D20 vs 2D10

[Inari]

I'm not a big fan of changing dice types. I know that 2d4 are better than d8 (5 average instead of 4,5) but many of my friends would much rather roll the worse d8, because 2d4 just always seems to get nothing special. You most of the time end up with the same result, which is often not so good (as I DM, I make unconciously all DCs need at least a 10 from the MOST skilled character, meaning that anyone other than the fighter needs at least a 12 to hit the creature, often even a 16).

My handle on critical hits has changed quite a bit over the last half-a-year. At first I just used the standard system, confirming critical hits but always making something bad happen on that one (I just thought something up). This was bad, because the bowstrings of the unluckiest archer I've met always snapped and the group started arguing. It was a fiasco. Later on I started using rules here and there about fumbling, but it always involved the 1, being a fumble.

Then I became smart, and stole the idea a DM of mine had. It involves "confirming" fumbles the same way a critical hit is confirmed. If the fumble is confirmed (hits) then it's just a miss. However if it isn't confirmed, then I confirm with my current fumble chart. There are though a few variables, like for example if it resaults in a fumble on the fumble confirmation then it's a "double" fumble meaning that two fumble results are rolled instead of one. It it's a natural 20, nothing happens (unless a Void point is expanded (yes, it is a Rokugan campaign)) then it's a normal hit. All actual fumbles (thus not confirmed) draw at least an attack of oppertunity.

Speaking from past experience of being hard to understand, I'll give an example:
Average Joe here is a 5th level fighter. His attack bonus is: 5 (BaB) + 3 (str mod) + 1 (focus) + 1 (enchantment) giving him a total of +10 attack bonus. He also has an AC of 16. His current opponent is Sam the ninja who's AC is 25 and attack bonus +8 (ninjas are 1337, aren't they). On joe's round he strikes at Sam. The player then rolls a 1 - a fumble. He rolls again to confirm it. In order not to fumble, he'd have to roll a 15 or higher, thus beating Sam's AC. Joe rolls a 14, just barely not enough, and fumbles. He gets lucky on my fumble chart, and only draws an attack of oppertunity on him.

Sam rolls his AoO, and conviniently gets a 1. In order not to fumble, Sam would have to get a 8 or higher. He rolls, and gets a natural 20, using a void point to make it a normal hit. And thus, he ends Average Joe's life (ninjas are 1337, yes? That katana's 1d10 damage and deadly poison really does make a lovely pair).

Hope that helps.

 

[Cbas10]

I seriously doubt I'm the first one to think about this, but my major house rule is that I don't use D20 for skill checks and attack rolls, but 2d10.
The reason is simple, a D20 is a linear system, while 2d10 follows a Gauss-curve of probability, meaning that you're far more likely to roll in the 8-14 region. Very high rolls and very low rolls are much more unlikely.

Interesting. I can see the merit in this.

- a 1 in 20 chance of fumbling is extremely silly. Imagine surgeons killing one in twenty people (and yes, you can't "take 10" because failure is unfavorable.... )Chances of roling a 2 in 2d10 is very low, and therefor i hear dreadfull gasps of horror when one of my players comes up snakeeyes, they know it's gonna hurt.

Without talking about the use of fumbles or when "Taking 10" applies, I can see what he means. I've seen and played in many games where 20's are always successes and 1's are always simple failures. This 2d10 system helps to show that people who are skilled in something (say, 10 or more ranks) will rarely screw up without outside negative stimuli. Similarly, under rolling a d20, characters simply will succeed at hitting near-impossible AC WILL happen 5% of the time. Not very "impossible" at all.

- The element of chance is still present, but it is lessened. Skill ranks are therefore more important. Skill checks are more predictable, if you know the DC you'll have a better chance of predicting success.

This is the aspect of the 2d10 proposal that I like the most. Most people do not perform skills with very random results. Sometimes they do better or worse than other times, but they generally perform with the same relative quality, given similar circumstances.

-Threat ranges are increased by 1, but you use the original rage for things like keen or improved crit, then add 1 (for example a bow would have a TR of 19-20, but with improved crit it would be 18-20).

I don't know if I would mess with threat ranges much, especially if the 2d10 system is used with Wounds/Vitality mechanics. That would take some playtesting to really comment on.

- Less likely chance of getting a lucky hit, so tough creatures get a lot tougher, watch out for high CR encounters.

Very true; a DM might even have to adjust a CR here and there to reflect this, but this would also require a bit of playtesting.

 

[Skikka]

Critical Success

Here's our group's approach to the problem.

Attack Rolls and Confirm Rolls:
20 on die = 30
1 on die = -9
That keeps the always miss / always hit rule in most cases but, for example, a commoner would not be able to hit the god of swiftness 5% of the time.
If you roll a natural one, you must roll again. If your second roll fails, you critically miss (something bad happens, i.e. you stumble, cut yourself, fall down, etc.)

Skill Checks:
20 on die = 30 if critical success
1 on die = -9 if critical failiure
20 / 1 threatens critical success / failiure.
If confirm roll result is the same, 20 becomes 30 and 1 becomes -9
Additional consequences of criticals can exist if DM chooses.

 

[wally]

Why only 2d10? Why not 5d4 or 10d2 or something?

wally

 

[Conaill]

Yup, 2d10 solves a lot of the problems you get with the flat probability curve 1d20 gives you. 3d6 is a lot better than 2d10. Then again, I tend to think that GURPS is a much better system than D&D too.

 

[Coredump]

Yup, 2d10 solves a lot of the problems you get with the flat probability curve 1d20 gives you. 3d6 is a lot better than 2d10. Then again, I tend to think that GURPS is a much better system than D&D too.

Using 1D20 does **NOT** give a 'flat probability curve'. People, pay attention to what you are rolling.

A 'probability curve' only applies when there are multiple possible 'results'. For the vast majority of situations in D20 games, there are ONLY 2 POSSIBLE RESULTS. Success or failure.

When trying to hit someone, if their AC is 25, and your bonus is 10, you need to roll a 15. Whether you roll a 4 or a 12 or a 7 or a 14 it DOESN"T MATTER. While you use a 20 sided die, there are NOT 20 results, there are only **2** results.(hit-miss) The game mechanic says that you will hit 30% of the time. You can use D% or a D10 or a D20 or Dwhateveryouwant. As long as you keep the 30% constant, the game will work fine. But as soon as you just randomly change (ala to 2D10) you have changed the game considerably, because you will no longer hit 30% of the time, but more like 25% of the time (estimate)

Now, lets say you were 19th level fighter, and killing orcs. And the DM decided that you were so kick-ass that you will just roll to see how many you kill in a round. *then* it matters to have a bell curve, because there are a number of different results. It should be more likely to kill an average number, but this is not the same when there are only 2 possible outcomes.

.

 

[Conaill]

Coredump, you crack me up! Of course 1d20 gives a flat probability distribution. Just as 2d10 gives a triangular one and 3d6 gives a rough bellcurve. Naturally the *outcome* is just binary, but that's not what we're referring to.

You're right that switching from d20 to 2d10 or 3d6 would have a significant impact on the game, because it changes the probabilities. What a lot of us are saying is that a mechanism based on a bellcurve has a lot of benefits over a straight d20. This is particularly true when considering opposed checks, like a simple STR-based armwrestling. With a straight opposed d20 roll, a STR 4 weakling still has a chance of beating a STR 40 giant.

 

[swrushing]

Let me take a crack at this...

whether you roll 2d10 or 1d20 is not going to change anything towards making the RESULTS of the action more or less likely. Sure, the 2d10 will roll 8-14 58% of the time while a 1d20 will roll the same numbers 35% of the time, but the factor then applied is "what does 8-14 mean" in terms of success and failure?

If i want a given task to succeed 25% of the time when the guy has +X skill at it, i will assign a DC wqhich means 25% of the die roll results fall in there. if the skill is +5, then i will assign a Dc of 21 for 1d20 and a DC of 19-20 for 2d10 (21%-28%.)

Changing the die rolls does not affect the probabilities of the results, unless you just ignore the DC changes that should accompany the shift in die rolls. if you do that, then you are buying a ton of basically random shifts in DCs you probably never had a problem with before.

It would be MUCH MUCH easier to just go thru and adjust the DCs you dont like and keep the very manageable and understandable 1d20.

By far the biggest problem with 2d10 or any multidie roll is that adjustments become variable in impact. if i say rain is causing a -2, the for some people that means rain is costing them 19 rolls out of 100, almos 20% but for others it is costing them only 9 rolls out of 100. Moreover, who gets 19 and who gets 9 is not readily obvious... its not "the skilled guys lose less" or even "the skilled guys lose more" its who is farthest from the "need a 10" crowd lose less. This applies of course to Dc assignment too. A Dc +5 higher might mean 54 fewewr rolls succeed out of 100 or it might mean 20 fewer rolls out of 100.

For me, I can take a flat d20, know every +1 is 5% and set the DCs based on knowing that each +1 DC or +1 modifier is adjusting the odds by 1 in 20 and then I know what i am doing. if i want the guys with moderate rolls to succeed more, say 60% vs 35%, i lower the Dc so that they will succeed on a roll of 9+, 60%.

2d10 is just d100 with uneven mapping of roll to result. 3d6 is just d216 with even more uneven apping between die roll and result. its the decision of "what rolls equate to success and failure" that then gives these systems the feel of more predictability. You can get that predictability without the uneven modifiers by just adjusting your DCs with flat d20.

Coredump, you crack me up! Of course 1d20 gives a flat probability distribution. Just as 2d10 gives a triangular one and 3d6 gives a rough bellcurve. Naturally the *outcome* is just binary, but that's not what we're referring to.

You're right that switching from d20 to 2d10 or 3d6 would have a significant impact on the game, because it changes the probabilities. What a lot of us are saying is that a mechanism based on a bellcurve has a lot of benefits over a straight d20. This is particularly true when considering opposed checks, like a simple STR-based armwrestling. With a straight opposed d20 roll, a STR 4 weakling still has a chance of beating a STR 40 giant.

 

[Coredump]

You're right that switching from d20 to 2d10 or 3d6 would have a significant impact on the game, because it changes the probabilities. What a lot of us are saying is that a mechanism based on a bellcurve has a lot of benefits over a straight d20. This is particularly true when considering opposed checks, like a simple STR-based armwrestling. With a straight opposed d20 roll, a STR 4 weakling still has a chance of beating a STR 40 giant.

There is a 1:400 chance that the weakling will win. But I understand your point. My point is that this type of situation is rare and may be better handled by something less....drastic than changing the basis the game is designed upon.

For instance, in your example, I think the real problem is the fact that the bonuses are so 'small' for having a high strength. Even considering a 10 and 18 strength, it seems (to me) to be way too easy for the 10 to beat the 18 in something. This is not a problem with the dice, but rather with the bonus. In an opposed check, I would double the bonus for each one. Now the same 1:400 chance exists for the Str4 weakling when competing with a Str18 person; which seems more realistic. Again, your complaint is with the percentages, notwith the dice used.

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. (except by roleplaying 'closeness')
Now, somethings, like jumping, may benefit from more of a curve, because then there are more possible results. But again, these are relatively rare.

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.

There are some problems, but there are better ways of fixing those than changing the dice.

.

 

[Coredump]

Let me take a crack at this...

WOW. This is one of the best posts I have ever seen. I only wish I had thought to say it as well. Thanks for taking a crack...