Me rolling for internet speed.

I've never actually played any RPGs. I wanted to but all my friends said I was a geek and wouldn't join in.

Computer games are better anyway.

@Bellerophon I am planning on running a one-shot (or maybe even multiple sessions) between the semesters, if you want to I can ping you when I set it up on rol20

@dot_Sp0T That would be great if you don't mind.

I have an original ADnD set with two of the dice. They didn't paint in the numbers in those days.

I'll just ping you on here :D

Thanks.

But I'll first need to finish writing about Edward and Jacob.

@Geo I get bored of computer games quite quickly.

I've been thinking about the XY problem since someone suggested one of my questions might be a case of it, and I realized that the stack exchange format actually encourages the XY problem. How strange!

@Geo That's an interesting statement. What makes you think that.

@sphennings Well obviously stack exchange has rules about what questions can be asked. Certain questions don't fit. So it wouldn't be at all strange for it to simply be impossible to ask X, and so substituting Y wouldn't happen by mere mistake, but by necessity.

This seems like it would mostly happen when X is too broad. You want to ask X, but you can't, so you think of a specific way to solve the X problem, if only you knew the answer to Y, which is naturally a more specific question. It's a recipe for the XY problem.

Of course, it's not an XY problem if the answer to Y really will solve X. That's what you have to do to avoid the XY problem.

I haven't seen too much Xy problem around.

Fun question for the mathematically inclined: what’s te probability density function for the result of choosing the top y of xdn?

@JoeBloggs Are you talking about dice?

Took me three days to figure it out. It seems so very simple, but noooooo.

What does "xdn" mean?

Yep. Roll x n sides dice, what’s the likelihood of the top y results adding up to any given number.

In the same way that 2d10 would be 2 10 sided dice, but with x and n.

Hmm. I see. So y is less than or equal to x, and you only uses the y dice with the greatest values.

Yeah.

I guess the way to solve that problem would be to start with n = 1 and move up until you see a pattern. Sounds tricky to me, not simple, but I've always been terrible at combinatorics.

My solution ended up with three nested summations and some of the gnarliest combinatorics i’ve ever employed.

I don't know how to display probability density but I'd guess that x(1/2n+1) is the highest likelihood and it falls off not quite equally towards each side.

There's no reason to expect a complicated question like that to have a simple expression as its answer.

@Bellerophon: basically you work out the probability of any given result and list ‘em, but mathematically.

But if you plot the function as a graph, you might be able to see a recognizable curve and simplify the formula.

@Geo: it’s a great problem because posing the question is simple, and people with a decent grasp of combinatorics or probability get suckered in.

@geo: sadly it’s not that nice at all.

I think requiring us to only use the y greatest of the dice is a terribly complicating requirement, so I don't see any anyone would expect the answer to be nice.

hey there @Bellerophon

I didn’t expect it to pe ‘nice’, but I didn’t expect it to be as utterly awful as it turned out to be. Like you I thought ‘oh, I’ll plot this, check the patterns, have it sorted in no time’

@Shalvenay Greetings.

Then it was three days later and I’d forgotten how to Human.

how're things going?

@JoeBloggs I am a bit curious to see what the plot looks like.

Heya

hey there as well @JoeBloggs, how're things going?

@geo: which one? X, Y or N?

@Shalvenay: not bad! Yourself?

OK here

Doing pretty well here.

@JoeBloggs I'd make the plot for Y, and look at how it changes when I change X and N.

I'm probably going to leave now. See you all tomorrow.

Or maybe not. I mean you might not be there tomorrow.

Or I might not.

@JoeBloggs No wait, the plot would have the value of the dice on one axis and the probability on the other axis. All of X, Y, and N are constants.

I’ll pull out the code I used for it and put some plots together tomorrow. They should be pretty.

You can put 3D ones together or fix the value of the dice at something sensible.

3D is pretty hard to visualize, though. It makes it hard to recognize a curve.

To be fair, for low values of x and n you can just directly enumerate it very easily. :-)

One of the best plots I ever saw was 3D and heatmapped. It looked beautiful, but was chuff all use.

Anyhow: my phone is about to die. Catch you all in the future!

Probability of the top Y of XdN? We want the percent chance of any particular total "Z?"