XKCD: It's more than a comic

I rather think it's a reference to "magic numbers" in programming, which you are told not to use ("magic numbers" are in general numbers that just are in the code without being explained, instead for example a constant with a self-explaining name)

Ah, I see. That puts it in context.

Even with that in mind I still think the main joke is someone hating a number.

Going back a couple of pages

The chance of a million-sided die coming up a 2 is equal to it coming up and other number from the range of one to a million, but that means that the odds are one in a million, whereas the odds of getting a 2 on a six sided die are one in six. This legitimately makes it much easier to guess the number of sides on a die from the data point of a single roll than it would be without that single data point (assuming a population of dice with highly diverse numbers of sides; in real life, you're better off always guessing six.)

The Doomsday Argument gives us information about when we can expect our species to go extinct, it's just that we have a lot of other sources of information, many of which are likely to contain significantly more data despite offering less precise predictions.

You're assuming statistical properties of the universe of possible dice based on external factors: namely, the much greater known distribution of six-sided dice in the population. We don't have any such information on the lifespan of the human race because we're a unique statistical object. There has never been a species like us.

If you take out the external known factors, then the die in your example is even more likely to be a two- or three-sided one — or four, actually, since you can't make a regular polyhedron with less than four sides.

But if you take any possible number of sides, from four to infinity, without prejudice, then one data point is insufficient to establish a statistical probability.

edited 10th Oct '13 9:04:57 AM by Fighteer

It's insufficient to determine with high confidence what the actual number of sides is, but it does indeed give you a probabilistic answer regarding the number of sides. The more data points you have, the higher probability you can assign to your estimate, but you have information which narrows down the space of your estimates quite dramatically with as little as one data point.

By the same token, a person observing a human specimen with no prior information, who randomly selects a five year old child, might conclude that our lifespan is statistically likely to be 10 years. That's nonsensical.

It's only nonsensical because you know the correct answer. It is statistically likely from the limited data they have. Obviously limited data is worse than plentiful data, but that doesn't make the prediction a fallacy or anything, it's just less accurate than it could be.

It seems to me that you're very biased against imperfect predictions. Or as I've said like three times already are assuming that this prediction is stronger than it's actually claiming to be.

No, it's nonsensical because it's applying a logical fallacy. Statistics is only as useful as the predictions that it makes. You can be perfectly correct on the math and still arrive at a wrong answer because you're using wrong assumptions.

In this case, the assumption is that, "The lifespan of a species or an individual can be predicted from a sample size of one."

edited 10th Oct '13 10:05:00 AM by Fighteer

I thought we already established that the doomsday equation was stupid but nobody could agree why, as Randall put it

The lifespan of a species can be predicted from a sample size of one, just with low confidence.

If you suggest that one might do this with humans by extrapolating from a 5 year old child, it sounds silly, because you know in advance that the prediction is incorrect (in fact, you know that the real answer would be outside an 80% confidence interval for the prediction.)

On the other hand, if I tell you that individuals from species on Earth generally have lifespans between 8 hours and 3000 years, and then tell you to guess the average lifespans of species a, b, c, d, e, f, g, h, i, j, k, l, m, n, o, and p, you're almost certainly going to do a lot better if, rather than leaving it at that, I give you the current age of one random specimen each from each species.

On a side note, the proper operation to predict the average lifespan of a species from the age of a random individual isn't the same simple x2 multiplication that you'd use for guessing the number of sides on a die, because species generally do not have flat distribution of their population through age cohorts, but even the x2 operation would still give you more accurate answers than maximum entropy.

edited 10th Oct '13 10:50:36 AM by Desertopa

Statistical analysis gives you a possible range of predictions from the specimen's current age to infinity. You'd do better to analyze its physiology, or feed it and wait for it to die naturally.

edited 10th Oct '13 11:10:04 AM by Fighteer

But not a probable range of predictions. You can narrow down to an 80% confidence interval for the species' average lifespan which will generally be less than an order of magnitude wide, where your starting range would stretch from zero to infinity.

edited 10th Oct '13 2:40:07 PM by Desertopa

#1276: Angular Size. This could easily be a What If?. I also love the recursive logic of the Alt Text.

Recursive?

Though that does make me realize the irony of the idea: we'd still be using lenses, they'd just be pointed down instead of up.

Recursive, because you'd need to explore the celestial sphere in order to map it to the Earth's surface, in order for people to be able to use microscopes to see it.

Ah, of course.

"Emergency Backup Earth" :)

Those names are callbacks to this strip.

Ah, Venus is no longer "football venus" or whatever that glitch was. Good.

Ayn Random

Anyone care to explain the alt-text?

Alt-text is a regex, one which specifically refers to people whose names contain only letters in the sequence PLURANDY. And I think I have insufficient context to really get the comic.

Ayn Rand is famous for her philosophy of Objectivism, which is basically super Libertarianism, specifically that the people with economic power deserve it because they're better than everyone else, who are implicitly parasitic scum. It was created as a reaction to her early life in Soviet Russia. So basically she hated communism with all her heart. In terms of the comic, her random number generator refuses to treat all numbers as equal.

edited 14th Oct '13 4:56:34 AM by Clarste

It is a terrible, terrible pun. Also, what's that "b" in the regex? Shouldn't it be "\b"?

And later in her life, she got cancer and became the "parasitic scum" she hated.

edited 14th Oct '13 3:36:25 AM by Medinoc