do any of you guys know the resulting probability distribution from the linear combination of two discrete uniform random variables

[Al Roker]

thanks in advance

 

[BP]

y=mx+b

 

[phoopes]

I think it's a discrete triangular distribution but I'm going to call in the big guns here

Estarossa and Astra I think

 

[Blizihguh]

>log out of my job teaching stats
>"golly gee it's nice to have a few days off for New Year's"
>log onto Smogon dot Com dot Smogoff
>see this thread


but yes Phoopes is right

 

[Comett7778]

Using smogoff to look up answers is something else

 

[az]

3

 

[az]

 

[TheMantyke]

this is another ligma set up, isn't it

someone's gonna come in here and say "I discretely distributed in your mom's triangle" right

 

[chroma_2903]

updog. the answer is updog

 

[mysticmarshtomp]

mitochondria is the powerhouse of the cell

 

[Astra]

I think it's a discrete triangular distribution but I'm going to call in the big guns here

Estarossa and Astra I think

probability and statistics is fake math i don't associate with those things

 

[bdt2002]

Wouldn’t the answer to your question depend on the outputs of your randomized discrete variables?

 

[Peum]

yes

 

[Blizihguh]

Wouldn’t the answer to your question depend on the outputs of your randomized discrete variables?

Good catch! With a discrete distribution, there's no reason the values need to be sequential. The triangular result is resting on the assumption that the values are equidistant from one another. For example, if you have two random variables A and B, both of which can be any of {1, 2, 6} with probability 1/3, the possible outcomes look like:
2 3 3 4 7 7 8 8 12
Which certainly does not look triangular.

In general I would say that an unspecified discrete uniform variable is likely to have sequential values. But almost everything in stats relies on your assumptions. Technically that's true of all math, but it's much more prevalent in stats. Like, if you're doing linear algebra or something, you don't start every problem by considering whether it's appropriate to model it using vector spaces. The point is, if you're doing stats, you should be cognizant of the assumptions you're making when you apply any rule, formula, etc.

 

[Blizihguh]

It's nice posting educational material on Smogon. On sites where it's actually on topic, you get spammed with comments like "you're a fucking moron for even trying to teach this person" or "nuh uh, one time in middle school my teacher said birds were a kind of fish, and I trust them more than you"