Why does 3.X use d3 but not d(any other odd number)?

[Luccan]

Just a thing that hit me while looking through the SRD. The small version of several weapons use 1d3 (down from 1d4), for instance. And weapons that deal 1d3 for medium creatures deal 1d2 for small creatures and shuriken go from 1d2 down to just one point of damage. Yet most other weapons reduce the damage die by an even amount, usually 2 (so 1d10 goes down to 1d8 for a small creature). So why is 1d3 used at all? And since it is used, why aren't other odd die numbers used?

[Fizban]

d2 and d3 are used solely because they let you cram in more "die steps" lower than the d4, for when small dice get smaller and such. They aren't seriously used for much of anything, and are easy to adjudicate. A d5 or a d7 doesn't really add anything but complexity.

[Alent]

It's also easy to "cheap out" on owning a D3 by simply using a D6 with some intuitive division. The same math works with a D10 to get a D5, but I don't think I've ever seen a D14. So there's likely an availability bias at hand, there.

[Luccan]

It's also easy to "cheap out" on owning a D3 by simply using a D6 with some intuitive division. The same math works with a D10 to get a D5, but I don't think I've ever seen a D14. So there's likely an availability bias at hand, there.

Thanks. Really should've thought of that. And then of course, there's no d18 or d22... Though now I really do want to see if there is a major change from instituting a d5. Or I guess you could do d#-1 for small weapon damage but that might screw with numbers too much. Oh well. Thanks again.

[Jormengand]

Some games really do use d5s (Only War and so forth use them) but I don't think that any real D&Dlikes use them. Anything which could be a d5 can usually have been a d6 or d4 anyway.

Notably, prismatic spray and related spells would probably use a d7 if such a die existed or could really be emulated.

[Kurald Galain]

Many computer games based on or inspired by D&D, including MUDs and Nethack, do use d7 and other non-existent dice ("XdY"). On a computer, of course, there's no real reason not to.

[Deophaun]

Notably, prismatic spray and related spells would probably use a d7 if such a die existed or could really be emulated.

What do you mean "if" it existed?

[Hecuba]

If you want to emulate a d7, roll a d8 and reroll on 8. This is generally your only option for emulation if the N in dN has prime factors other than 2, 3, or 5. (Alternately, you can use non-polyhedra but fair dice - which are generally prisims or tops* - or another fair generation option like a set of N cards).

The "reroll above x" option is generally thrown out of mathematical dice discussions because it makes the process non terminating, but in practice most tabes aren't going to care about the possibility of an endless string of 8s.

Using it on a d20 to emulate a d3, in contrast, is substantially more tedious.

*Edit: as the link above points out, edge designs can also work.

[Sian]

Yeah, a long bit of the way it's a question of practicability ... a d3 can be emulated quite easily by tossing d6s, but otherwise its often to tedious to use other odd numbers (sure, you could make an argument for d5s via d10s, but eh) when doing it with real dice ... well you could use it for sure via rolling a higher dice and rerolling rolls higher than x, but as said, tedious.

uncommon/physically impossible dice aren't uncommon in computer games through (usually classical rogue-likes)

[Jormengand]

Using it on a d20 to emulate a d3, in contrast, is substantially more tedious.

...unless you combine it with the technique for using a d(3X, X∈Z+, X>1), or in this case d18 to emulate a d3, in which case you only need to throw out a tenth of the possible results, instead of an eighth of them.

[Hecuba]

Yeah, a long bit of the way it's a question of practicability ... a d3 can be emulated quite easily by tossing d6s, but otherwise its often to tedious to use other odd numbers (sure, you could make an argument for d5s via d10s, but eh) when doing it with real dice ... well you could use it for sure via rolling a higher dice and rerolling rolls higher than x, but as said, tedious.

uncommon/physically impossible dice aren't uncommon in computer games through (usually classical rogue-likes)

It's not a matter of just being odd. It's a mater of factorization. Anything that has only 2, 3, or 5 as prime factors works (assuming you're willing to work with a d10/2 as a d5).

For example, lets say you wanted a d40. That is not the same as 4d10, for the same reason d12 is not the same as 2d6.
But you can get it with a d4 and a d10: (1d4 -1)*10 + d10 is, mathematically, identical to a d40.

To use another odd example, consider d75. That's 3*25, so we can first simplify to (d3-1)*25 + d25. d25 then becomes (d5-1)*5 +d5.
That ultimately is (d3-1)*25+(d5-1)*5+d5.


(Attentive readers and Modrons may notice that we are simply converting to base system that operates in a convenient die, then truncating the highest order of magnitude to cap.
This is the same principle that allows 2 d10s to work as a d100.)


Edit to avoid double post:

...unless you combine it with the technique for using a d(3X, X∈Z+, X>1), or in this case d18 to emulate a d3, in which case you only need to throw out a tenth of the possible results, instead of an eighth of them.

I meant specifically having someone roll a d20 and reroll above 3. d18 for d3, roll over on 19 & 20 is indeed how I would emulate a d3 on a d20 if needed.
But I did have someone try the former when this was first explained to him: I had to explain that it doesn't work well if you're re-rolling out 85% of all rolls.

[Kobold Esq]

For example, lets say you wanted a d40. That is not the same as 4d10, for the same reason d12 is not the same as 2d6.
But you can get it with a d4 and a d10: (1d4 -1)*10 + d10 is, mathematically, identical to a d40.

Given that this is exactly how percentile dice work, I'd hope this would be an intuitive solution to most players.

[Calthropstu]

The only other odd die that could be used is a d5. D6/2 round up is a d3. For a d5 you could use the d10, but there's no d14 etc.
A d15 could be made with a d30, but at that point it's a bit silly.

[Telok]

Am I bad person for wanting to make a player roll ed(PI) now?

[Endarire]

Various nonstandard dice exist in physical form. I own at least one d5 and d7. I've seen a d3. I've rolled friends' dice that were possibly d14s, d18s, d22s, and d100s (very big golf balls). Most people are familiar with the most standard of dice - d6, d4, d20, etc. I'd prefer not to need to pay extra when tabletop gaming to get d27s or d9s or such.

[Luccan]

Various nonstandard dice exist in physical form. I own at least one d5 and d7. I've seen a d3. I've rolled friends' dice that were possibly d14s, d18s, d22s, and d100s (very big golf balls). Most people are familiar with the most standard of dice - d6, d4, d20, etc. I'd prefer not to need to pay extra when tabletop gaming to get d27s or d9s or such.

Yeah, and that is a fair argument. I hadn't really thought of the larger numbers when I made the post. The real problem is that stuff like d14's aren't common, not that a physical d7 isn't common. If dice for the even numbers between 12 and 20 were more common, the rules would likely be different and I wouldn't have even made the thread.

[BloodSnake'sCha]

You can just make a long prism with the number corners(I don't think this is the right translation​, I used Google Translate) equal to the number of options you want +2.
Just roll it like a stick and look at that you got.

If I remember correctly someone was looking for a d54 in here and that was the solution.

[Hecuba]

Various nonstandard dice exist in physical form. I own at least one d5 and d7. I've seen a d3. I've rolled friends' dice that were possibly d14s, d18s, d22s, and d100s (very big golf balls). Most people are familiar with the most standard of dice - d6, d4, d20, etc. I'd prefer not to need to pay extra when tabletop gaming to get d27s or d9s or such.

Part of it is that 4, 6, 8, 12, and 20 have been around for historical lengths of time. The Egyptians were already making regular use of 3 of them when Euclid proved there were 5, and we have evidence of the Babylonians using aag least 6 and 12 as actual die.

There are other fair dice configurations (though the 100 sided golf ball pointedly is not): the d10 is valueable because we generally operate in base 10.

But platonic solids have the benefit of a couple thousand years of use, on top of being symmetric enough that they are intuitivelh fair and easy to roll.

[Goaty14]

I'm still looking for the 1d63 on my sap

[Kurald Galain]

I'm still looking for the 1d63 on my sap

Why would you expect a sap to deal so much dawizard?

[Calthropstu]

Why would you expect a sap to deal so much dawizard?

Don't fall for it... he's getting ready to tell you some sappy story and ask for gp donations.

[Knaight]

Basically any number of faces can be made pretty easily with a spindle die design, so it's not impractical to use them if you want a game to use custom dice. With that said, custom dice are enough of a hassle that they should be best avoided unless there's a really good reason to use them or they're easily mimicked with normal dice.

What do you mean "if" it existed?

For the prismastic X spells in particular a d7 where each face is colored a color of the rainbow could be a lot of fun.

[bahamut920]

I'm still looking for the 1d63 on my sap

It's right here, next to my scorpion whip's d43.

[Deophaun]

Basically any number of faces can be made pretty easily with a spindle die design

Odd numbered dies have the problem of there being no single top face. If you wanted a d63, you'd probably need two 63-sided spindles of different diameters, with the larger split in half while the smaller is numbered, rotated a half-face, and secured between them.

[John Campbell]

Or, y'know, just use software.

[Knaight]

Odd numbered dies have the problem of there being no single top face. If you wanted a d63, you'd probably need two 63-sided spindles of different diameters, with the larger split in half while the smaller is numbered, rotated a half-face, and secured between them.

Odd numbered dice will have a single top edge, so you can just use the d4 technique and label those instead of the middle of the faces.

[Deophaun]

Odd numbered dice will have a single top edge, so you can just use the d4 technique and label those instead of the middle of the faces.

True, but at large numbers that gets very hard to read.

[Hecuba]

Odd numbered dies have the problem of there being no single top face. If you wanted a d63, you'd probably need two 63-sided spindles of different diameters, with the larger split in half while the smaller is numbered, rotated a half-face, and secured between them.

The most elegant solution I've seen for this is a top with a second flatter n-hedron set right above the top sides, (and marked with the numbers) and offset such that the corner of each of the top n-hedron Matches the is positioned below the center of the labeled n-hedron.

I saw a set 10 years ago or so, and didn't see any again until last gencon: apparently there was a woodworker in Ithaca who was making them in small batches until he passed away.

Neither of the dudes would sell them, so I'm considering bribing a local woodworker to make some: I know we have at least two who de bespoke dreadles, so it shouldn't be too much of a stretch.

[Calthropstu]

I have always found the idea of odd dice rather... odd.

[BloodSnake'sCha]

Odd numbered dies have the problem of there being no single top face. If you wanted a d63, you'd probably need two 63-sided spindles of different diameters, with the larger split in half while the smaller is numbered, rotated a half-face, and secured between them.

Or just use a long prism with the number of sides as the number of options and ignore the two tops.