Combat system theory

[housellama]

[Warning: Potentially crunchy material contained within]

I need some help. I'm looking to create a new system, from the ground up. I've tossed around a lot of ideas, and I've got a good handle on setting and flavor and a lot of other things, but I can't go much further in the design process until I get a system. I'm coming at this backwards from most. Most of the time, you have a system and are building a world around it. I've got a world, and need to build a system underneath it.

Very little is set in stone. I haven't decided on an attribute based system vs a skill based system. I haven't decided on basic stats. Nothing number related is nailed down yet. There are a few things that are solid though, and that's because the system is going to be modeled to resemble combat in an MMORPG.

The easy bit is that it's going to be semi-realtime. It's going to run on a count. Attacks, movements, abilities, everything is going to be 'timed', and have a 'cooldown'. It takes X ticks and can be used again in so long. Everything is tied to that count. A character's general stats/durability is going to be tied to level, but most abilities are going to be set in stone. This ability does X+Y% of stat A (or skill, depending) and takes Z number of ticks.

The hard bit is figuring out how the math of combat should actually work. Do I want to use a D20 like system where Attributes govern derived stats that rule combat? Or do I want to go with a skill based system like the Roll/Keep system from Legend of the Five Rings? That gives finer control, but can be more complicated...

I can't really develop character stats until I know how I want to come at the combat system. In this case, it's easier to model the stats around the math than try to shoehorn the math into some made up stats. Until I know how the math works, I don't know how hard things will hit, so I can't give hitpoints. Since I don't have hitpoints, I can't even begin to think about a level table. Since I can't do a level table, planning development in anything other than the abstract is impossible. So I'm at an impasse. I need to know how I'm going to handle combat (and thus stat my characters) before I can really continue on this part of the game development.

So like I said, nothing is nailed down. Any suggestions are welcome. Use whatever tools you can suggest. Any stats, any skills, any dice, any random number generation, whatever. Please stay away from using published mechanics however. I need new, not existing. Inspiration is good. 'Borrowing' is not. Let's build something brand new.

This is a blank slate, ladies and gentlemen. This is your chance to throw whatever you want at the wall. Let's all see what sticks. The only thing I ask is that you keep the feel of the game in mind. If you have played WoW or similar games, that is the feel I am trying to reproduce with this combat system.

What I need help with is the math of combat and development to keep the players alive but still threatened by the things around them. I'm a game player, and I've designed a lot of campaigns, and even a lot of mechanics, but I'm very new at system development on this level. I need help, and this is the first place I thought of to come to.

Feel free to ask me any questions you want. I will be glad to answer what I can, either here or privately. Thank you for your time, ladies and gentlemen.

-Housellama

[rayne_dragon]

Well, I presume that like most game systems you're going to want to include a certain amount of randomness. If you're building a system up from scratch, it gives you the chance to decide how random you want the system to be. The larger the size of the die rolled the greater the randomness. If a bunch of dice are rolled you get more results closer to average and less at the extremes. Higher static modifiers decrease the randomness of results.

Generally, at some point, you'll have to start assigning arbitrary numbers. Making your first arbitrary numbers the number of dice and the number of sides on said dice allows you to determine how much randomness to include in the game, allowing you to flesh out what kind of static modifiers you want to add to the roll. Or just affect the success of the roll.

So, to start with, I'd say you might want to figure out what kind(s) of dice you want to roll for this game.

[housellama]

That is a good place to start. I've also worked out a few other things. Things I should have seen earlier, but sometimes you get so locked into a problem that you can't see the obvious.

I'm using a freeform class system, so players are free to change their classes more or less at anytime during the game (that they have time and access to the appropriate facilities, of course), and can have elements of multiple classes. They gain access to more skill levels gradually through experience. However, there is also a separate leveling system, and I want to make being a higher level MEAN something.

That means I'm going to have to tie SOMETHING to level, and the easiest thing is to tie stats to level. So stats will have an effect on hitpoints, and modify combat. But I want class to mean something too, since there are combat and non-combat classes.

So a hybrid system is what I'm going to go with. Stats either act as a base and skills modify them, or vice versa. Lets start with d10s. For now, let's assume stats are the base and skills modify that. Stats: Attack, 20 Defense 10, vs Attack 10 Defense 20. Weapon and defense skills 20 for both.

Stat + (Skill/10)D10. Under means a miss. Over is damage taken.

I'll run the numbers in a moment. Work is getting insane. But it's a starting place.

-Housellama

[housellama]

Okay, that didn't work. I need to look at this in more depth when work isn't absolutely insane.

-Housellama

[Wannabehero]

People in general prefer to deal with smaller numbers than larger, yet larger number ranges allow the game to be designed with greater precision in challenges and random chances.

This is why the d20 system is more successful than most. The d20 die roll lets random chances be determined in 5% increments, which allows a fairly high range of random chance to be handle in a number scale that is easily coped with by players (addition and subtraction in the 1 to 20 range is comfortable for quick mental arithmetic).

d10 systems are easier to compute on the fly by players, but you literally only have half as much space to move random chance around (because you are dealing with 10% increments of chance). d100 systems offer a huge range to fine tune random chances, but addition and subtraction in this large number range is, if never consciously understood to be, more difficult for players and tends to attract them less.

multi dice systems present the issue of probability distributions for roll results, making scaling random chances non-linear and more difficult for players to conceptuallize (in general, I personally recommend steering clear of using a multi dice system).

In summation: I recommend you pick your dice system first, and I also recommend using a d10 or preferably d20 system.

____________

If you are seeking to make a pen&paper RPG system based of of MMORPGs, I also recommend looking at some pen-and-paper MMORPG game ports, such as the WoW RPG, or D&D 4th ed. (seriously, it is very much like a p&p MMORPG)

[erikun]

The easy bit is that it's going to be semi-realtime. It's going to run on a count. Attacks, movements, abilities, everything is going to be 'timed', and have a 'cooldown'. It takes X ticks and can be used again in so long. Everything is tied to that count. A character's general stats/durability is going to be tied to level, but most abilities are going to be set in stone.

This seems like it could borrow inspiration from the weapon timing rules in AD&D 2nd edition. The way the system worked was that you rolled initiative, then added a number based on the weapon (or spell, for spellcasting) to determine turn order. The lowest number goes first, so a dagger than has a weapon speed of 1 would be far more likely to go before a poleaxe with a weapon speed of 8.

You could also use a "wheel" initiative that I've seen in various systems. That is, rather than having an initiative count from 0->infinity telling everyone what order they take actions in and then moving onto the next round, the wheel has initiative between 0 and 9, with the next step after 9 being 0 once again. Thus, deciding to delay or taking actions which are not instantaneous (frequently spellcasting) moves you along the wheel as initiative progresses, and your new initiative changes to when the delay/casting stops.

You would need to consider the magnitude of the numbers you wish to use. 0-9 works well, although it assumes that all actions will have a relatively small "weapon speed" - D&D rolled initiative every round, so a weapon speed of 8 wouldn't be considered crippling beyond attacking second all the time. Larger numbers would allow for more granularity - most swords and similar weapns wouldn't be stuck at speed 3 to 4 - but would require inflated numbers to use them.

You might want to consider how much faster, relatively speaking, you want specific weapons to each other. For example, is stabbing with a dagger twice as fast as swinging a hammer? Three times? If you get some idea of the ratios, it will be easier to determine what the numbers will be. (Note that a "dagger swing" isn't just the time to move left-to-right, but also find an opening and thrust properly.)

[BarroomBard]

Exalted also uses the circular tick system. It's handy, visually easy to comprehend, and works well. With that system, the wheel had values from 0-9 and no action took more than 8 ticks.

----

As far as dice systems, the most important thing to think about is what feel you want your game to have. If you want a more heroic, sword-swinging high adventure, you want to look to d20. If you want a more realistic game, you want to use 3d6 or so. If you use more dice, it gives you a bell curve, which tends to normalize results, rather than lead to extraordinary, heroic efforts.

Deciding what you want your game to feel like is also important when deciding on your stats. If you want the characters to always fear death, low hit points to high damage is what you want. If you want to model MMO's, you want a system that works with a very high percentile of hits to misses (how often do you really miss in MMO's or Final Fantasy?). In this system, damage is more highly variable than accuracy. Most tabletop games go the opposite route, mostly because you have less math to deal with.

The most traditional set up for table top games is Stat+Skill, and unless you have a really good idea, it's best not to mess with that formula. An important consideration when assigning numbers to your stats and/or skills is granularity. How much difference do you want between one number and the next number? In a percentile system, for example, there is not much difference between needing a 67 to hit and needing a 66 (only 1% difference). If you have a range from 1-20, however, the difference between one number and the next is a 5% change in success. If you have a range of 1-6, there is a huge difference between one number and the next, etc.

More granular systems (i.e., a larger range of values) tends to more mathematically complex and more realistic systems. Less granular systems tend toward being simpler in play and more abstract.

These are all very important considerations when translating your fluff into mechanics.

[erikun]

I wouldn't think the dice system needs to be overly thought out at this point, beyond the flat/curve distinction. The only difference between d% and, say, d8 is the size of the numbers involved.

You can also consider if you want a roll-over (like D&D) or a roll-under system. Roll-under systems may seem strange, but it is much easier to see your chances at a glance. In roll-over systems, you generally don't want your bonuses to be higher than the dice you are using.


Perhaps a better question at this point is: What do you want your system to cover? Just combat? Combat and social issues? War games? Political maneuvering? There is a lot more that a system can handle beyond hitting orcs, after all... assuming you design it to do so.

[housellama]

Having done some further research and reevaluation of what I've got, I've come to some conclusions.

First, since I'm modeling after an MMO, might as well use their system. For technical reasons, they keep the calculations made by the system to a minimum by using a table based system. Start with a base miss chance and modify it with attributes and skills from there. Take that number, roll a percentile and consult the table. That covers fumble, miss, hit and crit all in one go. That makes combat fairly simple and frees me up from being tied to a universal system. The hard part becomes making the table, but that's just a matter of brute force trial and error more than anything else.

I, and the players I am catering to, have always wanted to take apart one of those things. Or more precisely, put one together. You could call this a trial run. Easier to do something on paper and trial it here before writing and rewriting potentially thousands of lines of code... If the project ever gets that far.

The system will involve more that hitting things yes, BUT... I can deal with those things easily enough because they are not tied to level. Those are all strictly class and skill based. However, even though I'm aiming to make a very freeform system, I want level to MEAN something. If you've put the time and energy to make it to level 50, I want you to feel powerful. I want you to feel like something of a hard man, even if you're a non-combat character. That means that I have to tie level to combat somehow, and that's where my problems come in. I'm not sure how to balance that.

What I DO know is that while I want stats (which is what is tied to level) to mean something, I want skill to mean a lot more. A level 50 non-combat character should be able to kick the butt of level 30 monsters, but should still want a tank to hide behind going up against something near his own level.

I really do appreciate all the suggestions! I'm learning a lot during this project. Mostly from the ideas that end up in the trash, but that's generally how things go.

-Housellama

[DragonOfUndeath]

+2 Damage and To Hit every level? As well as increasing HP it would make levels important for combat. Ability increases would work well every couple of levels too.

[erikun]

First, since I'm modeling after an MMO, might as well use their system. For technical reasons, they keep the calculations made by the system to a minimum by using a table based system.

Well, good luck with that. Computers are great for running rediculously complex calculations in the background, and can easily handle tables with three (or more) axes. Neither of these are easily translated onto paper.

One nice thing about tables is that they can create a curve while still rolling a flat die. That is, you could roll 1d20 and consult the table, but the success-fail spread can make success more likely for more skilled characters.

What I DO know is that while I want stats (which is what is tied to level) to mean something, I want skill to mean a lot more. A level 50 non-combat character should be able to kick the butt of level 30 monsters, but should still want a tank to hide behind going up against something near his own level.

If you are tying stats to level, then perhaps use Stat-v-AC for determining success? For example, if a 50th level character has a stat cap of 50 and gains stat points each level, then you'd expect a level 50 fighter to have 50 STR. You can then just compare how well 50 STR hits a certain armor value (regardless of level), rather than how well a 50th level character of a specific strength value fares.

Then again, if you're taking apart a specific game system and trying to reproduce it on paper, perhaps my recommendations are rather pointless.

[housellama]

Well, good luck with that. Computers are great for running rediculously complex calculations in the background, and can easily handle tables with three (or more) axes. Neither of these are easily translated onto paper.
...
Then again, if you're taking apart a specific game system and trying to reproduce it on paper, perhaps my recommendations are rather pointless.

I'm not taking apart a specific game system. I'm looking at how those kinds of systems work, and attempting to reproduce it on paper.

Actually, the calculations is the point of using a simple table system. In an MMO, the computer is already keeping track of hundreds of thousands of different things at the same time. If it had to track multiple calculations for each hit for each player, that's even more calculations, and frankly the servers are already being pushed pretty hard.

Most of them have a base miss chance that's modified by one precalculated number on each character. Then it rolls a percentage and consults one table that remains the same for EVERYTHING, so it's loaded at the very beginning and stays in memory. It's similar to the THAC0 table. (Okay, what's your THAC0. okay, what's your AC. Okay roll. *check* Hit. *repeat* Miss. Etc.) The math to generate the table is a lot of statistics trial and error and hard core math and balancing. The part I need now is what goes into the number that modifies the rolls on the table.

Ability increases would work, because I plan to derive HP and attack stats from Abilities. Gives more freedom that way. I can increase the stat directly, or increase just the derived stat.

[Unrest]

Well, good luck with that. Computers are great for running rediculously complex calculations in the background, and can easily handle tables with three (or more) axes. Neither of these are easily translated onto paper.

How does that look like? Did you just mean 'with more columns'? Because I've been wondering whether there is something like a table with more than juz an X and Y, you know, like three dimensions on one table.

As for the topic itself, just one piece of (quite obvious perhaps) advice: first, go with whatever idea you have that sounds reasonable. Once you have a framework, start thinking up dozens of contexts ("what about touch attacks", "what if there's X in effect", "what about the possibility to do Y", "will this scale properly from level 1 all the way to 255" and son), and only then modify accordingly, trimming, cutting and scrapping without resentment. I find this way to be the best when devising something from scratch. You enter the labirynth many times, and it will take time, but you're not wasting all your time trying to think the whole path towards the center in advance - you just see whether you should go right or left as you go. And always keep to the right-hand side wall.

[erikun]

How does that look like? Did you just mean 'with more columns'? Because I've been wondering whether there is something like a table with more than juz an X and Y, you know, like three dimensions on one table.

Axes, as in the plural of axis. It would depend on the software you have available of course, but I don't see a problem with creating a table with an X axis (level), a Y axis (stat), a Z axis (weapon), an A axis (enemy dodge), and a B axis (enemy armor), and then cross-referencing them all to determine the number to beat in a random generator. It would be rather difficult to model it on paper, though, even in creating the program.

[Dingle]

What I DO know is that while I want stats (which is what is tied to level) to mean something, I want skill to mean a lot more. A level 50 non-combat character should be able to kick the butt of level 30 monsters, but should still want a tank to hide behind going up against something near his own level.

What else do you want?

You said you had a world, and want a system.
What is your world like?

What should be possible?

Are there Giants? Orcs? Guns? Aliens? Zombies? Magic?

How much better is the best fighter(MR. Epic) than the worst?

Could the best beat an army of average guys, or could a good hit from average Joe kill him?

Can Mr.Epic punch through a brick wall?

How different should PCs be from NPCs?


I can't help you if you don't tell me what you want

[housellama]

What else do you want?

You said you had a world, and want a system.
What is your world like?

I can't help you if you don't tell me what you want

I can tell you what I want, but I can't... well, more accurately won't, tell you about the world.

I am doing an experiment for the joy of doing so, and the knowledge that comes from such an experiment. For simplicity's sake, I am borrowing someone else's IP. My experiment is building a workable MMO on paper. Setting, system, classes, items, quests, factions, the works. If I can find volunteers to help me with concept art, I might even do that. I have avoided interaction with canon as much as possible. In creating the quests and factions and such, I have kept to the spirit of the setting without interaction with canon. I have used that IP because it's easy. It fits the ideas I had. The experiment is not about the IP, it's about the sum of the parts and learning how those parts fit together.

Since this is a not-for-profit experiment, which I don't expect to go any further than my hard drive, and I don't have any plans to do anything with it ever, I'm within fair usage. However, I am extremely sensitive to all of the legal issues surrounding the use of intellectual property. I don't want to go broadcasting that all over the place.

For the purposes of this discussion, it's modern day urban supernatural. Most people don't know about the existence of the supernatural. Guns hurt. Guns will be the main weapon of mortals. Some things may have supernatural strength or speed. Some things can do magic. Some may have access to some sufficiently advanced technology. As far as the system goes, the type of damage will differ. Some things will be more resistant than other things.

As far as threat levels go, the range should be +/-5 levels. A PC has an outside chance of beating a mob that is no more than 5 levels above him. A mob more than 5 levels below a PCs level shouldn't be a threat to him. The exception is endgame PCs. That should depend on gear level and experience.

-Housellama

[Dingle]

A D&D type system might be good for you.
In fact, you could take D&D 3.5 (BAB, Saves, and stuff) and change weapons/armour/magic items and magic a bit.

D&D CR works under the assumption that double the number of monsters is equivalent to +2 to CR
If you're not happy with this, D&D 3.5 is conveniently almost entirely all linear, so you can just multiply all numbers by 5 (including die size d20 => d100), and then split up the new levels to have 100 levels.


Considered creating a new
beginnings of (analytical quadratic approximation), but it's awkward

Spoiler

Show

a linear increase of several statistics each level should be simple enough, and if you do the initial values and increases right,you can get what you want.

ok, taking a to be damage increment and a0 to be lv1 damage
and taking b to be health increment and b0 to be lv1 damage


A lv x charachter does (a0 +ax)(b0 + bx) damage to a certain monster before he dies.
a lv x-5 charachter does (a0 +ax -5a)(b0 + bx -5b) damage to that monster before he dies.

if a lv x-5 charachter should do 1/2 as much damage, then:
a0b0 + (b0a+a0b)x + abxx = 2*( a0b0 + (b0a+a0b)x + abxx -5a(b0 + bx -5b) -5b(a0 +ax -5a)

a0b0 + (b0a+a0b)x + abxx = 2*( (a0b0 -5ab0 +25ab -5a0b +25ab ) + (b0a+a0b -10ab)x + abxx)

We have 4 variables, so we can create 4 equations by picking x = 6, 10, 15, 20 Then, Solving the equations, we can get values for a0, a, b0, b