Playing With Different Dice

[Apollonian]

I'd like percentile dice, but the amount of heavy lifting required to do the switch has always scared me off. I'm also not sure it would handle scaling.

[Nightharrow]

Originally Posted by ericbsmith

I wrote a couple articles on alternatives to d6 on my website GURPSLand. Polyhedral GURPS goes into detail about replacing the d6 with various polyhedrons while Cards as Dice discusses using a deck of cards as dice (which could be useful in situations where flat surfaces for rolling dice aren't handy). Both have details on the different bell curves that you get and compare them to d6 bell curves.

Thanks for the links, I took a look. Didn't see anything on alternatives to 3d+ which all have a normal distribution. My primary goal was to find the middle ground between a normal distribution and a flat line, which means 2dx. I could see a normal distribution still achieving the desired result, I just think it would take larger dice. Maybe even 3d10 would be okay. All of this would required a fairly large reword of the rules/scaling though.

[Nightharrow]

Originally Posted by Apollonian

I'd like percentile dice, but the amount of heavy lifting required to do the switch has always scared me off. I'm also not sure it would handle scaling.

Do you just prefer the flat distribution?

Agree that the switch would be a significant issue. Thats actually why we decided to go with 2d10, doesn't take much to convert it.

[edk926]

My last incomplete attempt at a homebrew was 2d10. I liked how it distributed. I never thought about converting GURPS into other dies. It doesn't seem like it would take too much of a change for many things, although you are going to have to rework damage quite a bit unless you keep damage d6.

[AlexanderHowl]

One problem with 2d10 versus 3d6 is that the probabilities are off when translating GURPS. The average roll for 2d10 is 11, meaning a 55% chance at failure for someone with a '10'. The average roll for 3d6 is 10.5, meaning a 50% chance of failure for someone with a '10'. 3- and 18+ are much more common as well (3% chance for 3- rather than a 0.5% chance and 6% chances for a 18+ rather than a 0.5% chance).

[Apollonian]

Originally Posted by Nightharrow

Do you just prefer the flat distribution

No, I like the ease of communication. What's your chance of succeeding? 55%. Whereas if you use a curve, then you need to actually work out the probabilities. And modifying the difficulty is very simple for the GM. +/- 10% is always going to change the likelihood by 10%, not 12.5% (10+1) or 1.38% (16+1).

On the down side, a flat distribution doesn't allow for stacking modifiers to have diminishing returns as elegantly as a curve does. Skill 10 + 6 from modifiers is great; another +2 is not nearly as useful as the first +2. (Obviously, you can use it to counteract negative modifiers, but...)

[ericthered]

You'll want to expand the range of character scores, but I think that's part of your goal.



Things where 2d10 will work fairly simply:

• Skill rolls
• Attribute rolls
• Reaction rolls

Things where 2d10 will technically work but you may not want to for other reasons:

• Hit location tables
• Other 3d6 tables (mostly related to tech and magic)

Numbers whose positions you'll want to adjust:

• Self-control rolls
• Malfunction scores
• The rule of 16 (or 14).
• Critical Hit ranges

And of course damage isn't really based on 3d6 in the first place.

[whswhs]

Originally Posted by Apollonian

No, I like the ease of communication. What's your chance of succeeding? 55%. Whereas if you use a curve, then you need to actually work out the probabilities.

Or look them up, or memorize them. 1/216, 4/216, 10/216, 20/216, 35/216, 56/216, 81/216, 108/216, 135/216, 160/216, 181/216, 196/216, 206/216, 212/216, 215/216, 216/216.

[Anaraxes]

Originally Posted by whswhs

Or look them up, or memorize them.

Or just develop a feel through play -- which is the closest to the character's actual perceptions. Outside of a few fictional characters, half of whom are making jokes, not many adventurers sit around calculating whether they've got a 37% or 48% chance of success. (They're much more likely to snap "Never tell me the odds" and blow a hero point if they have to...) It doesn't take long to develop the sense that a 6 is hard to beat, a 10 is fifty-fifty, and 13 is getting pretty likely.

Having a consistent change in probability for a particular modifier regardless of skill or circumstances is actually a detriment to verisimilitude. A novice will benefit from a small "+1" aid much more than an expert. And a little push helps change the outcome a lot more when things are balanced than when they're already askew and toppling. That same small nudge isn't then nearly as likely to make things come out your way -- you need to round up a lot of help. If you have a flat result from the die roll, then to get that effect you have to modify the modifiers on the fly, listing them all in a 2D table or run then through some formula to give the total diminishing returns. Rolling a bell curve means you can use flat modifiers while the result curve automatically adjusts how significant that +1 actually is.

[Nightharrow]

Originally Posted by Apollonian

No, I like the ease of communication. What's your chance of succeeding? 55%. Whereas if you use a curve, then you need to actually work out the probabilities. And modifying the difficulty is very simple for the GM. +/- 10% is always going to change the likelihood by 10%, not 12.5% (10+1) or 1.38% (16+1).

On the down side, a flat distribution doesn't allow for stacking modifiers to have diminishing returns as elegantly as a curve does. Skill 10 + 6 from modifiers is great; another +2 is not nearly as useful as the first +2. (Obviously, you can use it to counteract negative modifiers, but...)

Thanks for the reply, those are very good points. I guess if you went that direction you could just change every +/-x into +/-x% Does make it intuitive and easy to understand. If you every decide to give this a shot, you should @me, I would be curious to read what you come up with.