Best Role & Keep Mechanic

[Anderlith]

I've been looking to create my own RPG for a while now & I really like the idea of rolling a handful of dice & keeping a few. I'm wondering if the playground new of any good roll & keep based games they would recommend mechanics wise & the strengths/weaknesses of them.

[Eisenheim]

Seventh Sea is my favorite. It's lighter than any of the L5R editions I looked at. It's main weaknesses are that the magic is strongly tied to the setting and that the default char gen rules produce useless characters. If you up base traits to 2 and give everyone a bit of xp on top of their build points, it works great.

[Berenger]

SIFRP (A Song of Ice and Fire Roleplaying) does this, but I didn't play it long enough to have a real grasp of the strengths and weaknesses.

[Drynwyn]

Seventh Sea is my favorite. It's lighter than any of the L5R editions I looked at. It's main weaknesses are that the magic is strongly tied to the setting and that the default char gen rules produce useless characters. If you up base traits to 2 and give everyone a bit of xp on top of their build points, it works great.

I've found it's easier to just give everyone 150 starting character points.

Alternatively, say "Screw it", give everyone 300, and remove the starting caps on things, because that's how all the significant characters in the books were built and you don't feel like going through the obligatory leveling-up process in a high adventure game.

[Florian]

I've been looking to create my own RPG for a while now & I really like the idea of rolling a handful of dice & keeping a few. I'm wondering if the playground new of any good roll & keep based games they would recommend mechanics wise & the strengths/weaknesses of them.

The actual "Roll and Keep" mechanics as used in 7th Sea and L5R are quite solid.
The 4th Edition L5R core book actually mentions all flaws said system could have and proposes things to counter them.

[Eisenheim]

Drynwyn, my group varied between 100 and 150 build points, depending on how high powered the start was, and usually did 50 xp, to let people expand advanced knacks. We also made virtues free (vices still got you 10 back)

[Drynwyn]

Drynwyn, my group varied between 100 and 150 build points, depending on how high powered the start was, and usually did 50 xp, to let people expand advanced knacks. We also made virtues free (vices still got you 10 back)

Most of that would work out fine. Free virtues makes virtues TOO good a choice IMO, given how potent many virtues are, but to each his own.

[NomGarret]

Even L5R has changed the roll & keep system over the editions. From roll skill+trait keep ring (which is the lower of mental or physical aspects of an element) to roll skill keep trait, to the current roll skill+trait keep trait. With the emphasis abilities for higher skills in 3rd and 4th edition, the last system works better than the other two did.

I think a big trick is properly weighting the costs of one over the other. I love 7th Sea, but the upfront cost spread made traits a way better investment than knacks.

What I would like to see is a roll and keep system similar to FFG's Star Wars games. Roll the higher of skill or trait, keep the lower of the two. That way, raw talent and dedicated study are equally important and it's the combination of the two that takes characters to the next level.

[Magic Myrmidon]

The actual "Roll and Keep" mechanics as used in 7th Sea and L5R are quite solid.
The 4th Edition L5R core book actually mentions all flaws said system could have and proposes things to counter them.

Really? That sounds super interesting. Is it in the main book? What section?

[D-naras]

Really? That sounds super interesting. Is it in the main book? What section?

If I remember correctly, it had suggestions on what to do when someone reaches the maximum 10k10. I think, you never rolled more than 10 dice. For every 2 dice over 10 you would roll, you got to keep 1 more. So a roll of 12k3 became 10k4. Also, for each die over 10 you would keep, you added 2 to the total result. So a supposed 12k12 roll would become 10k10 + 6. That's all it says on the subject.

I've played L5R 4e for a couple of years and the roll system is fun but I've found that at some point, everyone becomes the same in terms of rolling. 6k3 for all the usual rolls which kinda bothered me. Also in L5R, traits are king. Skills cost much less experience but don't do nearly as much. For instance, I build my character as a Stealth expert, rolling 9k4 (-5 due to armor). At some point, the rest of the party invested in stealth (sneaking through the shadowlands does that) and suddenly everyone rolled 7k4 without armor penalties (2 of them being mages) and the other bushi rolled 8k5 which is plain better than 9k4. As mentioned before, you roll dice equal to you Trait+Skill rank and keep a number of dice equal to your Trait. Skill ranks range from 0 to realistically 7 while Traits range from 2 to 5.

If you want to use this as the basis of the system, you'd do good to find a way to make skills more important for skill rolls. I suggest making skill ranks cheap to raise and/or find a way to make skills add to the dice you keep (maybe keep 1 more for every 3 dice you roll?)

[Anderlith]

Well my original idea was your roll skill k stat, the more skilled you are the bigger pool you have. I want people to have big pools if they want, since you can only keep so many. Maybe 10 being large & 4 being average professional. My hold up is if what should I do if they have no skill ranks? Maybe give everyone 1/2 of your stat to your pool? So an untrained guy with 4 str & no skill climbing would roll 2k4? Or maybe roll stat keep 1 I.e. 4k1?

[VoxRationis]

So for those of you who play with roll & keep mechanics, I have a question: How do you estimate your odds of success? I'm used to playing with d20 systems, and those are quite easy to figure out, but the numbers are harder to crunch with a handful of dice, only some of which are retained, and some of which explode. Do you guys make probability tables in advance and staple them to the back of your character sheets? Do you just wing it?

[Magic Myrmidon]

I didn't do this when I played L5R, but when I play Savage Worlds, basically. I go to anydice.com, run the combinations of dice that I will typically be rolling, and just remember the probabilities of getting at least a 4, for example.

[Faily]

L5R Probablity Chart (http://lynks.se/probability/)

Emphasis matter a huge deal in statistics of L5R 4E, as you may re-roll 1s once on a roll if you have the appropriate emphasis. (In 3E, an emphasis let you add your skill rank as a static bonus to the roll)

So the aforementioned Sneak-specialist could've gotten ahead of the other party members' Stealth-increase by taking the appropriate emphases, and probably picking up the Advantage Silent (which also give a static-bonus to Stealth-rolls). Other Advantages like Touch of Yomi give +1k0 to one specific skill roll, or Prodigy gives +1k0 to all your School skills.

Thing with L5R is that it has the extra mechanic of Void points. You have a number of Void points per day, which you may use to do perform even better, either by gaining +1k1 on a roll, or temporarily have a rank in a skill you do not possess. I am personally very fond of the Void point mechanic, since when you feel like it really matters to your character they get that extra boost. In a way they can give 110% in L5R. :smallwink:

[Florian]

Really? That sounds super interesting. Is it in the main book? What section?

Its somewhere at the beginning of the rules section, one of those white text on black sidebars. Think it was labeled something about returning players.

Short version: They knew that rings used to more important than skill ranks, that's why they added skill mastery ranks to the individual skills, changed how emphasis works and modified a lot of schools to include an upgraded 3k1 void roll option for iconic school skill uses.

@VoxRationis:

That depends a bit what exactly for.
As the number of dice you keep and then add up is always the same, you basically know for each trait what the expected average should be and what maximum is reachable (sans exploding dice), i.e. Intelligence 4 is ~22/40.
Every additional die rolled beyond that raises the expected average by roughly +2.
Taking explosion into account, each added die raises the maximum by roughly +5 and the average by an additional +1.
(Int4/Lore4 is 8k4 and can be expected to be ~34/60)

Note that this is only a rough calculation, but actually all that is needed to decide if using a void point helps or to risk it and call a raise.

[Drynwyn]

So for those of you who play with roll & keep mechanics, I have a question: How do you estimate your odds of success? I'm used to playing with d20 systems, and those are quite easy to figure out, but the numbers are harder to crunch with a handful of dice, only some of which are retained, and some of which explode. Do you guys make probability tables in advance and staple them to the back of your character sheets? Do you just wing it?
Generally, I use a rule of thumb of your average result being the average on your kept dice + 1/5 the maximum result on the die for each unkept die (+1 for d6's, +2 for d10's, mainly.) This isn't SUPER accurate, but it generally suffices when rolling no more than twice as many dice as you keep.

[LibraryOgre]

Well my original idea was your roll skill k stat, the more skilled you are the bigger pool you have. I want people to have big pools if they want, since you can only keep so many. Maybe 10 being large & 4 being average professional. My hold up is if what should I do if they have no skill ranks? Maybe give everyone 1/2 of your stat to your pool? So an untrained guy with 4 str & no skill climbing would roll 2k4? Or maybe roll stat keep 1 I.e. 4k1?

You might also simply go with "Roll skill (minimum 1), keep stat." Or "Roll skill (minimum 1 with certain, relatively universal skills), keep stat."

It means that having a 1 in a skill isn't terribly impactful, since it's about what someone who is untrained might have... in essence, it become a bit of a point tax to advance further. But it lets you have a defaulting mechanic when necessary.

[Florian]

Well my original idea was your roll skill k stat, the more skilled you are the bigger pool you have. I want people to have big pools if they want, since you can only keep so many. Maybe 10 being large & 4 being average professional. My hold up is if what should I do if they have no skill ranks? Maybe give everyone 1/2 of your stat to your pool? So an untrained guy with 4 str & no skill climbing would roll 2k4? Or maybe roll stat keep 1 I.e. 4k1?

Before we advance too much in that direction, let us talk what the pool is for, how sucess or failure is determined by it.

For example, Roll and Keep uses a Target Number (TN) similiar to D&Ds DCs. You take all your kept dice, add the values up and compare that with the TN. You can call for a raise of the TN to incorporate additional effects.

Shadowrun 5 uses a variant of that, as each individual die must beat the Target Number. The number of "hits" generated in this way are then used to determine the effect.

[Anderlith]

Well I have two trains of thought on that.

My first idea was that the die type would change depend on what you were. I.e. Dragons had d20str & humans had d8s for everything. While elves would have d8s for everything but d10s for dex to show a physiological difference. This would show that a human & an elf could train their stat up to the same level but since the elf rolls xd10ky & a human rolls xd8ky the elf would always have a slight advantage. This system would count die numbers added together to reach a Target Number but it also involves a lot of math every roll.

My other thought is to only count 9s &0s as successes, & have advantageous creatures succeed on lower numbers. I.e. A human succeeds at 9-0 on dex & an elf succeeds at 8-0 dex. This would be easier as you would only have to count successes, but may not have the balanced range of effects as the previous method.

If this doesn't make since I can try to explain better.

Which sounds the easiest/interesting to use?

[Florian]

@Anderlith:

I understand what effect you want to invoke with that, but I think both methods are overly complicated in their own way.

Your first method will generate the same underlying problem that Savage Worlds has, when you want to include exploding dice (you should do that), namely that lower value dice explode more often, leading to the comical effect that your humans (or d6 on dex dwarves) will outperform your elves or dragons with sheer luck.
You'll have it easier if you stick to using the same mechanics and dice for all, but include things like racial traits that lower the TN for certain tasks, and/or give a free bonus to a sucessful task.

(Example: Elf Trait - All Dexterity based TN are 5 lower than usual.)

The second method, generating hits with higher value dice will lead to the same problem oWoD had, mainly needing an absurdely large die pool to generate hits for even the most mundane task.

[VoxRationis]

The second method, generating hits with higher value dice will lead to the same problem oWoD had, mainly needing an absurdely large die pool to generate hits for even the most mundane task.

No kidding. I've only played WoD once, but I came to the conclusion that we were the most pathetic vampires ever, on account of our characters routinely failing tasks that I felt reasonably confident I, a not-particularly-skilled mortal, could accomplish.

[Knaight]

The second method, generating hits with higher value dice will lead to the same problem oWoD had, mainly needing an absurdely large die pool to generate hits for even the most mundane task.

It depends on the specifics. For instance, Burning Wheel has a "shade" system, representing natural talent and such. Typically, one succeeds on a 4-6, which doesn't need huge pools at all. Those talented to the point of being borderline supernatural succeed on a 3-6, and have a good chance of throwing more dice anyways. The actually fully supernatural, in a way beneficial to the specific task succeed on a 2-6. This works pretty well, without ever needing huge quantities of dice. Only succeeding on a 6 on the other hand...

[NomGarret]

Well my original idea was your roll skill k stat, the more skilled you are the bigger pool you have. I want people to have big pools if they want, since you can only keep so many. Maybe 10 being large & 4 being average professional. My hold up is if what should I do if they have no skill ranks? Maybe give everyone 1/2 of your stat to your pool? So an untrained guy with 4 str & no skill climbing would roll 2k4? Or maybe roll stat keep 1 I.e. 4k1?

This is why I suggest roll (higher of trait or skill) keep (lower of trait or skill). Your version strongly encourages traits to be kept exactly equal to relevant skills, to the expense of concept.

[LibraryOgre]

This is why I suggest roll (higher of trait or skill) keep (lower of trait or skill). Your version strongly encourages traits to be kept exactly equal to relevant skills, to the expense of concept.

Not necessarily. In a roll and keep system, there's still advantages to having a high skill, though sometimes, your low attribute will hinder it.

Let's say we're using D6s, and we add dice. In a 5k3 situation, even if I roll two 1s, the other dice will preserve me. You'd wind up setting your difficulties based off the die size and average attributes.... so, if an attribute of 2 is average, an average difficulty might be 7. 2k2 will average to 7, but 3k2 is more likely to be 7, and 4k2 is even more likely. If a 6 attribute is normal, then the difficulty might closer to 20, and a roll 5k3 is more or less never going to make it.

It all depends on where you set the bar. A low attribute paired with a high skill will mean that you've got some caps on your ability, but you're not hopeless. A high attribute paired with a low skill will mean you apply everything you have, but it's not much.

[Anderlith]

The first example would not use exploding dice. & in both versions attributes would be hard to raise, encouraging skills over stat.

[NomGarret]

Oh there are certainly advantages to a higher roll than keep number. 4k3 will average out better than 3k3, but 2k3 is functionally identical to to 2k2.

[LibraryOgre]

Oh there are certainly advantages to a higher roll than keep number. 4k3 will average out better than 3k3, but 2k3 is functionally identical to to 2k2.

Sure, but that's why you tend to keep your attribute, instead of your skill. You can easily increase skills, and it also means that the untalented-but-skilled will have a disadvantage vs. the talented-but-skilled.

[Anderlith]

Honestly, roll skill, keep stat is just more realistic feeling. I don't plan on changing that

[Knaight]

Honestly, roll skill, keep stat is just more realistic feeling. I don't plan on changing that

What I've seen work here is roll skill+stat, keep skill. Talent is still worth something, but the talented and unskilled get trounced by the skilled but not naturally talented, which makes a lot of sense. Natural talent also counts for a lot more among competitions of the unskilled against each other, while at the same time also still mattering at the top. This also avoids the issue where you have NkM where M>N. The only real downside is that the dice pools get a bit big, and that slows things a bit.

[NomGarret]

Sure, but that's why you tend to keep your attribute, instead of your skill. You can easily increase skills, and it also means that the untalented-but-skilled will have a disadvantage vs. the talented-but-skilled.

Yes but the high skill low talent (4k1) has a significant advantage over high talent low skill (1k4). Now, you may feel this makes sense in a particular setting or genre, but do it consciously.

[Florian]

@Anderlith:

How about a small variation of it:
Number of skill ranks = size of pool. Attributs range from negative to positive (-2 to +2). Roll pool dice, keep highest value die add attribute, compare to TN.

[LibraryOgre]

Yes but the high skill low talent (4k1) has a significant advantage over high talent low skill (1k4). Now, you may feel this makes sense in a particular setting or genre, but do it consciously.

Why shouldn't they, though? It's a conscious choice, sure, but we're talking game design... they're all conscious choices. One of the key differences becomes advancement... the 4k1 gets only incrementally better by adding a +1k0, because their skill is really limited by their natural abilities. A 1k4 gets a lot better by going +1k0, because they're unlocking their talent.

[Faily]

I'm personally a fan of how most recent editions of L5R have done it with skill+trait - keep trait, in addition to their mechanics of Raises and Explosions.

Lets take attacking with melee weapons for instance, in this case with swords. It will use the Kenjutsu skill, and the Agility trait. Now, someone with no practice with the sword, but Agility 3 can roll three dice. But according to the rules of L5R, since its untrained, the dice cannot explode and they cannot call Raises (increasing the difficulty for better effect, like more damage). Their natural Agility helps to a certain degree, but a person with Agility 2 and Kenjutsu 1 will fare better because they can benefit from being trained in the skill.

In the end though, having a decent Trait is nescessary to back up being awesome in a Skill. It's hard to be an amazing performer if your Awareness is 2, because you lack that natural charisma. Sure you can train yourself to be technically good, but you lack stage-presence, charisma... that little extra something. While someone with less skill but raw natural talent might perform better. It's a rather delicate balance I find to upkeep Skill and relevant Trait to properly benefit from both. It also helps that skills have the added advantage of the Emphasis-mechanics, as well as the Mastery Abilities that unlock as you increase your Skill.

In short, you need both the talent and the skill to be amazing at something.

[Anderlith]

So I think I will go with a system of (Skill)k(Trait) using d10s & with variable target numbers for success based on race. I.e. an elf succeeds on 7s & above while a human is 8s & above. The number of successes you roll effect how well you did usually with a single success being needed. Tens explode.

If untrained in a skill you roll (Trait)k1. Tens do not explode.

My next issue is what traits should be in the game. Right now I'm thinking of a system where your base stats combine & average out into secondary stats. These secondary Traits are used for skill tests leaving your base stats for saving throws & such.

Strength - Muscles & physical power
Perception - Natural awareness & precision
Agility - Balance & speed
Toughness - Endurance & physical resistance
Wits - Resourcefulness & quick thinking
Intuition - Innate knowledge & smarts
Resolve - Willpower & grit
Grace - Composure & physical fluidity

Athletics (str + agi)/2 Used for physical skills (climb, swim, etc)
Dexterity (agi + per)/2 Used for delicate skills (mechanics, surgery, etc) & attacking
Insight (per + intu)/2 Used for seeing things & sensing motives
Intellect (wits + intu)/2 Used for knowledge skills
Charm (wits + gra)/2 Used for seduction & subterfuge skills
Presence (str + gra)/2 Used to command & direct people

Toughness & Resolve are both used together to determine a characters hp equivalent. I'm using hit boxes like in Top Secret. The is also a Fatigue stat.


I'd really like to have something like Reflexes but dont want to blow my system any more.

What's the playgrounds opinion?

[Knaight]

So I think I will go with a system of (Skill)k(Trait) using d10s & with variable target numbers for success based on race. I.e. an elf succeeds on 7s & above while a human is 8s & above. The number of successes you roll effect how well you did usually with a single success being needed. Tens explode.
...
What's the playgrounds opinion?
Honestly, this seems clunky. You've got too many moving parts, a weird hybrid of a roll and keep mechanic with an exploding dice pool mechanic, the generic necessary single success oddity, etc. If you're going to do roll and keep, I'd recommend doing it where you add the numbers together.

[Anderlith]

Darn... I didn't think about the roll & keep with a success based system :/ how would different races benefit then?

[Stubbazubba]

So I think I will go with a system of (Skill)k(Trait) using d10s & with variable target numbers for success based on race. I.e. an elf succeeds on 7s & above while a human is 8s & above. The number of successes you roll effect how well you did usually with a single success being needed. Tens explode.

What's the playgrounds opinion?

I can't get past this part. This is needlessly clunky. Roll and Keep is one thing, and Storyteller-style Hits is another, and you've mashed them together for no apparent reason. Is there some quirk of this chimera that justifies the extra book-keeping and resolution steps? How is this better than just rolling SkT, or just making a d10 dicepool of S+T and counting hits over a racial threshold?

In response to the original question, another Roll 'n' Keep system to look at is Cortex+, found in the current Firely RPG, the OOP Marvel Heroic RPG, Leverage RPG, and Smallville RPG, as well as the Cortex+ Hacker's Guide.

[Knaight]

Darn... I didn't think about the roll & keep with a success based system :/ how would different races benefit then?

You could just do something like skill bonuses. With that said, you should think about why you are choosing to have multiple species, and then build mechanics to support that. Take a look at Burning Wheel, and some of the specific mechanics it has to really get the Tolkien feel across (lifepath differences, but particularly Hate, Greed, and Grief).

[Anderlith]

Well my original idea was to have different races roll different dice for the trait, but had a few people didnt like that. So I was just deciding to change it...

In example an elf has a d10 Dex while a human has a d8 for Dex. So if they both have three ranks of Dex, & 4 ranks in Bows the roll would be 4d10k3 for the elf & 4d8k3 for the human

This was my first idea. You'll see it in my first post. I only changed my thoughts because people didn't seem to like it.

[Anderlith]

You could just do something like skill bonuses. With that said, you should think about why you are choosing to have multiple species, and then build mechanics to support that. Take a look at Burning Wheel, and some of the specific mechanics it has to really get the Tolkien feel across (lifepath differences, but particularly Hate, Greed, and Grief).

I'm only doing about five races; humans, genetically enhanced supersoldier humans, ancient magic humans (elves), magically sensitive genetically enhanced dogs (still dogs, not dogmen) & maybe one or two other races.

[LibraryOgre]

So I think I will go with a system of (Skill)k(Trait) using d10s & with variable target numbers for success based on race. I.e. an elf succeeds on 7s & above while a human is 8s & above. The number of successes you roll effect how well you did usually with a single success being needed. Tens explode.

If untrained in a skill you roll (Trait)k1. Tens do not explode.


Ok, a few things.

1) I do like the "Unskilled is (Trait)k1". I think it's a great solution.

2) So, variable numbers for success based on race... is that situational? Is it "Elves have a virtue by which they succeed on a 7 for singing and dancing checks", or is it "Elves are better, suck it, roundear"? 'Cause the second is going to be a lot harder to balance, while the first opens possibilities for edges, advantages, etc.

3) How does 10s explode work in this case? Is it "If you roll a 10, you get another die in your dice pool"? Because with a threshold based success/hits model, that's what makes the most sense, but I want to be sure.

[Anderlith]

Yes rolling a ten would allow you to roll the die again.

More, suck it round ear. They have a 1/10th extra advantage on agility & grace, & suffer 1/10 less in Toughness basically. It kind of balances out but allows the player to feel a difference.

While a human & an elf could have the same number of dots in a Trait the elf would always have that slight advantage. Honestly its just a thought I had to give another facet to characters. I never liked a flat bonus to a stat based on race or what have you. Having variable target number will allow everything to be set on the same "scale" yet still giving advantages to things that are physiologically different than human.

[LibraryOgre]

Yes rolling a ten would allow you to roll the die again.

More, suck it round ear. They have a 1/10th extra advantage on agility & grace, & suffer 1/10 less in Toughness basically. It kind of balances out but allows the player to feel a difference.

While a human & an elf could have the same number of dots in a Trait the elf would always have that slight advantage. Honestly its just a thought I had to give another facet to characters. I never liked a flat bonus to a stat based on race or what have you. Having variable target number will allow everything to be set on the same "scale" yet still giving advantages to things that are physiologically different than human.

Ah, see, that debility in toughness can be a big one, depending on what you're using toughness trait checks for. If they have an advantage on all dex checks, and a penalty on all toughness checks, then, yeah, that's less of an issue.

[Anderlith]

What's your opinion of Stub's ?... It does seem clunky to have roll & keep system with a success hit system together... I did t think that all the way through

[LibraryOgre]

I don't have a particular problem mixing the two... in some ways, that's what the new Shadowrun is doing, though they're making the "keep" dependent on your gear.

But there's a question I think you need to consider: What counts as a "die" you are keeping? Let's say you've got a 6k1. One of those six dice comes up 10 three consecutive times. You can only keep 1... but does that mean you can keep that one die, with its 3 hits, or is each subsequent roll considered a separate die?

[Cluedrew]

Actually I think ShadowRun has been using roll & keep for a longer time than that. (pool)k(limit) is a role & keep system, except that the keep limit is usually high enough that it doesn't come up very often.

As for the main topic at hand. Doesn't rolling Unskilled at (trait)k1 make getting your first rank in the skill (skill)k(trait)=1k(trait) worse than what you had before? At other low numbers you might still have that problem but at skill=0 you get one die from a pool of (trait) which is strictly better than rolling a single dice and keeping it.

[LibraryOgre]

Actually I think ShadowRun has been using roll & keep for a longer time than that. (pool)k(limit) is a role & keep system, except that the keep limit is usually high enough that it doesn't come up very often.

3e Shadowrun and earlier was a success system... you rolled X number of dice, and could have up to X number of successes.

4e Shadowrun had some keep mechanics... in spellcasting, you rolled Stat+Skill, but could only keep up to Force.

[Kuroshima]

First at all, if you're working with Roll and Keep, know that the system has a lot of statistical quirks. Do the math, specially when it's not trivial. Anydice (http://anydice.com/) is your friend. One very common pitfall is to assume that you can add a linearly increasing difficulty. You can't. Unlike single die systems, all dice pool systems will have a bell curve (though R&K has a particularly irregular one), so a +X is not worth the same at all points. It is wrong to assume that since XkY will beat difficulty Z W% of the time, 2Xk2Y will beat difficulty 2Z W% of the time. It might, or it might not.

I don't dislike dice pool systems or exploding dice per se, but my experience has taught me that authors often don't do the math, relying instead in gut feeling. That fails BADLY with either dice pools or exploding dice, and fails horribly with both. I nowadays don't give systems that use such mechanics a second look before checking online that the author did the math.

Word of advice, unless you're ready to really hammer the quirks out of this system, don't invent a new system that uses dice pools, roll and keep, and/or exploding dice. Don't even tweak one that does. The chances is that it will become degenerate with a ton of corner cases (if it wasn't already, in the case of almost all R&K systems i've seen).

[goto124]

What are exploding dice anyway? (http://m.youtube.com/watch?v=JGkRQjt9hVc)