08 Jul, 2011, Runter wrote in the 17th comment:
Votes: 0
Rarva.Riendf said:
Runter said:
I believe in telling players the exact correlation between statistics (like skill percent) and what they do.

Unless you give them the code, you just do like everyone else: 100% in a skill just means your chances are the highest possible.


Maybe that's how it is in your game. As I've said over and over that's not how I do it, and I advocate doing it differently.

100% in my game means you have a 100% chance of activation under whatever conditions activate it. As already mentioned, 100% second attack means 100% chance of an extra swing on melee. 100% dodge would mean 100% chance to activate the dodge skill when attacked by something that may be dodged. So I wouldn't give them 100% dodge. I'd give them 5% dodge perhaps. Why does everything arbitrarily need to be 100%?

And no, I don't give them the code to accomplish that.
09 Jul, 2011, Rarva.Riendf wrote in the 18th comment:
Votes: 0
Quote
As already mentioned, 100% second attack means 100% chance of an extra swing on melee.

Tell me if I finally get it: if you have a 5% chance of failing whatever your skill is, you show 95% to the player ?
It is a guenine question: I really do not get it. As without exposing the complete mechanics(not the code ok, but the algorithm) I do not see how you can do what you say you do.
09 Jul, 2011, quixadhal wrote in the 19th comment:
Votes: 0
Runter is making a distinction between correctly using a skill and having it be effective. You can know exactly how to open a lock and perform the operation perfectly, and still not succeed if your lock pick breaks in the process. Having 100% at dodge means you might always get to perform a dodge, but doesn't HAVE to mean you always succeed at it… if you see a giant semi-truck hurtling towards you, you may make your dodge roll and start moving out of the way, but you might not be fast enough.

This differs from how AD&D mechanics have worked, as there the dice roll is applied to the end result. You aren't rolling to be able to dodge, you're rolling to actually dodge the attack. As such, different situations apply different penalties to your attack roll. In the case of the semi-truck above, you may be rolling at a -15, meaning even if you had a perfect skill (a 20 would still work), you now have to roll a 5 or less because of the difficulty.
09 Jul, 2011, Rarva.Riendf wrote in the 20th comment:
Votes: 0
Ok, so if my game use URANGE(0,skills/2,100) > random_percent I should show 50% to the player so it not 'reeks of hiding mechanics' ?
0.0/39