What is this for.
Don't lie to me.
>>43892501
For sexual. Like everything else.
Playing 2HG mtg.
To generate a random integer between 1 and 30, inclusive.
A lot of 1980s RPGs had random tables with 30 possible results.
Back in "the day", required die sets weren't as standardized.
>>43892501
For throwing at unruly players
>>43892610
Are there other such random integer generators?
Its a novelty really unless you wanted increments of 3.3333333333333333333333333333333%
>>43892501
For playing Dungeon Crawl Classics, a game which loves its weird die sizes.
There are also the AX2 products, like their d30 Sandbox Compendium, which has a bunch of tables based on the d30.
>>43892768
And AFTERMATH!, apparently.
>>43892501
I use it in CoC. I someone empties a tommy gun, and they get a hit, i roll to see how many rounds hit something.
My friends and I used it in our DnD games for years. Once per session, a player can roll a d30 to attempt a task that would normally be impossible for his character to perform, but is not unrealistic or game-breaking; a roll of 30 lets it happen. It's basically a luck die, but there were some incredibly awesome moments when we used it, since it was extremely rare for any of us to land on 30.
Speaking of weird dice.
I have a d12, except, instead of an 11 and 12, it has a 0 and crown on it (I also have another one with a star on it instead of a crown) I have no idea what it is used for except as a d12-1, but it sure is interesting.
I presume it is from some oddly specific game, but have never found out what it is.
>>43892501
Hit location.
Rolled 11 (1d30)
>>43892658
>d30
>Are there other such random integer generators?
Yes. Your magic internet machine routinely uses them. As does the 4chan dice-device, as I have demonstrated.
Also, you can do 1-30 with a d10 and a d6. The d10 is the ones place (range 1-10), while the d6 is treated as d3-1 for the tens place (range 0-2). It's basically a reduced version of percentile dice.
>>43896950
6d6-6 actually produces a better spread though.
(this was how early japanese d&d translations handled d30 rolls btw, due to the japanese not having heard of plato and the platonic solids that form all other good dice)
>>43897116
>Better
Depends on what you want.
For one thing 6d6-6 produces a number between 0 and 30, not 1 and 30. And if you want to assign EQUAL chance to all 30 numbers, (perhaps rolling on a random table like >>43896946 ), the bell curve doesn't do what you want to in that case.
A way to get an actual d30 with only d6s is (1d6-1)*5 +1d6, reroll all 6s on the second dice only.
>>43896950
>>43897116
>>43897263
Would 3d10 really be THAT bad?
>>43898149
Number from 3 to 30.
>>43898149
Range is 3-30 for starters so no 1s or 2s, and the average is 16.5 which is above the normal average of a 1d30 of 15.5 on top of that numbers in the average range will be a lot more common because of the multiple dice being used.
>>43898149
See for yourself.
Rolled 6 (1d30)
>>43896946
>rolling for 12
>>43897116
>due to the japanese not having heard of plato and the platonic solids that form all other good dice
Rolled 22 (1d30)
>>43896946
roll
>>43898149
If you have the d10's, roll one d10, it's the ones place. Roll a second, 1-3 it's +0, 4-6:+10, 7-9:+20, 10: reroll
Gives exactly even 1-30 spread.
>>43892501
Random tables with 30 items on them.
>>43892501
Random day of the month, assuming you're in a fantasy world that has the same number of days in each month.
>>43896946
Oh Christ, I had a GM that used that chart for EVERY combat roll.
>Every time we rolled a 12, the group experienced sympathy pain.
>>43897116
>6d6-6
0-30, and bell curve. A proper d30 is 1-30 with uniform distribution (all results have 1/30 chance to happen).
>43898149
>3d10
Range is 3-30, probability increases around the mean. Not a uniform distribution.
>>43892501
Based statisticians of /tg/, what dice could be approximated/emulated using coin flips? I mean, with a close enough probability distribution to the real thing, that is. Like, say, can a d20 be emulated using coin flips?
Just wondering if there could be a DnD-like system that diceless people can use/poorfag tabletop version.
>>43901674
It's kind of clunky, but you can do it.
>>43901753
Oh, and if you want to speed it up slightly, you can use multiple coins and just declare their columns. Like a penny for the first coin, a nickel for the second, dime for the third, and quarter for the last one.
>>43901805
>>43901753
Ooh, nice, that'd come in handy. Muchas gracias.
>>43901753
>>43901805
Well, now if I get stranded on an island with only a pocket of change and a buddy, we won't be bored.
>>43892658
Aside from the "standard" d4, d6, d8, d10, d12, and d20, there are also a rather odd d5, and several developments of the double cone d10 into a d14 and d16, as well as a state lottery-related d34. You can get pyramid-sided d6 variants that are a d24.
Then there is the Zocchihedron, a d100 only slightly less likely to roll off the tale than a golf ball.
>>43898427
>that site
thanks based anon
>>43898149
Math is hard
>>43892501
So you can roll a number between 1 and 30?
I was unaware that people didn't know the purpose of dice. Seems like a pretty self explanatory concept.
![48178599bc[1].png](https://i.imgur.com/rBPNs0N.png "48178599bc[1].png")
>>43898427
>that site
Bruh, how does this shit even work.
>>43903043
It has complete documentation in the links on the sidebar.
What are you trying to make it do?
>>43903094
I was trying to make it do:
((1d10)d(1d10))d((1d10)d(1d10))
Apparently that takes too much time to compute.
>>43893327
makes sense
![RCaet0m[1].jpg](https://i.imgur.com/b5tbHse.jpg "RCaet0m[1].jpg")
>>43903043
I could be playing with this shit for hours
>>43903352
You know what I would legitimately pay money for? A service that lets you draw out a range of dice values, and their distribution, and will attempt to generate a (reasonably simple) dice function that matches it to a decent degree.
>>43903160
>((1d10)d(1d10))d((1d10)d(1d10))
>Apparently that takes too much time to compute.
Yeah, you don't want to crash the poor guy's server. I wonder if he'll ever come out with a downloadable version that runs off the user's computer instead of his.
Though it is entirely possible to make programs that do this stuff. I know I made some relatively basic functions for shadowrun diepools. Maybe someone already has something for it.
I was waiting for an appropriate thread to ask this question.
What would be the best possible emulation of a d16 roll using the six standard dice?
I don't think 1d10 + 1d6 would be too accurate, and neither would 2d8 or 1d12 + 1d4. But which one of these is the least inaccurate? And furthermore, is there a better option that I'm just not thinking of?
>>43904539
Roll 1d20, reroll if you get 17 or higher.
>>43904539
(1d4-1)*4+1d4
>>43899452
>rolls 12
>collective table groan
>>43904626
Nailed it. Distribution is exactly the same. Also only needs a pair of d4s.
>>43905132
How intriguing. Would it be possibly for a mathematically-minded anon to be able figure out the probabilities of rolling the various sides of any of the Archimedean solids?
>>43905132
16 is an easy one because of how it is a multiple of two common types of dice (or actually, in this case, the same die twice)
The harder shit is like, the higher prime numbers, which really can't be done without picking a slightly bigger value and having all values over the die size be a reroll. (Like, say, a d37, would be best done as a d40 (1d4-1)*10+1d10, but you reroll BOTH (this is important to not skew the odds), if you rolled a 38, 39, or 40)
>>43904626
B-but how do you figure this shit out, math wizard?
>>43896946
Roll
>>43892501
Time displaced Cyberpunk Sniper in the world of DnD. Roll for damage
Man I wish I remembered my Inductive Reasoning college course involving mutually exclusive variables...
Then I could work game probability down to a T .
Like how do you measure dice probability for games that involve dice pool "hits" rather than a sum value?
>>43896946
>Where did the bad man touch you timmy?
Roll for bad touch
>>43905558
that got a chuckle out of me.
Rolled 7 (1d30)
>>43905558
>>43892640
They were selling packets, including a d4, d6, d8, d10, d12, and d20 in the the early 80s. They were tiny low-impact crap, and TSR drove the sale of them. Anything else was just to be cool and drive sales.
d2, roll d6/3
d3, roll d6/2
d4, roll d8/2; leave the caltrops at home
d5, roll d10/2
d6
d8
d10
d12
d16, roll d8 and a d2
d20
d24, roll d8 and a d3
d30, roll d10 and a d3
...
>>43892934
Yes, but they really wanted you just to have d6s and d20s.
AFTERMATH! was a mess of a system where play could ground to a halt, and three books would be searched because you knew that there was a way to resolve a stick of dynamite being thrown into a room involving concussive force and over-pressure.
>>43905307
Factoring.
Basically you take the number you want and factor out the numbers of common dice types. A lot of numbers (or at least the ones we care about) can be reduced to these simple values. Once you have those factored dice types, you then basically apply the same principles as you do when you roll a d100 with 2d10. You have one die be "big" (the tens digit d10) and another die be "small" (the ones digit d10). You then multiply the big die, by the size of the small die, this is done so that no combination of the two dice will make the same result. Then you add the "big" and "small" dice together to get your final result.
However, this, by itself, creates an error. to take our d100 example, if we were to follow the steps laid out, and were to roll a 1 on both dice, this would get 11 (1*10 + 1), instead of 1, and likewise, if we were to roll a 10 on the "big" die, that's 100 right there before adding the little die.
To compensate for this, we subtract 1 from the result of the big die before multiplying. This means that the minimum result of the big dice is 0, and the maximum result the big die can get on its own is the maximum of the small die can get short of being our maximum result.
A few other neat things:
Which die is big and which is small doesn't matter. Take the d40 I proposed >>43905241
(1d4-1)*10+1d10 and (1d10-1)*4+1d4 produce the same probability results. (See http://anydice.com/program/7213 )
You can also daisy chain this formula, like, say I want a d64, and I only have d4s. If I use my d16 formula as my "big" dice, I can use a third d4 as my little dice ( d64 = (1d16-1)*4 + 1d4 =((1d4-1)*4+1d4-1)*4+1d4 ) (see http://anydice.com/program/7214 )For a d64 2 d8s would be better since you can do that in 2 dice.
Hopefully that makes sense to someone other than me.
>>43896946
>Hit location.
>shows 20 locations
Besides, Aftermath used a percentage system to determine hit location. Facing and aiming then modified the location.
>>43897116
>the japanese not having heard of plato and the platonic solids that form all other good dice
Amusingly, all the d20s being made in the early days were being made by JSA.
>>43905553
>Like how do you measure dice probability for games that involve dice pool "hits" rather than a sum value?
tl;dr: It involves binomial distribution functions.
In anydice: Let's say for Shadowrun, where you roll d6s and count fives and sixes, you put in "output [count {5,6} in Xd6]" where X is the size of the diepool. You can do opposed tests by taking one and subtracting it from another, though I haven't yet found a way to make it consolidate results less than zero. There are ways to do exploding dice, but I'm not remembering them right now. Also note: Anydice times you out if you try to do diepools higher than 29, and it won't simulate more than 20 explosions per die.
In Excel, and also programming languages: I use cumulative and partial binomial distributions, then I check for errors against pic related (don't kill yourself if it's off by like .2 percentage points; that's probably just a difference in the binomial distribution function being used). Pic related is from someone on reddit called dev_grrl (or something like that) for shadowrun 5e diepools. In a language called R, I've managed to make functions for Shadowrun success-tests, opposed tests, and extended tests.
Note for programming: If you want to, yes, you can just simulate the rolls like 100,000 times and approximate it. Personally though, I've been trying to get the exact probabilities, since the model is already given.
I believe the 1st edition AD&D DMG called for its use in determining the day of the month. Forgotten Realms (I think) has 30 days in every month, plus a 5 day period. Krynn has a 360 day year.
For some reason, I recall it also had something to do with barbarians or monks in some early edition. My memory of that age is fuzzy. Or maybe it was a treasure table... I know for sure that the d24 was used in those. You'd occasionally run across them in published modules, too. And I think there was an entire book of d30 tables somewhere along the line....
It was also useful as a substitute for the d3.
>>43904539
Roll a d20. If the result is 17-20, roll again until it is not.
>>43905697
so essentially you're using base 4 with the d4 and so on?
>>43892501
Sex. Sweaty nerdy sex.
>>43905168
Stop.
>>43905456
Around 6d4+1 damage, calculating a 7.62C sniper rifle coming from the Cyberpunk 2020 universe to the D&D universe (Where a 1d6 damage dagger is a 1d4 damage dagger)
>>43909390
http://anydice.com/program/7218
>>43908688
Kinda, that's really just how shit happened to work out for a d16.
Really it is just breaking the number down into factors equal to the dice size.
For example, lets take a d600, we want to factor that out into dice we can actually roll. 600 is 6 times 100. We choose 6 as our "big" number because multiplying by 100 is much easier than multiplying by 6. So we get (1d6-1)*100+1d100.
Now, most of us don't have those golf ball d100s, so we are actually rolling 2d10 like (1d10-1)*10+1d10 **
Thus, since 1d100= (1d10-1)*10+1d10
We can say that 1d600 = (1d6-1)*100+1d100 = (1d6-1)*100+(1d10-1)*10+1d10
We can also simplify the math a bit if we would like to
(1d6-1)*100+(1d10-1)*10+1d10 =
1d6*100 -100 +1d10*10-10+1d10 =
1d6*100 + 1d10*10 + 1d10 - 110
Which is technically less operations, but might be less intuitive to some people than just reading every die but the final die as 1 less.
**Technically, most of us are actually rolling (1d10-1)*10+1d10-1, since we we treat the 0 on both dice as a 0, and just have the special rule that rolling double 0 equals 100, but really it doesn't change the probability at all.
It's also used in the game Morton's List.
Unfortunately, no print copies of the game exist anymore. I heard all the remaining copies were sacrificed somewhere in Mexico in the end of 2012.
>>43909318
Why, afraid you might like it?
>>43910777
Yes.
>>43905767
Are you... are you stupid?
>>43912693
No, just big boned.
>>43901753
All all statisticians this bad at application?
>>43913386
Most big-boned people can count.
>>43913832
Elaborate.
What did he do wrong, and how would you fix it?
Because the chart looks fine to me
>>43913832
>>43913899
>what did coin-flip anon do wrong?
Seconding. It might take 5 coin-flips to emulate a d20 (or more if you have to re-roll), but aside from that the probabilities seem to work out.
>>43913899
>>43914029
Not sure what >>43913832 was talking about, but since the probabilities seem manifestly correct, they were probably talking about the methodology used, ie., there are some timesaving methods that could be used rather than just brute forcing it like that.
>>43904539
d8 + a coin toss. Heads, add 0. Tails, add 8.
>>43913870
nice meme
>>43912366
>Yes
Laughed for a hot minute
>>43914173
Oh, so basically doing shit like this instead of the d5 presented in the chart to reduce the frequency of rerolls?
I mean, I guess that likely does work out to be less coinflips over all, but honestly the chart the first anon provided is much simpler (even ignoring my shitty paint skills), and thus easier to remember and derive for the average person.
>>43914267
>>43914531
that's also a good meme
>>43914410
>average
>usage it like that in a statistics thread
>>43914556
You know what Imean
>>43904005
>Yeah, you don't want to crash the poor guy's server.
Thank you, but it runs on AWS.
>I wonder if he'll ever come out with a downloadable version that runs off the user's computer instead of his.
Probably not, since everyone has internet in their pocket these days. It is web based for that reason.
>>43914540
>>43914671
epic
>>43914684
>>43914707
http://twitter.com/dril/status/458950527616024576
>>43914267
>>43914531
>>43914540
>>43914671
>>43914684
>>43914707
>>43914738
Just fuck already.
>>43914784
nice /tg/ meme, /b/ro.
>>43914267
How is this a meme?
>>43915391
Everything is a meme.
>>43915391
Everything is a meme.
Rolled 9 (1d30)
>>43896946
rollin
>>43915404
>>43915414
Then your use of "meme" as a descriptor is meaningless.
>>43915513
No, anon. You are the memes.
>>43915533
THE MEMES, JACK.
>>43898427
i dont get it desu, it should be proportional too
i always sucked at maths tho
>>43912693
The image presented showed 20 hit locations; not something that a 30-sided die would be used for. The rules use a percentage die to determine hit location anyway. Since this is a role-playing board, I will assume that you are playing a moron.
>>43916413
Look again. The image shows 30 hit locations; they are numbered in the image, 1 to 30.
In case you still don't see it, the dummy's left hand is #30 (its left, but our right).
>>43916533
welp, it's official. I'm an idiot and I apologize.
>>43916602
>>43916533
epic samefag bro
>>43916666
Nice meme.
>>43916923
you are the meme
>>43916933
We are all the meme.
>>43913832
It's not my image, dude.
>>43897116
>>43898149
>>43898427
>>43901805
>>43901753
You can also stop flipping once you've locked down to some result.
For example if you are flipping for d20, and get heads then tails, you are locked to 17 and higher. If that's enough to pass a skillcheck, you pass. (Although still need to check for 20 for combat crits ofc)
And on the other side, two times tails means 8 and lower, if that's not enough to pass, you can stop.
>mfw this thread
>>43905168
using wolfram alpha
if the goal is to get the probabilities of the shapes themselves, i guess you could calculate the areas of the respective regular shapes, and then multiply those shapes by the times it appears, additionally you need to do this to get the total area if you plan on putting different numbers on each face. Those would produce a distribution that had a 3 stepped down form
on this shape
(name: great rhombicosidodecahedron)
there are 30 squares
there are 20 hexagons
there are 12 decagons
since the length of all sides are the same it makes getting the areas easy
(assuming unit edge length of 1)
squares have A = 1 u^2 = 1
hexagons have A = 3*sqrt(3)/2 = 2.598...
the decagons have, A = 5/2 sqrt(5+2 sqrt(5)) a^2 = 7.6942...
The surface area is 30 (1+sqrt(2 (4+sqrt(5)+sqrt(15+6 sqrt(5)))))=174.292...
chance of landing on any square square is 30*1/174.29... = 0.172...
chance of landing on any hex is 20*2.598.../174.29... = 0.298...
chance of landing on any decagon is 12 * 7.6942.../174.292... = 0.5297...
>>43924787
for individual sides
chance of landing on a specific square square is 1/174.29... = 0.005737563...
chance of landing on a specific hex is 2.598.../174.29... =0.01490...
chance of landing on a specific decagon is 7.6942.../174.292... = 0.044145...
not sure how to make the graph through anydice, working on that
>>43924811
I'm not really aware of the physics involved or a concept that will pair down all the trivialities out, so really this is just a estimation
but I would guess, it being close to a sphere, the reality makes the decagons far far more likely than these estimations with them reducing the momentum way more than the other sides, while the smaller shapes less
Regular shapes can effectively ignore these issues because of their uniformity
>ctrl-f calendar
>0 results
Step up your game.
>>43924895
Actually there was, see >>43905803
>>43924895
>Step up your game.
Okay.
>ctrl-f month
>>43899267
>>43905803
>>43924904
>>43924908
I can't find the table in the 1e DMG.
>>43905803
>has 30 days in every month, plus a 5 day period
I find it hilarious that Exalted, touting itself as going far and away from D&D, has the same arrangement.
Although with three more months, but still, neat 30 days and then 5 special days.
>>43922771
>>43897116
I think it's important to remember that it's more likely for people to have d6s around and is thus an easy/no extra cost way to roll a lot of dice for the pretty numbers.
>>43925031
If it was less out of whack than 6d6-6 then sure. But that's just way too much.
Something like d12=2d6 is passable, but a six-die curve is noooope.
With d6s only, you could do d30 as (1d6-1)*6+1d6 and reroll everything over the cap. Although thats' only d30, actually common dice like d20 are far out (4d6-4 is waay curved. Like dem creets? Well you can forget about them, 4d6-4 has sixty times lower chance of a 20 than d20.)
>>43925111
I was considering cases where you WANT a bellcurve. Crits would obviously not be max everything, so probably degrees of success. Exploding each die that rolls 6 might be interesting, if extremely time consuming to keep count.
but probably stick to (1d6-1)*5+1d6-1 for uniform chances. Not too hard to find 2 d6s for some impromptu gaming.
>>43925163
>I was considering cases where you WANT a bellcurve.
That means designing a system with mechanics of a bucketful of d6s in mind rather than substituting d6s for dice you couldn't find.
We have plenty of the former, GURPS, FUDGE, Shadowrun.
The latter screws up mechanics, 4d6-4 won't be good in D&D.
>but probably stick to (1d6-1)*5+1d6-1 for uniform chances. Not too hard to find 2 d6s for some impromptu gaming.
That has some irregularities, but is passable, yeah.
d30 is a rare occurrence anyway.
d20 is much harder. You could go a little ways with (1d6)*4-1d6 and rerolling everything that goes overboard but it still has differences and is quite annoying.
>>43925238
>That has some irregularities, but is passable, yeah.
....damn, hadn't realized. The closest I can come up with is 1-36 or 0-35 with uniform chance.
Maybe in a system with babylonian counting?
>>43925286
You only get perfect uniform chance on (1dX-1)*Y+1dY type formula. If your needed pseudo-die is not X*Y of any combination of dice you have (or X*Y*Z*etc for more expansive formulas), then you're not getting your perfect distribution.
>>43925375
hmmm, what about:
5*(1d6-1)+1d6
ones digit 6 explodes: add 5 and roll again
uniform, with a bit of positive skewing.
>>43925499
>ones digit 6 explodes: add 5 and roll again
I am not at all sure how that will help.
You're still trying for d30? Exploding the six will burst you out of range, skewing everything else up too.
>>43925554
It's close enough to uniform to be treated as such, while allowing for crits that don't deviate TOO wildly from regular values.
>>43925578
Really not seeing the benefits.
>>43925597
is the third one really exploding? I thought there wasn't a way to do it on anydice.
Shouldn't it be more curved anyway, with the higher values having a slight upward trend?
>Really not seeing the benefits.
the benefits are mostly a compromise between the rush from getting a crit, balance, and convenience.
I mean, there ARE people who like exploding dice, right?
>>43925630
>there ARE people who like exploding dice, right?
No.
>>43925630
Hrm, I tried approximating via custom die like d{1,2,3,4,5,(5+1d6)} but I forgot to test it and seems like it doesn't work.
..Oh it actually has explode die function. Lemme play with it for a bit.
Ok
function: customplode N:n {
if N = 6 { result: 5 + 1d6 }
result: N
}
output [customplode 1d6]
Seems to create the exploding die you want.
Plugging it in...
>>43925708
...Actually you know what gets PERFECT d30?
(1d6-1)*5+1d5 with 1d5 being "1d6, reroll if 6"
There. DONE.
>>43925727
yeah, I thought of that BEFORE the exploding dice. Probably what I'd use for folks like >>43925662
>>43925708
but I didn't know about customplode, I need to check that out.
>>43925764
>>43925708
>explode is literally right there in the function library
....I SWEAR that wasn't there a few versions ago.
>>43925764
Customplode is just what I named an edited function, it's not a standard anything. Check "function library" to see standard explode functions and what I modded it from.
>>43925708
edited it a bit
set "explode depth" to 4
set "maximum function depth" to 4
function: customplode N:n {
if N = 6 { result: 5 + [customplode 1d6] }
result: N
}
output (1d6-1)*5+[customplode 1d6] named "d30 ctm explode"
output (1d6-1)*5+[explode d6] named "d30 std explode"
>>43925499
Just do 5*(1d6-1)+1d6 reroll any 6 on the second dice. It's the exact same thing as a d5 that way.
>>43923138
You can do that on the d4, d8, d3, d6, and d12, but be aware that the "reroll" result on the d20 table is important. If you don't discard BOTH the heads+heads and heads+tails+heads results, and just take it to be a 17 or higher, you're inflating the odds of those four numbers.
When you have heads+tails, there's still a 50% chance that you've got something less than a 17.
>>43892501
I think the old game Reich Star used the d30.
>>43893327
But that gives each number of hits an equal probability, rather than each shot an equal probability to hit. That means each bullet that hits magically makes other bullets in the spray more likely to hit.
>>43925111
>4d6-4 has sixty times lower chance of a 20 than d20.
Why would you say "sixty times lower?" What the hell, man? You can't say "one sixtieth?" You have to multiply things down now?
URRRGGKK-H@4*Y=G:A!faiS4aL+++
NO CARRIER
>>43903394
...That's not a bad idea. I have a programmer friend that I can pitch the idea to. He might go for it.
>>43892501
Hi Danielle
>>43892501
For THAT GUY to use instead of the D20 he's supposed to roll because he's mad at the GM but won't talk about his problems like an adult
