D20 vs 2D10

[Conaill]

The other point, which is made SO much better by swrushing, is that you will need to change just about every DC and AC to deal with the change in percentage chance to succeed. But most don't even mention that.

Only if you want the end result to stay exactly the same. What's the point changing the dice you roll if you're going to map it back to what you would have gotten with a d20 anyway?

Yes, using 2d10 or 3d6 will make dice rolls much more "reliable" in the sense that you tend to get results within the middle of the range. Forget about doing a straight attack on that AC 30 monster if your attack bonus is only a +15. But finding some situational modifiers to improve your chances suddenly become much more important. And that puny +1 sword now looks a lot more attractive as well.

 

[swrushing]

[/QUOTE]

Only if you want the end result to stay exactly the same. What's the point changing the dice you roll if you're going to map it back to what you would have gotten with a d20 anyway?

The point is to scrutinize which DCs you want to change and which ones you do not. making a blind change of dice without looking at the DCs means you are likely changing some you did not want to. Were you INTEITNIONALLY deciding to change the chance of untrained spotting scries down? Were you intentionally meaning to drop the chance to make an untrianed heal check to stabilize someone from 30% to 21%?

Note i am not arguing whether these were good or bad but simply, did you actually stop and think of these odss for these specific things and decide "yeah they need to go down" or did the odds for these just get caught up in the "i want fewer crits" wash and get changed unintentionally?

If you are not going to look at Dcs and dont mind basically random changes to DCs creeping in, then hey, your game, your problems.

On the other hand, if you are going to look at the Dcs and decide which ones need changing and which ones dont and which way the ones that do should go, you can do that a lot easier by simply changing the DCs and keeping d20.

Yes, using 2d10 or 3d6 will make dice rolls much more "reliable" in the sense that you tend to get results within the middle of the range.

Which is meaningless to a game until you start mapping those 8-14s to "success/fail".

Forget about doing a straight attack on that AC 30 monster if your attack bonus is only a +15.

Need a 15 on the roll, eh?

d20 makes that 30%
D100-2d10 makes that 21% or a little better odds than the equivalent of needing a 17.

is it really in your experience that the difference between "need a 15" and "need a 17" is sizeable enough to make it "forget about it"?

Or do you just not know the odds results on 2d10?

If i wanted that monster to be that 9% or so tougher to hit, why not just make it AC32 in the first place and not make me reconsider every Dc and Ac in the book?

But finding some situational modifiers to improve your chances suddenly become much more important. And that puny +1 sword now looks a lot more attractive as well.

You apparently did not pay attention before.

the value of +1 is constant in d20, +5%, one more hit in 20.

in 2d10, its variable. it could be as little as +1% or it could be as much as +10%. You dont know where in the mix you are getting it until you know both the Dc and the bonus.

In your own example above, the +1 sword would add +5% to the d20 roll, but the 2d10 roll would see it add +7% so the "much much more important" and "a lot more attarctive" is the difference between 5 more hits out of 100 and 7 more hits out of 100 in your own example.

2 more hits out of 100 seems like a whole lot less than earning "mucn mucn more important" acclaim.

 

[ichabod]

But I stand with my statement, probability is based on results, and there is not 'distribution curve' if there are only two outcomes. Even in your example, there is only win-lose; no curve.

Rampant Pedantry Warning

There is no curve in any case. In terms of probability, two outcomes and twenty outcomes fall into the same class: discrete (univariate) probability distributions. Mathematically, curves are continuous. Thus a probability "curve" would only apply to a continuous probability distribution, assuming it was even a well defined term in probability, which it isn't. And which you'd never get from rolling a finite number of dice. Generally in probability you talk about cumulative distributions, which are defined for both discrete and continuous cases, with the discrete case including both two result cases and three or more result cases.

I think even if you want to provide some discrete curve definition, you still can't distinguish between two and twenty. In both cases, you can interpret them as numerical values (and in fact you usually would for analysis), and thus have a change across results. Only if there is one outcome does this fall apart, but then there's not much point in talking about probability.

Even if you want to come up with some funky definition of discrete curve that includes three or more results but not two results, it becomes a question of where you define the probability. If you define it as the probability that an attacker with a +6 bonus will hit a defender with AC 17 (a function of two set values), that's two results. If you define it as the probability that an attacker will hit a defender (a function of two variable values) then you have to think in terms of at least 400 values (20 if you first map the two values to one, which is reasonable in this case).

End Rampant Pedantry

 

[Conaill]

ichabod: "rampant pedantry" indeed! I just knew someone was going to get a conniption because I inadvertently used "curve" to talk about a discrete distribution. Not that this has any importance whatsoever for the current discussion, of course.

swrushing and Coredump: I get the feeling neither of you have ever played a non-d20 game? Having a straight 5% per point on your roll is convenient, but it's hardly essential to having a good game mechanic. Personally, I like the fact that the probabilities drop off much faster at the extremes. And no, it's not just an issue of getting more or less crits. Yes, it's non-linear. That's the whole point!

As for the numbers you quoted, I was thinking of a 15 on 3d6, which is quite a bit less than 21%, IIRC (let's see... 4.6% actually). Nice if you get it, but not something you would want to waste your time on if you have other options. Also, 3d6 has a standard deviation which is about half of 1d20, so a +1 bonus will on average have twice as much effect under 3d6 as under 1d20. The reduced spread of values is less important to me than the nonlinear distribution, but this does give a reasonable heuristic for adjusting values for bonuses and penalties.

These are just quick notes. Real life is getting in the way, so I can't afford to continue this discussion in great length. I suggest you try some game systems that use a 3d6 mechanic - god knows there's plenty of those. As for using 3d6 in D&D, you're probably right that it might be more effort than it's worth. And I mean that in a number of ways: (1) IMHO, D&D isn't worth much to begin with ; (2) you will need to do some playtesting to resolve some of the consequences of this change; (3) it's probably better to switch to a different system altogether (see (1)); (4) the main benefit of the d20 system is that it's a standard, if you start tinkering with the base rules, you lose that benefit; ...

 

[swrushing]

[/QUOTE]

swrushing and Coredump: I get the feeling neither of you have ever played a non-d20 game?

and i get the feeling you are willing to base assumptions on no pertinent info.

Having a straight 5% per point on your roll is convenient, but it's hardly essential to having a good game mechanic. Personally, I like the fact that the probabilities drop off much faster at the extremes. And no, it's not just an issue of getting more or less crits. Yes, it's non-linear. That's the whole point!

I prefer knowing how my modifier affects things when i apply it and having it apply a consistent impact. Sure, i dont have to know what I am saying to runa a good game, but it helps.

As for the numbers you quoted, I was thinking of a 15 on 3d6, which is quite a bit less than 21%, IIRC (let's see... 4.6% actually). Nice if you get it, but not something you would want to waste your time on if you have other options. Also, 3d6 has a standard deviation which is about half of 1d20, so a +1 bonus will on average have twice as much effect under 3d6 as under 1d20. The reduced spread of values is less important to me than the nonlinear distribution, but this does give a reasonable heuristic for adjusting values for bonuses and penalties.

Well i have been focusing on the topic of the thread, comparing d20 vs 2d10. i got way to many years of HERO under the bridge to get worked up about 3d6 again.

These are just quick notes. Real life is getting in the way, so I can't afford to continue this discussion in great length. I suggest you try some game systems that use a 3d6 mechanic - god knows there's plenty of those.

I suggest you stick to assumptions you have any iotas of evidence to make. Telling someone who started with HERO in the early 80's and has Gmed much more HERO than any other game including all the DNDs combined he should try some 3d6 games is not a way to make your point.

Contrary to your assumption, my preference for a single die resolution is not born of not having experience with others. Its born of having a lot of experience with others (3d6 in particular) and rejecting them as not as useful.

As for using 3d6 in D&D, you're probably right that it might be more effort than it's worth. And I mean that in a number of ways: (1) IMHO, D&D isn't worth much to begin with ; (2) you will need to do some playtesting to resolve some of the consequences of this change; (3) it's probably better to switch to a different system altogether (see (1)); (4) the main benefit of the d20 system is that it's a standard, if you start tinkering with the base rules, you lose that benefit; ..

ahhh, OK, and here i thought you were actually involving in the discussion, instead of just taking the opportunity to swing in some good ole fashioned anti-dnd nonsense.

My bad. Wont happen again.

But it does make your inbred presumption of "must never have tried anything else" more understandable.

Enjoy your games.

 

[swrushing]

for those not here for errant DND bashing, and actually considering a 2d10 add-em pyramid probability system...

do you really want to explain to your players that your chosen resolution system can never do 50/50?

just a thought.

 

[timfire]

for those not here for errant DND bashing, and actually considering a 2d10 add-em pyramid probability system...

do you really want to explain to your players that your chosen resolution system can never do 50/50?

just a thought.

That's only half true. In the case of skill vs. DC, then yes, the closest you get is 55/45. But in a contested roll between two characters with equal modifiers, then the outcome is 50/50.

I personally like the idea of 2d10 (I heard of that house rule before this thread), and as soon as my group starts up again we're going to try out a 2d10 mechanic for a while.

 

[DrZombie]

First of all, thank you guys for responding. There are a lot of interesting approaches to my first post, wich, I must say, was a bit vague due to chronic sleep deprivation.


Second : WHOA, hold your horses guys, this is not something to start swearing at each other about. It's just a game, and you're not a better person for liking or disliking a certain number of dice, or a roleplaying system.


As to previous posts:




Quote: Originally Posted by Conaill

Only if you want the end result to stay exactly the same. What's the point changing the dice you roll if you're going to map it back to what you would have gotten with a d20 anyway?


The point is to scrutinize which DCs you want to change and which ones you do not. making a blind change of dice without looking at the DCs means you are likely changing some you did not want to. Were you INTEITNIONALLY deciding to change the chance of untrained spotting scries down? Were you intentionally meaning to drop the chance to make an untrianed heal check to stabilize someone from 30% to 21%?


I agree that basically all of the DC's are affected. But, in my game, that doesn't appear to be a problem (yet?). The reason being,it works two ways, if it's harder for them it's harder for the "bad guys" as well.
The major reason for using it (for me) is to "dampen down" the extreme ends of dice rolling.( i would have quoted whoever came up with that phrase, but i lost it in the texts).

And, face it guys, chances of anyone untrained stabilising a mortally wounded person are just a bit less then 30% anyways, or I'd be out of a job.

Cheers