Single Success Systems vs. Multiple Success Systems

[c_stone_bush]

I am working on the mechanics for a new game and am trying to decide on a system for task resolution. Which do you think I should use:

a) The player rolls a single die (or multiple dice and add them up), applies any modifiers (possibly stats, skills, equipment, etc.) and complares it to a relatively high target number. If the modified roll equals or beats the target number the task is successful. The more the die roll beats the target number the greater the success.

b) The player rolls several dice, has a pool of modifiers that they may split up and apply to the dice individually, and compares the modified resuly of each die to a relatively low target number. Any dice that equal or beat the target number are a success and the more successes you get the better you perform your task.

I guess that I am looking for help on the pros and cons of a single success system compared to a multiple success system.

[John Kim]

I guess that I am looking for help on the pros and cons of a single success system compared to a multiple success system.

First of all, I discuss some in my extended essay on dice mechanics, on my http://www.darkshire.net/~jhkim/rpg/systemdesign/">System Design pages.  

For a moderate number of dice, dice pools are easier to handle.  You don't have to keep numbers in your head because you can count by sorting the physical dice.  And comparisons are easier than addition or subtraction.  

My personal beef with dice pools is that they don't scale well.  They can't go below one, and one or two dice usually have odd behavior.  And dealing with more than six dice becomes increasingly unwieldy.  Moreover, the shape of the probability curve changes for different numbers of dice -- usually in ways that are not intended.

[Valamir]

My personal beef with dice pools is that they don't scale well. They can't go below one,

Actually its fairly easy to go below one with a die pool.  One simply starts to add more dice and taking the worst roll instead of the best.  Its not a perfectly smooth distribution without hiccups, but it gets the job done.

[c_stone_bush]

I like the idea of taking the worst roll for negative die pools Valamir, ver interesting. But perhaps I didn't explain my "dice pool" idea as clearly as I could have. Here is what I meant:

The GM tells you that the target number to hit the orc in front of you is 6, and that you need 3 sucesses. You have a stat of 3 (meaning that you roll 3 dice) and your modifiers are +4 (for your skill of 4 in swordweilding). You roll your 3 dice (d6's for this example) and get a 3, a 5, and a 6. The 6 is a success, and since you have +4 to distribute amoung your die pool you modify the 3 to a 6 (using 3 points) and the 5 to a 6 (using the last point), giving you 3 successes. Your character hits the orc.

My question is, since the player is only attempting one action, what would the reason be for requiring multiple successes?

[DNAphil]

Using multiple sucesses for a single action is one way of looking at how sucessful a task is.  Several systems have used this concept, that one sucess is needed to complete the action at hand, but that the more sucesses you rack up the more sucessful your action was.

Personally, I am more of a single die guy.  It may be more limiting then a pool of dice, but havng come from games where I have had to roll handfulls of dice, it gets to be annoying.  You have dice spill about, you need more surface area to roll.

I belive that In Nomine from SJG used three d6's for action resolution.  The first two were to acheive the target number, the final die was a scale of how sucessful you were.

Both systems have been done before.  Having played games with both, I am for single dice resolution.  It is faster, and gets the answer quickly and back to the story.

My two cents...and my first Forge Post!

[Mike Holmes]

Welcome Phil. Is that just an indication of your preferences? Sounds like it unfortunately.

OTOH, really all that you're going to get is preferences until we have some context, Stone. That is, not knowing what your game is about, or any of the details, all we can give are opinions and generalizations (which John has nailed down). It may turn out that there's something about your game that's really telling in terms of this.

Can you give any relevant details that might help? First and foremost, do you have a mechanical use for "successes" generated by a multiple success generating system? If not, then that says a lot right there. Just having it as an indicator to the GM about narration probably isn't enough of a reason to complicate the system.

See what I'm saying? What can you tell us?

Mike

[Brian Leybourne]

I think you really need to be asking your setting this question, not us.

There's a propensity among game designers to try to come up with "the coolest new die system". Hell, I'm guilty of that myself, in fact often I get no further than that :-) But really, the die system needs to support the game and setting, and integrate well. THAT is the most important consideration (and is what Mike is trying to get across as well; I'm really just saying "me too" to Mike's post).

Some of the best systems out there are the ones that "fit" the game setting and genre so well, like Deadlands' use of poker chips and cards for a wild west game, Sorcerer's simple D6 system for a game that is heavy narrative and needs a system that doesn't get in the way, D&D's heavy system that trys to cover every possible rule because, really, it's a tactical wargame pretending to be an RPG, and so on.

So look back at what the game will actually be about, and you may find that the answer is staring you in the face. If it's really not coming to you, tell us what the game is about, and we may be able to make suggestions that are not simply personal preferences, which is all you're going to get right now.

Regards,
Brian.

[c_stone_bush]

As far as context goes, I don't have anything yet. This was going to be a pretty universal system so I figured that any of my character/setting ideas could ues the system. I was mostly wondering why for example Shadowrun uses the multiple success system and D&D uses the single success system. Preferance? I guess thats probably the case. Someone on another board said something that I thought was interesting:

This has two sides; either you're trying to perform as complex a task as possible in a fixed time, or you're trying to perform a task of fixed complexity in as much time as it takes.

The fixed-time scenario is usually used for things like combat, where extra successes can represent a cleaner hit, more force behind the blow, or perhaps a called shot to avoid armour.

The fixed-complexity interpretation is useful for longer-term tasks. For example, chopping down a tree isn't a difficult task, so the target number should be low (it's hard to miss hitting a tree). It should take a long time though, and one lucky swing shouldn't drop it in a single blow. So you can give it a TN of 2 but require the character to accumulate 20 successes to fell the tree.

So I guess its just a question of whether I want to calculate how sucessful an action is by how much they beat a target number by, or how many successes they get. By the way, I'm so happy that I happened on this forum. Now I actually have something to do at work!

[Brian Leybourne]

This gets said a lot around here, but that's only because it's a fairly good point.

If you're simply gunning for a universal system anyway, why not just use one of the many universal systems that are already out there and free, such as FUDGE. You have the advantage that it's already been designed in some depth, is very flexible and adaptable, and since you don't have to blow time working out every nuance of the die system you have more time to make the actual game(s).

What I'm saying is, what will your universal system offer than others don't already (except the fact that it's your baby, and that's not a small thing by any means).

Brian.

[M. J. Young]

My question is, since the player is only attempting one action, what would the reason be for requiring multiple successes?

Up front: I am not a fan of dice pools. I like straight probabilities, or even rolls with curves in them that are easily managed.

So why would anyone use dice pools?

I think in part it's an effort to obfuscate the odds. Seriously, I think it's done so that players can have a feeling that they know how the probabilities are shifted, but can't actually know.

A simple dice pool has at least three dials: how many successes do you need, what number constitutes a success, and how many chances to you have to roll it? You increase the number of successes needed, and the chance of success goes down; you increase the target number (on a meet or beat system) and the chance of success goes down; you increase the number of dice rolled and the chance of success goes up--but without doing some rather complex math at the table, you've no idea how much impact any one of these changes is going to have. Is it better to roll 4d6 to get three successes where 4 is the target, or 5d6 to get two successes where the target is 5? You can't guess reliably; therefore, it is much more difficult to make in-game strategic decisions based on how it will impact the odds.

Finally, Brian has a point. I know why we designed Multiverser--there was something we didn't think any "generic" or "universal" or "multigenre" system had ever done successfully, and we thought we could solve that problem (and I believe we did). What is it that you will accomplish by your system that hasn't already been solved effectively?

--M. J. Young

[efindel]

Some semi-random thoughts:

A "roll plus adds vs. threshold" system gives guaranteed success at some point (when minimum roll + adds is greater than or equal to the threshold).  If you want there to always be a chance of failure, you have to resort to rules of the "a natural 1 is always failure" sort, or involve rerolls.  The same is true for failure; if maximum roll + adds is less than threshold, you've got guaranteed failure, barring special rules.

A dice pool vs. fixed difficulty per die, on the other hand, always has some chance of success and failure, unless the per-die threshold is automatic or impossible for a die.  (And, of course, you can also do open-ended rolls with dice pool dice -- e.g., Shadowrun).

One characteristic of a dice pool, used in the early games that had them, is that it is a *pool*.  If a character is going to do two actions, you can have them split the pool up.  The Riddle of Steel makes use of this in its combat system -- a single pool is split between attack and defense.  You can also use the pool as a pool in another sense -- dice used from it "go away" and are refreshed later.  

Now, you can do similar things with a "dice + adds" system... dividing up adds, having used adds go away to be refreshed later.  With a dice pool system, though, if you have enough dice around, then you've got a convenient physical representation at the table.  You could use poker chips or such with an adds system, but to actually make a roll, you need to know exactly how many adds you have... which may require stopping and counting chips.  

Some variant systems should be noted:

- separate die for level of success.  Someone already mentioned In Nomine, where 2d6 determined success/failure, and a third one the degree.  Chivalry & Sorcery 3 & 4 do the same, but with 2d10 (which may have been being used as 1d100 -- not sure off-hand) for success/failure, and 1d10 for degree.  This is quick, simple, and easy to read.

- "best X" or "worst X" dice from a set.  This has been mentioned as a way to do very low probabilities with a dice pool, but it can also be used in a non-dice-pool way.  For example, the system for Don't Look Back totals the bonuses and penalties to an action to get a modifier, usually in the range -5 to +5.  The player rolls (3 + asolute value of modifier)d6.  If the modifier was positive, the best 3 are added together.  If it was negative, the worst 3 are added.  A table gives degrees of success/failure.  (Basically, 10 or worse is failure, with lower numbers being worse.  11 or better is success, with higher numbers being better.)

Both of these methods share a trait:  the possible range of results never changes.  No matter how low your chance of success, a success of the highest level is always possible... and no matter how high your chance of success, a failure of the worst level is always possible.  In the first sort of system, degree of success/failure is independent of the chances.  In the second sort, it changes along with the chances... if you're rolling 8d6, keep the best 3, a successful roll is more likely to be a "high" success than it is if you're rolling, say, 5d6, keep the worst 3.

Lastly, I'd like to mention opposed dice pools.  Sorcerer and Donjon use one method of this -- two sets of dice are rolled against each other and compared to each other.  Universalis and Story Engine use another method -- two dice pools are rolled with fixed difficulty, and the one that has the most successes is the winner.  One possibly interesting aspect of the latter method is letting players cancel opposing successes -- e.g., you could go "reckless" and use all your successes to inflict harm on someone, saving none for defense, or you could use some defensively to take less harm, but do less damage yourself.