@Glen_b I infer that you found it

Yep.

I've removed the offending remarks (the accusation that Bill didn't know enough Bayesian stats to conclude the question was unclear was equally bizarre) and offered to close the question myself if that would help.

@Glen_b do you think it's exhausting to exist in a world where everyone is simultaneously dumber than you and out to get you? I'm wondering if we should offer psychiatric references along with software help in our reasons for closing.

heh

If I find any topic I know better than Bill, I'll let you know.

I'm confident that the only hope I have of laying claim to that mantle is in the realm of heavy metal

and even then, I'm not sure

I can't think of any. I might come close in a few niche areas here and there but outside that? Nah

At least on stats/mathematics related topics.

:-p

remember the time that the power was out at his house, so he wrote one of the most comprehensive answers ever?

No, I don't think I recall that but my memory is bad.

I wish my memory was as good as it was six years ago, I'd probably remember stuff like that. I don't even recognize a lot of my own posts.

Ooh, hang on I'll go read it.

Ah, yes, I have read the post

maybe it's because I grew up on the Internet, but I'm really good at remembering where things are -- weird turns of phrase to find things -- but not what they actually say

I had forgotten about the comment underneath it.

Don't ask me what year my children were born; I have to work it out each time from their ages (at least I usually remember that)

lol

riddle me this: why do I spend so much time thinking about math when most of my problems are with software?

e.g. the 10 hours I spent today trying to get OSX, python and a module to play nice together

I often have to work out my own age from my date of birth. Goodness knows what this says about my own memory ...

xkcd on linear regression:

I foresee manuscript reviews that consist of nothing but this link.

@StephanKolassa I like linear regression joke :)

@Silverfish: That's one of the signs of entering into middle age.

Uh-oh, another statistically dubious xkcd strip. Residuals that are merely large are not much evidence against linear regression being an appropriate model.

It is perfectly consistent with a linear model to have a big error term.

@Kodiologist I understood the point of that XKCD comic to concern effect size and predictability. If the fit changes little across the range of the regressors, there's not much effect--one is tempted to say "so what?". In a nice classical scatterplot like this, such a low R^2--however significant it might be--shows that most of the variation in the response is modeled as random error (unpredictability).

@Glen_b Re "find any topic": I see I have you well fooled. :-) I'm sure I know next to nothing about statistics; it's too vast and dynamic a field. But perhaps that puts me in a good position to assess what makes sense and what doesn't, what is clear and what is unclear. So user "Jill Sellum" got her criticism exactly backwards: my relative inexperience with Bayesian analyses prevented me from automatically filling in the gaps in her question and helped me see it was nonsensical.

@NickCox @Scortchi @whuber Thank's for your advice on the closed question thing. In the future I will flag a moderator if I think I can put together an answer on a closed or closing question. I'm not sure why I didn't think of that before.

@GeneralAbrial Maybe the causation is the other way: most of your problems are software because you spend so much time thinking about math. Whenever a math or stats problem comes up, you've already thought about it.

@GeneralAbrial I just received a Bathory shirt in the mail. Hammerheart!

@MatthewDrury Just got my Bongripper shirt yesterday. they finally restocked their merch.

Gonna listen to that, are they inspired by Sleep? I have Dopesmoker on vinyl.

That record is awesome.

@MatthewDrury it's similar -- I'd say more along the lines of Karma to Burn or Pelican.

Oh, I like Pelican.

I'll check it out.

@MatthewDrury I saw Sleep live a few months back. it was a great show

I bet.

except that they only played a fraction of Dopesmoker. :-\

Well it would be the whole show to play the whole thing right : )

wellllll... yeah...

it's also like 90pct of the reason anyone would attend the show

Hahaha, fair enough.

I saw Slint recently and they played all of Spiderland. Which is kinda a similar situation, where the band has this one seminal completely mind shattering album.

Well, recently is like a year ago or something.

named after the Game of Thrones captain of the gold cloaks?

Probably the other way around.

I would like to star this entire conversation. Almost completely incomprehensible.... Anyone like Beethoven or Mozart?

@NickCox do it

Too much like work!

I like beethoven, but have trouble with Mozart.

I keep working on classical music, it's been one of the most persistently difficult generes for me.

But ill get there.

@NickCox perhaps we can meet halfway? youtube.com/watch?v=4cqQEMBU3D4

Cirith Ungol playing Toccatta in D minor on electric guitars

Cirith Ungol I recognise as a place in LOTR. I was reading that in 1965. It seems they're a rock group too....

yes, IIRC, one of the teeth of mordor

@whuber Yep, that makes sense, too.

@NickCox I play through 31 of the 32 piano sonatas every two to three years, in different orders. (Guess which one is left out? :-) It's different every time.

2^31 is quite a lot.

I'm wondering what the distribution of completion times of a deterministic process on a computer looks like. It's some constant (minimum possible completion time) plus a lognormal (mean = average delay?) or something like that?

@NickCox I love Beethoven and I like Mozart. I especially like Beethoven's Op. 81a and Op. 106 piano sonatas.

@AaronHall It depends on the distribution of inputs. If the computer is not having to respond to other inputs--interrupts, messages, etc--and you give it the same input time after time, then in principle the completion times will be identical, right down to the number of clock ticks.

Well I mean that there's some amount of noise.

various processes getting prioritized by the OS or competing for resources at random times.

@AaronHall Right: so your question comes down to what one might expect about the distribution of the noise. That depends on the system--how it's connected to other systems, what demands are being made on it, how its peripherals operate, its architecture, and more.