Dice Pool Mechanic vs. Single Die

[painandgreed]

Now that I think about, that would seem to me to be the ultimate goal of a die roll mechanic - to incorporate as much "processing" into the roll itself, and reduce special cases of how to read the die and modifications to the final roll as much as possible.

Not always. You can try and incorporate all the processing into a single roll, and even assuming that you avoid a system that breaks under certain situations (as simple systems tend to do) but it may appear lackluster to the players. WoD 2.0 went to a single roll mechanic for combat and that is the main reason that my plyers and I won't switch. It's great for a simple, quick playing game, if that's what you want, however, we all want some depth to our combat when it happens. Sometimes players don't want to make one simple roll to determine success because it all seems almost anticlamtic and removes the challenge (even if the odds for success remain the same).

 

[jester47]

Wil,

Your last post seems completely nonsequiter with what you quoted. I am talking about the way burning wheel works vs the way d20 works. What are you talking about?

In burning wheel your target is say a four. you roll 6 dice 3 of which come up as successful. the GM set the number of successes to acomplish said task to be 4. You fail.

I am asking how this is really any different from you roll a d20 add your skill ranks and attribute modifiers and come up with a 28. The GM set the DC at 30. You fail.

Burning wheel makes the claim that the GM should not be able to just set the difficulty of a task. But clearly the GM does to an extent

Thats pretty much my argument. What you are saying seems to be in a completely different vein. Are you offering an alternate system that does not have the DM/GM "problem?" I am generally curious as to what your posts about SilCore are all about.

I really don't mean to be so obtuse.

 

[Wil]

Wil,

Your last post seems completely nonsequiter with what you quoted. I am talking about the way burning wheel works vs the way d20 works. What are you talking about?

In burning wheel your target is say a four. you roll 6 dice 3 of which come up as successful. the GM set the number of successes to acomplish said task to be 4. You fail.

I am asking how this is really any different from you roll a d20 add your skill ranks and attribute modifiers and come up with a 28. The GM set the DC at 30. You fail.

Burning wheel makes the claim that the GM should not be able to just set the difficulty of a task. But clearly the GM does to an extent

Thats pretty much my argument. What you are saying seems to be in a completely different vein. Are you offering an alternate system that does not have the DM/GM "problem?" I am generally curious as to what your posts about SilCore are all about.

I really don't mean to be so obtuse.

Because, rolling six dice is statistically much different than rolling 1 die. If I roll 6 dice and require at least one of them to be 4 or higher I have a tremendous chance - well over 80% - that one of those dice will be more than 4. If I roll 1d20, I have precisely 1 chance - 15% - that the result will be 17 or higher. It's called a bell curve, and single die systems lack it.

Regardless of *how* BW counts the successes, in terms of probability it is very close to Silcore - namely, the higher the skill level, the more dice are rolled and the more predictable the results. Someone with a high skill level will predictably do better than someone with a low skill level. In a single die system like D&D, skill level doesn't matter on the raw die roll - there's always a 5% chance of hitting any individual number.

I also think there's more to the BW's "the GM doesn't set the difficulty level" than meets the eye, because my understanding is that the character's beliefs (or strength thereof) affect the difficulty level. So if the difficulty levels are hardcoded in the system (i.e., you always need higher than 4,5,6 depending on the skill level you chose) and are not changed by the GM, and the strength of the beliefs that the player chose for the character affect the chances of success, then in a lot of ways the difficulty is set by the player. If you choose to attempt a task that goes against your beliefs and it makes the task more difficult, the player has decided on the difficulty of the task. The number of successes required are a different axis entirely, and there's no claim that I can see that the GM doesn't do *anything*, just that the difficulty level is dictated by the player's choices.

Also, a tangental difference I can see is that If I have rules to handle extended tasks (meaning you need to accumulate successes), rolling 3 successes when I need 4 may mean mean that on the next roll I only need 1 more success. Single die systems normally are not used for this sort of thing (although you could by saying "Give me five rolls that beat DC 15).

In the end, single die systems are good for what they're good for, die pool or roll and pick systems are good for what they're good for - but it isn't "the same thing." The results between the two have quantifiable differences.

 

[IcyCool]

Wil, the BW system sounds like it has much more in common with Shadowrun and White Wolf(Storyteller). You roll from a dice pool, and count successes. I can see from your posts that SilCore is similar in that the number of successes is unimportant, only that you get one. How does requiring a greater number of successes affect the curve?

 

[Wil]

Wil, I'm not trying to be difficult, but could you perhaps focus on the BW system? It sounds like SilCore is different, and so comparisons to SilCore aren't all that helpful in this thread. SilCore sounds interesting and all, but the BW system sounds like it has much more in common with Shadowrun and White Wolf(Storyteller). You roll from a dice pool, and count successes. I can see from your posts that SilCore is similar in that the number of successes is unimportant, only that you get one. How does requiring a greater number of successes affect the curve?

I can only go off of the same stuff that is on the BW website, so maybe the BW guys can come in and correct me - but either way, whether it's "roll xd6 and keep the highest" or "roll xd6 and count dice that are higher than y", statistically it generates different numbers than "roll 1d20 + bonuses to beat DC". So my point was that the path taken to get to the numbers in very different.

I do think I misunderstood the OPs point - I thought that he was saying that the methods produce identical results, when in reality he was stating that the two operate the same way ("roll vs. target number"). In that, he is correct - even BW seems to have some amount of GM input into what is required to succeed at a task.

 

[Wil]

Here are the probabilities for BW:

http://www.burningwheel.org/pdf/bw_probabilities.pdf

Here are the probabilities for SilCore (for comparison purposes):

http://www.dp9.com/PDF/SilCOREdiceprob.pdf

To me, the spread of the numbers *is* the reason to use a die pool (or roll and keep) vs. single die. There may be no functional difference in the end, but using that logic I can ask, "Why use D&D instead of Palladium?".

You'll notice that in BW, requiring one additional success for someone with skill level 2 reduces his chances of succeeding from 75% to 25%. You'll notice that beating a difficulty of 4 vs. a difficulty of 6 with skill level 2 in SilCore results in about the same decrease in performance - meaning adding a success in BW might be said to be the same as adding 2 to the threshold in SilCore. To get the same decrease in performance in d20 (75% to 25%) on the raw die roll, you would need to increase the DC from 5 to 15. Of course, all that means is that d20 has less granularity then BW.

So to me the answer to "Why use one over the other?" will simply be player preference, as the mechanically it can be shown that BW and d20 do work on different assumptions, regardless of similarity in the final outcome.

 

[painandgreed]

There may be no functional difference in the end, but using that logic I can ask, "Why use D&D instead of Palladium?".

Books with indexes?

 

[Wil]

Books with indexes?

Well, there is that...once again proving that the end result might be the same, but the path to get there can be very different

 

[jester47]

I do think I misunderstood the OPs point - I thought that he was saying that the methods produce identical results, when in reality he was stating that the two operate the same way ("roll vs. target number"). In that, he is correct - even BW seems to have some amount of GM input into what is required to succeed at a task.

Thanks for clearing that up! I think you completely misunderstood me!
But I have some bad news for you. You still don't.


Ok, I am going to call up your other example because in it you get really close to what I am talking about:

Because, rolling six dice is statistically much different than rolling 1 die. If I roll 6 dice and require at least one of them to be 4 or higher I have a tremendous chance - well over 80% - that one of those dice will be more than 4. If I roll 1d20, I have precisely 1 chance - 15% - that the result will be 17 or higher. It's called a bell curve, and single die systems lack it.

So you would say that if you needed one success you had over an 80% chance of getting it if you rolled 6d6.

My point is that if the chance of success is 80% then that means that there is a 20%chance of failure. This can be modeled just as easily on one d20 roll. If a character has 4 skill ranks, and a +2 stat bonus, then to get a 80% chance of success the DC needs only to be an 11. I guess my question comes down to why all the dice-robatics? d20 just seems so much more elegant. I have a roll bonus of +6, if somthing has a 90% chance of failure then the DC should be 25.

I think the difference is that you are really concerned about how the stats look on the way there and I am saying "well if the bottom line is X% and you know that, why not do it on a linear roll?"EDIT: Ok so it would help here if I had read your last major post first! Doh!

The simple answer is that people just like their pie made different. Still, I am a minimalist, so rolling lots of dice should be reserverd for special moments like Fireball, Sneak Attacks and Criticals.

 

[Wil]

So you would say that if you needed one success you had over an 80% chance of getting it if you rolled 6d6.

My point is that if the chance of success is 80% then that means that there is a 20%chance of failure. This can be modeled just as easily on one d20 roll. If a character has 4 skill ranks, and a +2 stat bonus, then to get a 80% chance of success the DC needs only to be an 11. I guess my question comes down to why all the dice-robatics? d20 just seems so much more elegant. I have a roll bonus of +6, if somthing has a 90% chance of failure then the DC should be 25.

I think the difference is that you are really concerned about how the stats look on the way there and I am saying "well if the bottom line is X% and you know that, why not do it on a linear roll?"EDIT: Ok so it would help here if I had read your last major post first! Doh!

The simple answer is that people just like their pie made different. Still, I am a minimalist, so rolling lots of dice should be reserverd for special moments like Fireball, Sneak Attacks and Criticals.

Err, while it does boil down to preference, the actual probabilities involved don't work quite like that.

Example:

While you have an 80% chance of rolling 16 or under on 1d20, there is an equal chance of chance of hitting any one number: 5%. So if you need to roll a 16 exactly, there is still just a 5% chance of doing so. On top of that, the die doesn't "remember" your rolls - while the probability of rolling (for example) 20 twice in a row decreases, there is still a 5% chance of rolling that 20 with each throw. In other words, the odds don't change of throwing 20, just the probability.

The same holds true with die pool systems. Each d6 that you throw doesn't "know" what the other d6 is going to be, and the outcome of any one die has no influence on the others. There is a 1 in 6 chance of rolling any one number. But, because you're looking for each one to be higher than the others, it changes the probability of getting a particular number. Hence, when I'm throwing 3d6 and looking for the highest number, there is no longer a straight 16% chance of getting any one. It instead looks like roughly like this:

1 - .5%
2 - 3%
3 - 9%
4 - 17%
5 - 28%
6 - 34%

So the difference winds up being that, yes, while there is an 80% chance of rolling 16 and under in a d20 system and a ~75% of rolling 4 and under in a d6 roll and keep system, there is still a 5% chance of hitting any one number on the d20 but the probabilities vary significantly in the roll and keep. While the first roll on d20 might be a 16, the next one is just as likely to be another 16, or a 1, or a 3. Rolling 3 dice and keeping the highest, there is more of a chance that both rolls will be a 4, or a 3 - the rolls are more predictable than with the single die.

Sorry if it seems like I'm beating this into the ground, it's just I want to make sure that it's clear that while the results are the same, the path they take to get there is very different. Single die systems are much more unpredictable than multiple die systems - even rolling 2d10 as percentile has a small (rather flat) curve behind it. Try rolling, say, 4d6-3 a few hundred times and note the difference in the rolls you achieve than rolling a straight d20 and you'll see what I mean.