Discussion of skills for the XNA MUD project.

My thought was that skills should be somewhat broad, rather than specific. A character wouldn't need to learn how to use a specific weapon, rather a category of weapon, Sword, Bow (although this would probably break down into regular bow and crossbow since they're so different), Dagger, etc. This would make it much easier to code and play, as I would rather target a more non-hardcore RPG player.

Spellcraft would be a required skill for magic-using characters.

Other skills:

Hide
Sneak
Lockpick
Disarm Trap
Perception - gives a better chance of noticing hidden object, traps, etc.
Armor Use - how to fight while wearing armor
Dodge - passive skill

_________________
Bored? Head on over to my blog and see what I'm up to.

Microsoft XNA MVP

How about passive defensive skills
IE:
Block/Parry
Dodge

have Armor skill determine what types of armor the player can use.
IE:
1 -> 5 - Leather
6 -> 10 - Studded Leather
11 -> 15 - Scale

Yeah, I like this idea, the only problem is we want people to actively use skills to level up. So, I don't think this will work.

_________________
In brightest day, in blackest night. No evil shall escape my sight. Let those who worship evil's might, beware my power... Green Lantern's light!
Twitter!

Dodge could be a good passive skill. It would make people decide where to use points for skills - take away from offensive skills to keep from getting hit or put it in offensive skills to try to kill mobs before getting damaged.

With regard to armor, any armor should be able to be worn but if you don't have enough skill in wearing armor your actions will suffer.

_________________
Bored? Head on over to my blog and see what I'm up to.

Microsoft XNA MVP

Yeah, any skill that requires much dexterity would be penalized if you're wearing armor that is too bulky for you. For example, if you use magic (which may require physical incantation), the armor penalty might cause your spells to fail xx% of the time. Having skills like "Heavy Armor" would help alleviate that.

Also, would each weapon type (i.e. blade, blunt, pierce, exotic, etc.) have its own skill?

_________________
What most people don't understand about "enlightenment" is that it is not an end-goal; but where you find yourself just before taking a new "first step."

Why not have spot/search? Spot for objects that might be in plain view, but not easily seen (which would make it a passive skill), and Search for hidden objects/traps/doors/etc. (a manual skill).

Or would Perception lump these two skills together?

_________________
What most people don't understand about "enlightenment" is that it is not an end-goal; but where you find yourself just before taking a new "first step."

Perception could easily lump both together. It could always be active in a 'passive' mode, for obvious objects (or at obvious to a person who is [insert skill level here] perceptive). You could then perform searches, which would flip it to an 'active' state to allow for finding hidden things. This would also be when using the skill levels it up.

Having an active and a passive state for skills may work for other skills as well... it could even work for armour in a way. The armour skill is 'passive' when you are wearing armour that is comfortable (your armour level is lower than your skill level) and 'active' when you are wearing armour that is uncomfortable... I see a bell curve for the peak skill increase there though... a level or 2 higher and it's only slightly uncomfortable and you won't level up very quickly... a few more levels and you will learn quicker as it's pretty damn uncomfortable... and then it drops off sharply because your armour is just too bulky and uncomfortable. Penalise for wearing armour above your level as well, and you have a way of training up armour skills (in a way).

_________________
God must love stupid people, he made so many.
theraje: 'God doesn't love stupid people, they're just easier to make'
http://sharedillusions.blogspot.com

This would be on a case by case basic. Depending on the type of sword, it case be used as for slicing, piercing, or even blunt (say you want to knock an opponent out without injuring him). Right now I'm thinking combat skills should be as few as makes sense. It also depends on what weapons we have. If there's no flail in the game there's no need for a skill to handle it.

As for perception, what work said is exactly what I was thinking. For searching you'd have a bigger bonus if you were actively searching, but a person with a higher Perception would have a better chance of noticing something while just walking along than a person with a lower Perception. Both would still have a chance though.

_________________
Bored? Head on over to my blog and see what I'm up to.

Microsoft XNA MVP

My own ideas here:

skills are related to weapon types like this:

melee (possible subclasses: sword, mace, stave, etc)
ranged (archery, throwing knives, slingshot, etc)

you could have each subclass with it's own skill, or a simple dichotomy melee and ranged skills.

Each weapon may have two or three ways of using it. Each way creates a different kind and amount of damage. eg,
-- a sword can be swung generating 20 points of normal damage,
-- a sword can be slashed, generating 40 points of slash damage,
-- a sword can be lunged, generating 40 points of piercing damage,
-- a sword can parry a melee attack, absorbing some % and reflecting some %.

Damage and armor types should pair up on some sort of table. ie, Light armor has best defence against piercing damage (pointy fast projected impacts). Normal armor has best defence against normal damage (blunt heavy impacts). Heavy armor has best defence against slash damage (sharp cutting impacts). Siege damage (explosive or large scale) impacts all armor equally, while armor offers no defence against magic attacks.

eg,



what these figures represent is the effective armor.

So if I have 14 units of heavy armor, and get hit with a lunged sword (piercing) for 50 damage points, then the armor is effectively 50% = 7 units of effective armor against this weapon's damage.


Then we calculate damage taken = function(50 damage points, 7 armor)

I'd probably suggest a more DnD-esque split of weapon skills myself... Have broad classes for swords, axes, maces, etc. and then have specific skills for the more exotic weapons (specific skill for shurikens if we have them, for example).

_________________
God must love stupid people, he made so many.
theraje: 'God doesn't love stupid people, they're just easier to make'
http://sharedillusions.blogspot.com

I don't understand the logic behind this. Heavier armor would provide better defense against all types of attacks. My take is that the more armor you have the less damage you take from any attack.



This was my original idea and I think probably the best way to go that offers a bit of flexibility and ease of development.

_________________
Bored? Head on over to my blog and see what I'm up to.

Microsoft XNA MVP

Whatever my opinion counts for, I don't think that skill points are the most natural way to advance a character. I feel that weapon skills should be automatically gained through use of a weapon, rather than through arbitrarily assignable points, else like we say, you could advance a skill without ever practicing it.

I feel that customization should be through non-weapon skills, like which spells to learn for a mage or cleric, or special attack or defense moves for the fighter class: All things which cost mana (or equivalent).



The logic is that armors are paired against weapon attack types for maximum effectiveness. So that you can change your weapon attack method (or weapon) and armored gear to suit your combat situation. But it is still true that the more armor, the better defense you have.

Without biasing like this, there is no logic to having different classes of weapon and different classes of armor.

iirc, in dungeon master, there was two special armor/weapon classes: lyte and darc.

Both were associated with very powerful creeps. Lyte armor was effective protection against lyte attacks, but lyte armor was relatively ineffective against darc attacks, and vice versa.

So if you were in lyte gear and came up against a creep which made darc attacks, you'd have to change your armor if you wanted maximum protection.

Like in real life, some armors are better in some situations. Other armors are better in other situations. Chain mail will protect you from the blade of a sword, but an arrow will split the chain open and penetrate. Leather will generally be better at absorbing an arrow than chain mail, but can be cut through easily by a sharp slashing blade.





this would relates to this part of the table

I agree, but the code necessary to implement a true skill-based advancement system would be more difficult than a point-based skill advancement system. This can be somewhat helped by only allowing points to be put into skills that the character already has. To learn a new skill would require a trip to a trainer in a town and training would cost money, not points.


This would mean that characters would have to carry around a lot of different sets of armor and many weapons, which means a bag of holding inventory system, which I hate (although I understand why it's done). A character shouldn't be able to carry an entire dungeon's worth of loot around. Limiting what can be carried would mean slower progress in advancing the character monetarily, which I like.


The bias can come from class/skill costs and armor penalties for different actions. Fighter types would naturally learn to use armor easier than mages.


Not really:


I see chain as an addition to lighter armor, not instead of.



Good discussion so far.

_________________
Bored? Head on over to my blog and see what I'm up to.

Microsoft XNA MVP

Yes, but we can exaggerate reality a little. If the progression of armors is too linear and straight forward, I'm not sure it would be as fun. If armor B is always better than armor A -->

Besides, we can still use the armor values to make some armor classes better than others. So a typical light chain vest may be 20 armor, and equivalent light leather vest may be only 10 armor. So we can still have that chain is usually better, but there may be some situations where leather is equal or better.

Like vs piercing weapon:
light chain vest (20 normal) becomes 10 effective.
light leather vest (10 normal) becomes 10 effective.

vs blunt weapon:
light chain vest (20 normal) becomes 20 effective.
light leather vest (10 normal) becomes 7 effective.

So the front line units generally will take the normal armor because they take the (normal) melee damage. While those at the back of the party can take the light armors since they will only ever take damage from ranged (piercing) attacks (unless they're dungeon noobs)

It isn't always the case that many armors sets need carrying, but shuffling units around to change who is tanking and who isn't, or shuffling armor sets between different players. That would be the emergent game play.

Better in what way? Heavier armor provides more protection but has its penalities - it requires more skill to wear and maintain combat effectivity.


It sounds like you're thinking MMO-style, which this isn't. The issue the player is going to have to think about is - heavier armor means putting more points into armor skill which means less points for other things so do you get better armor or increase other skills. I'd rather have the armor calculation be a simple damage reduction against all attacks rather than having to create a table for armor/attack type combos.

_________________
Bored? Head on over to my blog and see what I'm up to.

Microsoft XNA MVP

Actually I'm thinking dungeon master style, but a similar format is used in some mmos, yes.



It's easy to do.
Each time a weapon class is used, do a probability roll P to see if the skill is developed. Once X development points are earned, you gain a skill. This moderates the rate of learning skills: Absolute minimum = X uses of weapon. Expected number = X/P uses of weapon. Standard deviation will be reasonably tight, as long as P isn't too small. Say, if P is around 5%, it should works pretty well at that.

Alternatively, in WoW, I saw they used a 5 skill points per level, per weapon class. That is, at level 4, a warrior could develop their sword skill upto 20. Each weapon use had a probability roll P to see if the skill is developed. So after 5/P uses of the weapon, weapon skill was expected to be maxed out for that level.

It's still easier to just allow the player to configure his character. I'm fine with not allow the character to learn a skill through assignment of skill points. We would just have trainers in the town level that the player would have to visit to learn a new skill. After that he can assign points to it.

_________________
Bored? Head on over to my blog and see what I'm up to.

Microsoft XNA MVP

Another important thing we'll need to decide is the relative power of consecutive power levels.

Like Level 20 fighter vs Level 21 fighter. How much difference is there between them?

This is relevant to character stats and also for the stats of the monsters, the power of loot, and how much loot we need (because if it's rated at -2 levels it is useless?), and also spell powers and how many power levels each of those is replicated into (1->6 or 1->12 for example).

• On one hand there are duals 1 level X vs 1 level Y, (given that Y>=X) and the outcome we would like for the conflict.
  
• Secondly, there is extrapolating the relationship across 100 levels which could inadvertently create ridiculously large character stats.
  
• Thirdly, there are biased conflicts (eg, 3v1) and what level difference should the two parties have for a balanced confict (unpredictable winner).
  
  eg, should 3x lvl 10 warriors be able to win if they team up against 1x lvl 20 warrior?
  

It isn't straight forward. Not something we can just guess at. But I'll talk more about this tomorrow.

There shouldn't be much difference from one level to another. It should take a bunch of levels to really feel the difference. It also depends on character configuration.


Yes.

_________________
Bored? Head on over to my blog and see what I'm up to.

Microsoft XNA MVP

3x lvl 10 warriors have a 50:50 chance of winning if they team up against 1x lvl N warrior. What is N?