In the Elthos game test we've discovered a well known and important
fact: Magic Spells are Powerful. They may, in fact, possibly, be
too powerful compared with the other character classes.

This of course has everything to do with the kinds of spells they can
cast and how those spells are structured. For example, at the moment
I have on my list the proverbial Sleep Spell. It is, indeed, quite
powerful. It allowed our party of adventurer's the ability to Sleep
three kobolds, a goblin, and in the next melee an Ogre, who was then
minced to pieces by the surviving characters... yes, before that
happened three of the Characters in the party had been killed by said
Ogre. It was a fast and brutal battle, which is an effect of
the "One Die System", and the fact that the party members and the
opposition were all roughly averaging second level. The entire
combat took a total of 8 melees (roughly 48 seconds), and took place
within a range of 60'. No missile weapons were used by the party
members, and Sleep was the only attack that had a decisive effect.
Sleep, as it currently stands, is a Range Spell, where the range is
10' x Level of the SpellChanter. In this case that gives Vespasian,
our fourth level Fighter SpellChanter, a range of 40'. The Sleep is
an Area Spell and will put down anyone (including friends) within the
forward facing range of the SpellChanter. He points his hands and
everyone within 40' of him falls asleep if the Sleep Spell works on
them. Each Character has a chance of saving vs the Spell based on
their Mystical Resistance. The SpellChanter, can, if he chooses,
add Mystic Power to the casting of the spell to do any of the
following (or combination thereof) - 1) Increase the range 2)
Increase the Attack Level 3) Increase the Effect (in this case they
would sleep longer).

The crafting of Magical Spells has to take into account the long term
implications of how powerful the spell will become as the
SpellChanter raises in levels. As the Character goes up in Levels
the Range increases, as does the Attack Level, and the Effect. This
needs to be balanced (to a certain degree) against the effectiveness
of the other Character Classes (yes, I know that in *some* RPGs they
have done away with classes and yes, I know that there are reasons
for this and that some people prefer classless games - however, this
game has classes and I'm not on a quest to change that). So
Fighters at 3rd Level should not be dwarfed by SpellChanters of 3rd
Level in terms of overall Power and ability.

The problem is that it is not easy to mathematically determine who
has the overall greater capacity. For example, what if the
SpellChanter has taken no offensive spells, but instead chose
Invisibility and Polymorph. Powerful indeed, but not in an
offensive combat sense. Conversely, SpellChanters are limited in the
total number of attacks they can make (they will run out of Mystic
Power after some number of attacks), whereas Fighters in the Elthos
ODS (One Die System) do not have that limitation and can potentially
continue dealing damage indefinitely (in the Elthos Prime system,
they would run out of Fatigue, but I'm not using that rule in the
ODS, although I could make a simplified version of it and add it to
the supplimental rules section, possibly). Also to be considered is
that Fighters can wear Armors that deflect or absorb damage, whereas
SpellChanters are limited in the armors they can wear. And so on.

So what I'll be working on is looking at each spell, both offensive
and defensive, and trying to determine how to evaluate the relative
power level of SpellChanters in comparison with the other classes.
It's a challenge.

Any thoughts on this would be welcome. Thanks.

- Mark

Quote from: VBWyrdeThe problem is that it is not easy to mathematically determine who
has the overall greater capacity. For example, what if the
SpellChanter has taken no offensive spells, but instead chose
Invisibility and Polymorph. Powerful indeed, but not in an
offensive combat sense. Conversely, SpellChanters are limited in the
total number of attacks they can make (they will run out of Mystic
Power after some number of attacks), whereas Fighters in the Elthos
ODS (One Die System) do not have that limitation and can potentially
continue dealing damage indefinitely (in the Elthos Prime system,
they would run out of Fatigue, but I'm not using that rule in the
ODS, although I could make a simplified version of it and add it to
the supplimental rules section, possibly). Also to be considered is
that Fighters can wear Armors that deflect or absorb damage, whereas
SpellChanters are limited in the armors they can wear. And so on.

So what I'll be working on is looking at each spell, both offensive
and defensive, and trying to determine how to evaluate the relative
power level of SpellChanters in comparison with the other classes.
It's a challenge.

Any thoughts on this would be welcome. Thanks.

- Mark

The mature technology is to write a Monte Carlo simulation.  That is not guaranteed to be an accurate reflection of gameplay.

A more elegant solution would be a Horne logic or other highly elegant logical means by which the rules could be proven to uphold a set of invariants before they were put into play.  Discovering such a system is not easy, and few mathematicians would even entertain the concept.

So ... I'd say you pretty much have to code it up as a discrete-event simulation.

Quote from: riprockThe mature technology is to write a Monte Carlo simulation.  That is not guaranteed to be an accurate reflection of gameplay.

A more elegant solution would be a Horne logic or other highly elegant logical means by which the rules could be proven to uphold a set of invariants before they were put into play.  Discovering such a system is not easy, and few mathematicians would even entertain the concept.

So ... I'd say you pretty much have to code it up as a discrete-event simulation.

Thanks for the comment.  That's pretty much what I thought.  I think attempting Monte Carlo or Horne Logic would be a bit more than I can manage at this point, to be honest.  So in this case, it seems that the objective of "Game Balance" is a bit of a lark.  I mean for any RPG where the we might claim "balanced design"... can we prove it?  I think the answer is, maybe, but it would be very difficult, and in a practical sense, not really.   So balance in RPG design must be considered an opinion, rather than a fact.   One could say, "this game feels balanced to me", and that would be a fair statement, but that's pretty much the best we can do.  

Given that, however, I don't think I want to abandon the balancing of my system, but I do want to keep in mind that any sort of balance is going to be a rough estimate, and based more on feel than percision.

Quote from: VBWyrdeThanks for the comment.  That's pretty much what I thought.  I think attempting Monte Carlo or Horne Logic would be a bit more than I can manage at this point, to be honest.  So in this case, it seems that the objective of "Game Balance" is a bit of a lark.  I mean for any RPG where the we might claim "balanced design"... can we prove it?  I think the answer is, maybe, but it would be very difficult, and in a practical sense, not really.   So balance in RPG design must be considered an opinion, rather than a fact.   One could say, "this game feels balanced to me", and that would be a fair statement, but that's pretty much the best we can do.  

Given that, however, I don't think I want to abandon the balancing of my system, but I do want to keep in mind that any sort of balance is going to be a rough estimate, and based more on feel than percision.

Bear in mind the difference between design and testing.

"Balanced design" means "the designer thought it looked balanced."

"Testing" means "we bribed two dozen adolescents with pizza, got them to playtest our game, and the results were so depressing we drank ourselves to sleep that night."

Now, if you can bribe *math students* with pizza and get them to rules-lawyer the game with their mad optimization skills -- then maybe you can combine the two.

Quote from: riprockBear in mind the difference between design and testing.

"Balanced design" means "the designer thought it looked balanced."

"Testing" means "we bribed two dozen adolescents with pizza, got them to playtest our game, and the results were so depressing we drank ourselves to sleep that night."

Now, if you can bribe *math students* with pizza and get them to rules-lawyer the game with their mad optimization skills -- then maybe you can combine the two.

WOW!  What a CONCEPT!   I *have to* try that!   !

Er... or maybe not.  The problem with a perfectly balanced system may be that it actually just might not be all that fun ... maybe... Or conversely, what Mathamaticians think is fun may not correspond to what my pizza devouring adolescent market thinks is fun.  Hmmm...

This kind of ties in to my Experience Calculations thread in a way.  For example, young kids may require that the rate of advancement in terms of levels be substantially higher, and at lower risk, than adult players.   They may also have vastly different expectations about what magic is, and what its for, and how it should be used, by whom, when and why.   Girls may have totally different expectations than boys.   Magic in particular is tricky that way.   And don't even get me started on Clericy!

Quote from: VBWyrdeWOW!  What a CONCEPT!   I *have to* try that!   !

Er... or maybe not.  The problem with a perfectly balanced system may be that it actually just might not be all that fun ... maybe... Or conversely, what Mathamaticians think is fun may not correspond to what my pizza devouring adolescent market thinks is fun.  Hmmm...

This kind of ties in to my Experience Calculations thread in a way.  For example, young kids may require that the rate of advancement in terms of levels be substantially higher, and at lower risk, than adult players.   They may also have vastly different expectations about what magic is, and what its for, and how it should be used, by whom, when and why.   Girls may have totally different expectations than boys.   Magic in particular is tricky that way.   And don't even get me started on Clericy!

Theoretically, I think pen-and-paper RPGs with very predictable GMs can be described as Markov Decision Processes.

http://en.wikipedia.org/wiki/Markov_Decision_Process

Given the unpredictability of human GMs, real-life games probably don't qualify as consistent MDPs, and a genuine MDP might only be found in a consistent, but boring computer game.

If you have any math geek friends, you might ask them if they know a way to use MDP analysis to test your rules.  Pen-and-paper games might be too loose, but the technique works on some software applications.  A trivial board game, like a dumbed-down version of Monopoly, ought to be easy to verify.

Anyhoo, I need to go read your Experience Calculation thread to see how this thread ties to it.

But you're definitely correct in that people have very different intuitions about what magic should be, what gods should be, what heroes should be.

Quote from: riprockTheoretically, I think pen-and-paper RPGs with very predictable GMs can be described as Markov Decision Processes.

http://en.wikipedia.org/wiki/Markov_Decision_Process

Given the unpredictability of human GMs, real-life games probably don't qualify as consistent MDPs, and a genuine MDP might only be found in a consistent, but boring computer game.

If you have any math geek friends, you might ask them if they know a way to use MDP analysis to test your rules.  Pen-and-paper games might be too loose, but the technique works on some software applications.  A trivial board game, like a dumbed-down version of Monopoly, ought to be easy to verify.

Anyhoo, I need to go read your Experience Calculation thread to see how this thread ties to it.

But you're definitely correct in that people have very different intuitions about what magic should be, what gods should be, what heroes should be.

Thanks for the thoughts!  

Well short of MDP, I still need a way to design a magic system that is reasonably well balanced.   What I notice as I play test is that the items which I rough drafted (spells) to be at certain levels don't quite achieve the desired effect in terms of balance.  I could try to adjust them but I think that as time goes on and new spells get added it will become harder and harder to feel reasonably sure that they are at the right level for their effects unless I come up with a system for creating spells that lets me determine the level it should be at mathematically in some way.

What I've come up with are some basic rules for spell building based on my study of my existing spell list (not much different from the standard D&D fare).

Breaking down my spells I have determined their components are:

Category
... Protection
... Detection
... Elemental
... Energy
... Transmigration
... Transmutation
... Necromancy
... Summoning
... Enchantment
... Illusion
... Shamanism
... Will

Geometry
1... Point (touch)
2... Line (bolt, beam, ray, person, object)
3... Circle
4... Square
5... Sphere
6... Shell/Shield
7... Cone
8... Cube
9... Area (can be amorphous, or organic shape)
10.. Wave (area in motion)

Range Units
1... Touch
2... Feet
3... Yard
4... Field (300 yards)
5... Mile
6... Region (20 miles)

Duration Units
1... Instant
2... Second
3... Melee (6 seconds)
4... Round (36 seconds)
5... Minute
6... Hour
7... Day
8... Month
9... Year
10.. Decade
11.. Century
12.. Millennium

Complexity/
.5.. Basic
1... Simple
2... Elaborate
2.5 Intricate
3... Byzantine

With these components I think I should be able to map out what Level each spell is and that will allow me to give spells out the Spell Chanters in a way that keeps them somewhat within bounds for their own level (ie - they will get spells that have a corresponding and appropriate power for their level).  

Now each Category of magic has levels, so I can't use that as a basis for determination of the spell level.   For example every category has a set of Level 1 spells.  I may be able to do this using a Multiplier system where take the value in the lists, multiply them and then assign the spells a level according to their value divided by 10 (truncated).

Fire bolt (Elemental)
Geometry: Line (2)
Range Units:  Feet (2)
Duration Units: Seconds (2)
Complexity: Simple (1)
= 2x2x2x1 = 8
Spell Level:  1

Polymorph Self (Transmutation)
Geometry: Person (2)
Range Unit: Touch (1)
Duration:  Day (7)
Complexity:  Complex (2)
= 2x1x7x2 = 28
Spell Level:  2

Sleep (Necromancy)
Geometry: Circle (3)
Range Units:  Feet (2)
Duration Units: Hour (6)
Complexity:  Simple (1)
= 3x2x6x1 = 36
Spell Level:  3

Lightening Bolt (Energy)
Geometry: Area (9)
Range: Feet (2)
Duration: Seconds (2)
Complexity: Elaborate (1.75)
= 9x2x2x1.75 =  61.2
Spell Level:  6

Seems to work.  I use Complexity as the multiplier which can fine tune the result at the end to get it to the Level it needs to be (ie - a short range powerful spell like Lightning bolt should *not* wind up being 3rd Level).  

Well, ... I'm only so-so happy with this system.   I think there may be a better way to do this.   My primary goal is to ensure that my spells aren't too powerful, or too weak for their Level, and that there is some coherency to that decision.  Of course the use of Complexity makes it arbitrary again.  But if I leave it out then I wind up with spells at levels that feel somewhat ... uncomfortable.   Maybe I should drop complexity and just live with the results.  Yup Lightning bolt is 3rd Level.   That's ok...?

What do you think?  Is there a better way to get the result?  Am I off in the ozone in terms of direction here?   Or does this seem to work out ok?Thoughts/Suggestions/Recommendations appreciated!   Thanks!  

Quote from: VBWyrde...
Well, ... I'm only so-so happy with this system.   I think there may be a better way to do this.   My primary goal is to ensure that my spells aren't too powerful, or too weak for their Level, and that there is some coherency to that decision.  Of course the use of Complexity makes it arbitrary again.  But if I leave it out then I wind up with spells at levels that feel somewhat ... uncomfortable.   Maybe I should drop complexity and just live with the results.  Yup Lightning bolt is 3rd Level.   That's ok...?

What do you think?  Is there a better way to get the result?  Am I off in the ozone in terms of direction here?   Or does this seem to work out ok?Thoughts/Suggestions/Recommendations appreciated!   Thanks!  

Draw on the power of distributed debugging.  Publish your complete system for calculating the requirements of spells.  Invite players and GMs to email you with their tweaks, complaints, and bug reports.  This is much cheaper than bribing math students to playtest your system, and it produces a similar volume of bug reports.

Quote from: riprockDraw on the power of distributed debugging.  Publish your complete system for calculating the requirements of spells.  Invite players and GMs to email you with their tweaks, complaints, and bug reports.  This is much cheaper than bribing math students to playtest your system, and it produces a similar volume of bug reports.

Not a bad idea.   I'd like to, of course, rough out a system that works for me in this regard before I release it to the wild.   So far I am finding that it seems to work reasonably well for creating new spells at appropriate levels.   At least it is providing me with a methodology, even if it isn't perfect.  

I'd certainly be curious to hear any feedback on the proposed system.  The purpose of it is simply to provide a means for determining what formulaic spells are appropriately assigned to specific levels for Spell Chanters to learn.