Creating my own system

[Immabozo]

So I was thinking about creating my own system and had some ideas and I would love to hear feedback.

EDIT, Genre, fantasy

HP are a set value for a given race, gained per age category (baby, toddler, teen, adult), plus con and then negatives for older age categories. This way, race is a very important choice, hp wise, and a main consideration in racial balance.

AC is mostly measured in miss chance and DR, with a numerical AC that must be rolled over, in order to deal damage. Skills and feats have a huge effect in defense, with feats like:

Parry, roll an attack versus your opponents attack roll, plus dex, plus weapon category (heavy weapons get a minus, light weapons get a plus), TWF get a second roll.

Riposte: After a successful parry, you may make a single attack at the enemy whose attack you parried, as long as they are within reach.

Attacks rolled 3D6 +str or dex (whichever is more advantageous)

A tank would be vital and "tanking" made possible with feats like:

"Arrogant Bastard" Whenever you successfully parry an attack or deal damage, you obnoxiously taunt an enemy to attack you. To break off attacks from you, if an enemy has an intelligence, they must succeed in a will save equal to you charisma modifier, cumulative for each time you taunt them in this way.

and "Taunt" if you speak a language your opponent understands, you can taunt them as a free action that you can only make once per each attack you have in a round, if they fail a will save (DC charisma based) they cannot attack a target other then you until the end of your next turn (if you die, they still attack your body until the end of what would-be your following turn).

Magic would be changed and fully restored to how I view how magic should work.

Mage, at level 1, chooses learned spell caster, or natural spell caster. The two magic systems work entirely differently.

Natural, may use only naturally known spells and requires a feat to get next-level spells (only eligible to be taken at a certain level) and does not incur a spell casting penalty for wearing light or medium armor. A spell has a fixed effectiveness per caster level.

A Learned spell caster progresses learning spells of a new level as soon as spells of that level are available to him (and taught to him, if only taught to himself, with a written scroll or other written source) and the power of such spells (and if they work at all) is based on a knowledge check in the magic's school. Generalist wizards will not be able to make knowledge checks as reliably or as high as specialist wizards. May not wear light or heavier armor.

On a 3D6, an auto-fail (like rolling a nat 1 in 3.5) will be if and of the 3 D6 come up "1" and auto success is if all 3 D6 come up "6". However, if you are trained in the task you are trying to accomplish (melee, ranged, casting a spell, disarming a trap, diplomacy, etc) whether from a feat, class feature, etc, an auto-fail is only if all 3 D6 come up "1" and an auto-hit/threatened crit (like rolling a nat "20" in 3.5) if any of the 3 D6 come up "6".

I know it is far from a finished idea and balance hasn't even started, but what do you think of this rough 1st draft of my idea?

EDIT: for the undoubted "The change to the 3D6 math is thrown off substantially" responses, If you are untrained, you are going to be more likely to fail, if you are skilled, you are going to be more likely to succeed. Also, another feat would be the countering of the other fighter/spell caster/whatever's spike in auto-success/crits, with making something like requiring an attacking melle/ranged/spell caster to roll 1 more "6" to crit/autohit you, or 1 less "1" to auto miss you.

Cause a highly skilled combatant vs a noob or a mildly skilled combatant, is a slaughter, unless the latter gets real lucky, but a highly skilled combatant vs a highly skilled combatant is on a whole 'nother level!

So a monk, with his unarmored, agile combat, might require only 1 "1" to auto miss him, while a heavily armored knight, with his magically enhanced defenses and shield, might require 3 "6"'s to auto hit/crit him, regardless of how that roll is normally lowered for you.

Making squishy caster abnormally so and heavy tanks remarkably tough

[The Rose Dragon]

This feels less like creating your own system and more like hacking D&D. Which, to be fair, can result in games quite distinct from D&D (M&M, for example), but in most cases, I would suggest starting with things like genre, scale and dice system, and piling stuff on those basics.

[Immabozo]

This feels less like creating your own system and more like hacking D&D. Which, to be fair, can result in games quite distinct from D&D (M&M, for example), but in most cases, I would suggest starting with things like genre, scale and dice system, and piling stuff on those basics.

I covered dice systems at the end, genre is fantasy, I guess I could add that at the top, but I do not get what you mean by scale

EDIT: also, I think some of my phrasing lends itself to that idea. But I do believe I have started where you suggest

[The Rose Dragon]

Scale means how powerful the PCs are supposed to be in comparison to the baseline of the setting (not the system). For example, Weapons of the Gods and Qin: the Warring States are both wuxia games. But in the former, an experienced warrior can easily fight a couple dozen fighters of the fifth rank, while in the latter, being outnumbered three-to-one is nearly always a risky proposition if your opponents know how to hold a sword.

[Immabozo]

Scale means how powerful the PCs are supposed to be in comparison to the baseline of the setting (not the system). For example, Weapons of the Gods and Qin: the Warring States are both wuxia games. But in the former, an experienced warrior can easily fight a couple dozen fighters of the fifth rank, while in the latter, being outnumbered three-to-one is nearly always a risky proposition if your opponents know how to hold a sword.

AH, I see what you mean. In mine, it would be more realistic, a highly skillled fighter CAN hold his own against several mid level fighters, but it would cause him to sweat. Thus the idea behind not gaining HP per level, but depending on race * age categories + con for HP number. High skill will give more ability to fight, cast, etc, while not making it impossible for lower level characters to hit you, relying on skill to minimize danger and armor is a very minor factor.

EDIT: not very minor, but more minor

[meschlum]

Math is fun!


Odds of automatic failure (untrained): 125 / 216, so somewhat less than 58.33%

Automatic success is 1 / 216, so less than 0.5%

Otherwise, you get an average roll of 12, with sigma^2 = 2.


Given that combat (via parry) appears to be opposed rolls, there are 6 dice involved, so we get a sigma of 2 (sqrt(2+2)) and can sort of assume a normal distribution.


What happens when an automatic success on an attack meets an automatic success on a parry?


So let's look at combat statistics.

Untrained and Trained fighters, one on one.

U attacks U

Attacker has roughly 40% chance of a successful attack, average 12 total (dice show 2 to 6, so average is 4 for each)

Defender has a 40% chance of a successful defense, average 12.

If attack and defense are successful, the attack beats the defense roughly half the time.

So the attacker connects with probability 40% * (60% + 40% / 2), or about a 1 / 3 chance.

If the attacker or defender have special weapons granting bonuses or penalties, this affects the 40% * 40% ~ 1 / 6 times that opposed rolls matter, shifting the odds from the default of about 1 / 12 of connecting (between 0 and 1 / 6 of hitting).

An attacker with a massive bonus goes from ~33% to ~40% of hitting, via a bonus of +4 or more.

A defender with a massive bonus goes from ~33% to ~25% of being hit, via a bonus of +4 or more.

Bonuses don't matter much, and there is a lot of ineffective flailing about. More to the point, you have a strong incentive to try aimed attacks: a -6 penalty to your hit roll if more than worth it if it doubles the damage you do.


T attacks T

Attacker has roughly 40% of a 'normal' attack, averaging 9, and 60% of a critical attack (dice show 1 to 5, so average is 3 for each).

Defender has roughly 40% of a 'normal' defense, averaging 9, and 60% of a critical defense.

If a critical defense blocks a critical attack, the attacker will hit with probability 60% * 40% + 40% * 40% / 2, so about 1 / 3, as before.

Weapon bonuses or penalties will again allow the odds to shift between 2% and 40%, with the same incentive to go for difficult aimed shots if possible. Essentially, there is no difference between untrained and trained people when they fight people with the same skill level.

If a critical defense does not block a critical attack, the attacker hits with probability 60% + 40% * 40% / 2, so about 68% (2 / 3) of the time. Bonuses shift this between 60% and 75%. Trained warriors wound each other much more often than untrained, and the incentive to go for aimed shots is even higher - a massive penalty to hit is worth it if it increases your damage by 25%.


T attacks U

Attacker has a critical attack ~60% of the time, and a normal attack otherwise, average of 9.

Defender can only block normal attacks, and has a normal block ~40% of the time.

When there is a normal attack and a normal block, the attacker averages 9 and the defender averages 12, with a sigma of 2. So the defender has a solid 80% or so probability of defending...

Chance of hitting is 60% + 40% * (60% + 40% * 20%), so 86%. With bonuses and penalties, it can range from 84% to 99%, though it is very difficult to raise over 92% or so.

Again, a strong incentive to go for aimed shots, and the notable irony that if neither party gets an automatic success / failure, the untrained mook is going to roll better than the trained warrior.


U attacks T

Chance of hitting is 40% * 40% * 80%, so about 13%. Weapon bonuses and penalties move this between 1% and 16%, though it's unlikely to get less than around 8%.

Aimed shots may not be advised in this specific case, as the odds of hitting can drop fairly fast with attack penalties - though the untrained attacker has an effective +3 bonus (12 average on a normal attack versus 9 average on a normal defense), so it depends on the damage boost granted.


Verdict

Except when an untrained person attack a trained foe, criticals are going to be the main driver for combat outcomes. Untrained versus trained gets a fairly large combat bonus when not getting an automatic failure or meeting an automatic parry, which also means criticals matter to the trained fighter.

Combat will be extremely random and risky if critical attacks give extra damage (since they represent more than half of all hits).

Combat between trained warriors will be very fast and lethal if critical parries do not block critical attacks (and it's always slow between untrained opponents).

Aimed shots, in the form of penalties to hit in exchange for increased damage, will be used by everyone (except, maybe, untrained grunts attacking a trained foe), since criticals guarantee they'll connect.


Bonuses of +/- 4 are overwhelming when criticals do not occur, so you need to give small modifiers for strength, dexterity, and weapons or they'll swamp the non-critical outcomes.

Strength is likely to be a god stat, as it helps to hit (minor, thanks to criticals) and helps damage (major, since common criticals mean you hit a lot).



And that's what the core engine of your combat system is like. You can change parts of it by giving more options, die shifts, and so forth - but the fundamentals are there.

[Immabozo]

Verdict

Except when an untrained person attack a trained foe, criticals are going to be the main driver for combat outcomes. Untrained versus trained gets a fairly large combat bonus when not getting an automatic failure or meeting an automatic parry, which also means criticals matter to the trained fighter.

Combat will be extremely random and risky if critical attacks give extra damage (since they represent more than half of all hits).

Combat between trained warriors will be very fast and lethal if critical parries do not block critical attacks (and it's always slow between untrained opponents).

Aimed shots, in the form of penalties to hit in exchange for increased damage, will be used by everyone (except, maybe, untrained grunts attacking a trained foe), since criticals guarantee they'll connect.


Bonuses of +/- 4 are overwhelming when criticals do not occur, so you need to give small modifiers for strength, dexterity, and weapons or they'll swamp the non-critical outcomes.

Strength is likely to be a god stat, as it helps to hit (minor, thanks to criticals) and helps damage (major, since common criticals mean you hit a lot).



And that's what the core engine of your combat system is like. You can change parts of it by giving more options, die shifts, and so forth - but the fundamentals are there.

I'm not going to pretend to understand what just happened, but I'll take your verdict into account.

I'll try to fix the problems you suggest tomorrow, since my girlfriend is just about to get here

[Mr.Bookworm]

I'd advise you to read this essay, which is about D&D take-offs (okay, '90s D&D take-offs in specific, but it's applicable to this). (http://www.indie-rpgs.com/articles/9/)

The first two parts in particular seem very relevant to what you're trying to do here.

EDIT: I'm not telling you to not create your own system, mind you. I just think that when creating an RPG system, you just always need to have in mind what you're trying to do with the system. Do you want to create a system that supports gritty, granular dungeon crawls? One that supports high magic campaigns where everyone is decked out in full Christmas tree regalia? A wuxia-inspired game where fighters float from tree to tree dueling?

Whatever you're trying to do, your mechanics should be focused towards that goal and support it during gameplay. The gritty game doesn't need twenty million magic items statted out, the high magic game doesn't need a long list of mundane weapons, and the wuxia game doesn't need rules for knights wearing plate armor.

"D&D, but better" isn't really any sort of coherent statement to make with a game.

[Immabozo]

I'd advise you to read this essay, which is about D&D take-offs (okay, '90s D&D take-offs in specific, but it's applicable to this). (http://www.indie-rpgs.com/articles/9/)

The first two parts in particular seem very relevant to what you're trying to do here.

EDIT: I'm not telling you to not create your own system, mind you. I just think that when creating an RPG system, you just always need to have in mind what you're trying to do with the system. Do you want to create a system that supports gritty, granular dungeon crawls? One that supports high magic campaigns where everyone is decked out in full Christmas tree regalia? A wuxia-inspired game where fighters float from tree to tree dueling?

Whatever you're trying to do, your mechanics should be focused towards that goal and support it during gameplay. The gritty game doesn't need twenty million magic items statted out, the high magic game doesn't need a long list of mundane weapons, and the wuxia game doesn't need rules for knights wearing plate armor.

"D&D, but better" isn't really any sort of coherent statement to make with a game.

Fantastic advice, thank you

[erikun]

I'm not going to pretend to understand what just happened, but I'll take your verdict into account.

I'll try to fix the problems you suggest tomorrow, since my girlfriend is just about to get here
Remember, though, that there is around a 50% chance of auto-success or auto-failure on any single roll, meaning that the mathematical results don't matter.

In the Trained attacks Untrained example, it was stated that the defender has a 80% chance of successfully defending... but that assumes cases where the roll matters. In truth, the defender will auto-fail around 50% of the time and the attacker will auto-succeed 50% of the time, meaning the 4-in-5 chance of successfully defending only applies to the other 25% of the time.

Your defender successfully defends around 20% of the time overall, the attacker rolls higher 5% of the time, and the remaining 75% of the time the defender either auto-fails or the attacker auto-succeeds (or both).


And yes, I'd like to second the advise Mr. Bookworm provided. Consider what you want the new system to do, and make sure your changes work towards that goal.

[Immabozo]

Remember, though, that there is around a 50% chance of auto-success or auto-failure on any single roll, meaning that the mathematical results don't matter.

In the Trained attacks Untrained example, it was stated that the defender has a 80% chance of successfully defending... but that assumes cases where the roll matters. In truth, the defender will auto-fail around 50% of the time and the attacker will auto-succeed 50% of the time, meaning the 4-in-5 chance of successfully defending only applies to the other 25% of the time.

Your defender successfully defends around 20% of the time overall, the attacker rolls higher 5% of the time, and the remaining 75% of the time the defender either auto-fails or the attacker auto-succeeds (or both).


And yes, I'd like to second the advise Mr. Bookworm provided. Consider what you want the new system to do, and make sure your changes work towards that goal.

I definitely see what you are saying and am trying to figure it out.

[kyoryu]

And yes, I'd like to second the advise Mr. Bookworm provided. Consider what you want the new system to do, and make sure your changes work towards that goal.

Totally agreed with that. If you know what you want the system to do, then you'll know when you've hit that mark. It gives you something to validate against.

One thing that I often find helpful in situations like this is to just sit down and write a fake transcript of play - don't include any mechanics, though, just what the players do and what *types* of decisions they make, and maybe what's going on in their heads a little bit. That'll at least give you a good idea of the kinds of decisions players need to make, and what elements should be in the game.

And that's another point - while the math in a system may be interesting, what's more interesting is how the players interact with the system and each other. The fake transcript can also help you zero in on these types of things as well, since you'd omit the math anyway.

[Immabozo]

Totally agreed with that. If you know what you want the system to do, then you'll know when you've hit that mark. It gives you something to validate against.

One thing that I often find helpful in situations like this is to just sit down and write a fake transcript of play - don't include any mechanics, though, just what the players do and what *types* of decisions they make, and maybe what's going on in their heads a little bit. That'll at least give you a good idea of the kinds of decisions players need to make, and what elements should be in the game.

And that's another point - while the math in a system may be interesting, what's more interesting is how the players interact with the system and each other. The fake transcript can also help you zero in on these types of things as well, since you'd omit the math anyway.

I wouldn't have thought of it that way, thank you