[ 3 / biz / cgl / ck / diy / fa / ic / jp / lit / sci / vr / vt ] [ index / top / reports ] [ become a patron ] [ status ]
2023-11: Warosu is now out of extended maintenance.

/sci/ - Science & Math

View post   
File: 44 KB, 440x442, 1312259307816.jpg [View same] [iqdb] [saucenao] [google]
3493289 No.3493289 [Reply] [Original]

Sorta interesting question on /tg/.
>>>/tg/15786990

Basic question is, what gives you a better chance for a dice roll? 1 20-sided die, or two 8-sided dice?

>> No.3493310

>>3493289
Expected values: 10.5 and 9 respectively. 20-sided die wins.

>> No.3493313

The two 8-sided have avg=9, max=16, min=2
The 20-sided has avg=10, max=20, min=1
I'mma go with the 20.
brb TI-84 simulation.

>> No.3493333

Holy fuck how is this even being discussed?

Expected values, fuck.

E(20 side die) = (1/20)(1+2+...+20)

E(two 8 side dice) = 2E(8 side die)= 2(1/8)(1+2+...+8)

You crunch the (very easily computable) numbers and the higher one is the one you wanna roll.

I trust the guy who said 10.5 and 9

(this is likely the calculation he used)

Fucking babby stats kids, babby stats.

>> No.3493341

>>3493333
Some of us didn't like stats.

>> No.3493357

I think the more interesting question is how does your choice change if you have an expected number or range of numbers to beat.

>> No.3493359

>>3493341

It's just the average, you idiot.

Do people seriously not know how to determine a geometric mean?

>> No.3493361

>>3493313
Crap, 10.5 I mean.

Results after 500 rolls:
d20-5302 (52 higher than expected)
2d8-4514 (14 higher than expected)

>> No.3493369

>>3493361

WHY ARE YOU USING A CALCULATOR PROGRAM?

YOU JUST ADD UP THE NUMBERS AND DIVIDE BY THEM BY THE AMOUNT OF NUMBERS

TROLOLOLOLOL

GET OUT

>> No.3493373

>>3493369
Simulations are more fun.

>> No.3493381
File: 172 KB, 400x341, 1285463299376.png [View same] [iqdb] [saucenao] [google]
3493381

>>3493369
OH GOD I'M SO SORRY PLEASE FORGIVE MY EMPIRICAL EVIDENCE I WILL NEVER SUBMIT ANY AGAIN.

>> No.3493429

The question isn't about the higher average, which is obvious. The pertinent element is the number required to pass the test.

>> No.3493439

>>3493429
>The question isn't about the higher average, which is obvious. The pertinent element is the number required to pass the test.

Sorta mentioned here
>>3493357
But no one seemed to notice.

>> No.3493450

>>3493429
Which is absent from this example, given that the /tg/-er didn't set a DC (or, Difficulty Check) for the die rolls in question. Certain situations in these games require varying DCs: the higher the number, the more difficult the task is to complete. Without knowing the ranges of the possible DCs that the /tg/-er would utilize in his game...

>> No.3493480

>>3493450
>Certain situations in these games require varying DCs: the higher the number, the more difficult the task is to complete. Without knowing the ranges of the possible DCs that the /tg/-er would utilize in his game...

I think it's safe to assume that you might need to roll anything for the sake of argument. So you might need a 1 or you might need a 20 or anything in between. Given that, I think the d20 still ends up being the better choice most of the time, but I'm tired and not thinking straight.

>> No.3493785

>>3493333

Stop being such an arrogant asshat. It's people like you who give /sci/ a bad goddamn name. Here's babby statistics for you.

For every possible roll you can perform with the 20 sided die, you've got a 0.05 chance of rolling any specific number.

With the two 8 sided dies, the probability of landing numbers that would be considered "territory" of the 20 sided die is actually higher since there's a larger amount of possile combinations. For example, 13, which sounds like 20 side die territory, actually has a 0.0125 higher probability to be rolled by the 2 dies.

Wolfram makes it easy; http://www.wolframalpha.com/input/?i=2+8+sided+die

http://www.wolframalpha.com/input/?i=20%20sided%20die&t=wvg01