Talk:Doomsday argument/Archive 2

This is an archive of past discussions about Doomsday argument. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

Archive 1

Archive 2

'Math'-free

Not only is 'Math' an unnecessary abbreviation, it is also a US-specific term. This can all be avoided by expanding the term, at no loss of readability. —Preceding unsigned comment added by Tescoid (talk • contribs) 16:07, 14 July 2008 (UTC)

Huh?

The argument in Carter's article cited in the lead has very little to do with the one in this article, beyond the fact that they are both Bayesian arguments about the evolution of mankind. Carter's argument concludes that there has likely been only one or two determinant "unlikely" steps in human evolution so far, and that most of the other unlikely steps weren't determinant (e.g. there's no a priori reason why civilization couldn't have been created by egg-laying animals). He concludes absolutely nothing about the future lifespan of mankind given its current age (he only assumes that it'll be shorter than the lifetime of the Sun). What's going on? -- Army1987 – Deeds, not words. 17:07, 30 December 2008 (UTC)

Generation length rather than lifetime?

Shouldn't we be using the generation length rather than the lifetime in the calculation of how long humans will be around in the lead? Warren Dew (talk) 03:02, 28 March 2009 (UTC)

How long is a piece of string?

The article is misleading, lending credence to an absurd argument - as pointed out by others above. Imagine we are on a piece of string; can we say anything about the overall length of the string, given an estimate of our position from one end? No, we have no knowledge of the string's overall length and can say nothing about it, not even a probability. The article ought to be clear that the DA is flawed and invalid. Dare I add that this sort of thing is dear to Philosophy's heart: an proposition with profound implications, supported by mathematical technobabble. Barlow (talk) 06:40, 6 August 2009 (UTC)

If you take a piece of string and cut it at a point chosen uniformly at random, then tell me the length L of one of the pieces, then it is reasonable for me to state with 95% confidence that the length of the original string was less than 20 times L. That is the Doomsday Argument in a nutshell. Yes, it is a very simple argument dressed up in a load of "mathematical technobabble," including ill-defined "principles" and nonsense "priors" that don't sum to one. The fact that so much effort has been put into this, and that papers on it have appeared in Nature, is astounding. Anonymous, Wed Feb 3 10:09:27 CST 2010. —Preceding unsigned comment added by 69.137.68.170 (talk) 16:10, 3 February 2010 (UTC)

Edit to John Eastmond's Many Worlds paragraph

I've taken out the paragraph referring to my many-worlds refutation of the Doomsday Argument. I think the paragraph goes against the wikipedia directive that one should not promulgate unpublished research. John Eastmond (talk) 09:28, 15 October 2009 (UTC)

Fail in second sentence

The article says, "Simply put, it says that supposing the humans alive today are in a random place in the whole human history timeline, chances are we are about halfway through it." Sorry, you failed probability. Given a uniform distribution, it's not any more likely to choose a value near the middle than near the edges. —Preceding unsigned comment added by 128.83.61.200 (talk) 20:46, 10 November 2009 (UTC)

Problem with the infinite expectation integral - slicing Adam.

The integral used to claim infinite expectation for N (the total number of humans ever) assumes that the variables in question are continuous. But clearly, they are not! They are quantized by 1/N. And the moment we assume "f" to be in the range (0,1], we are implicitly assuming N to be finite number.

Put another way, the expected infinity of N actually comes from slicing up the first human in countless ways, and counting these slices over and over again (think of the graph of the function n/f - the infinite area under the curve comes from the part where n, f are close to 0). —Preceding unsigned comment added by 212.253.35.93 (talk) 12:43, 29 May 2010 (UTC)

And another thing on this section: it claims that the expectation value is infinite (and the mathematics is correct), but that doesn't mean that the 95%, or even 99.999%-confidence upper bound is infinite so you can still state that the number of humans that will ever exist is less than some finite number with arbitrary confidence. — Preceding unsigned comment added by 131.111.185.74 (talk) 22:32, 19 October 2011 (UTC)

Bogus Premise

This looks like a complete misuse of probability to me. We aren't marbles in a box waiting to be sampled and put in a life. There is no box and there are no marbles. We are the result of the circumstances unique to each one of us, environmental and heredity, and there is no other time we might have been selected as we didn't and couldn't exist until those particular circumstances existed and then we did. To be born at some other time and place would make us some other person, not the same person sampled into a different life. This whole thing is just more pseudo-science posing as real science, and needs to be noted as such. Jlodman (talk) 23:12, 12 August 2010 (UTC)

Cite pls k thnx bai. --Gwern (contribs) 05:41 26 August 2010 (GMT)

Philosophy

The doomsday argument is different from the usual statistical problems. It is not clear that you can argue about it in a statistical/mathematical way. Maybe you have to go to the foundation of science, which is philosophy, if you want to rebut it. By things like this citation by William Kingdon Clifford, ‘It is wrong, always, everywhere and for any one, to believe anything upon insufficient evidence.’ Reko (talk) 21:26, 28 September 2010 (UTC)

Bad graph

I think we need a graph which shows the recent deceleration in population growth. Ginger Conspiracy (talk) 07:00, 22 October 2010 (UTC)

Introduction is horribly written and is complete nonsense

After reading the first 5 paragraphs of this article (i.e. the introduction, which should CLEARLY explain the subject), I found myself completely confused. The 3rd paragraph attempts at some sort of mathematical argument, but the argument is nonsensical.

Quote: "Denoting by N the total number of humans who were ever or will ever be born, the Copernican principle suggests that humans are equally likely (along with the other N − 1 humans) to find themselves at any position n, so humans assume that our fractional position f = n/N is uniformly distributed on the interval (0, 1] prior to learning our absolute position."

1) "Humans... (along with the other N - 1 humans)..." This does not make any sense! Does the author mean a particular human?

2) "...To find themselves at any position n..." What position? Position in time? Numerical position.. as in the order of your birth?

3) "...so humans assume that our fractional position f=n/N is uniformaly distributed on the interval..." WHAT?! "humans assume that our..." This is very sloppy language. And I don't understand how a fractional position (using this definition) is "uniformaly distributed" on an interval... again, very sloppy.

4)"...prior to learning our absolute position." How is the fractional position different from the absolute position? I assume the author means "assumed fractional position" and "absolute fractional position". Anyhow, how does one "learn" their absolute position?

Also, in the next paragraph,

5)"...If we know our absolute position n, this implies an upper bound for N obtained by rearranging n/N > 0.05 to give N < 20n." THIS IS RIDICULOUS! Here's a simple contradiction: lets say we repeated this same EXACT argument 5000 years ago. Then n would perhaps be 50 million. Does that mean there is a 95% chance that the total number of humans to ever live would be 1 billion? That estimation is off by at least 2 orders of magnitude even now... and population is still growing...

That's 5 reasons why this article (or at least the introduction) needs a complete rewrite. —Preceding unsigned comment added by 76.93.217.205 (talk) 12:06, 22 October 2010 (UTC)

Assumption that any N is equally likely!

"f is uniformly distributed on (0, 1] even after learning of the absolute position n. (This is equivalent to the assumption that humans have no prior information about the total number of humans, N.)"

Sentence in brackets is wrong. I'll proof with an example, assuming f to be me and n=60 Billion. If f would be between 0.1 and 0.2 then N is between 300 and 600 billions, on the other hand if f is between 0.2 and 0.3 then N is between 200 and 300 billion. But if we do not know N, we should treat each possible N with same probability (excluding N<n). Because many possible Ns are very high, this can be only done by assuming f to be close to 0. q.e.d. --134.76.233.140 (talk) 12:45, 22 February 2011 (UTC)

Please say in Introduction that this is pseudocience!

The article is misleading people, it has to assure in the very introduction it's considered a fallacy by mainstream scientists.

There is 10% chance we are in the first decil, 10% in the second decil, and so on. Wherever we are, the probability is very tiny (infinitesimal, if we consider integral calculus).

We have to be somewhere in the mankind timeline, it's like the weak antropic principle.

The mathematical / statistical argument presented is absolutely invalid, as noted by many comments in this discussion. — Preceding unsigned comment added by Jbbinder (talk • contribs) 13:52, 14 November 2011 (UTC)

For something to be pseudoscience, its pseudoscientific claims have to be presented as actual science. In this case, I'm not sure that there's a clear assertion by those who have independently published the idea that this actually predicts some kind of meaningful doomsday or something observable. That is, it seems like those who argue its statistical validity do not claim that any meaningful conclusions can be drawn (i.e. science) other than those about statistics itself or existentialism or philosophy, etc.

That said, I agree that the intro seems to have some lingering confusion as to whether the argument is scientifically meaningful. That the argument was subsequently championed" seems to imply that, where I think Leslie was more arguing for its statistical validity and what precisely it means (rather than whether or not it's "true"). That kind of thing should be clarified. SamuelRiv (talk) 07:04, 15 November 2011 (UTC)

By the standard WP terminology this falls neither under 'science' or 'pseudoscience'. The DA isn't pseudoscience in the normal sense because it doesn't purport to explain anything about reality. The DA isn't science, because it makes no synthetic claims (it is purely analytic). Since it is an exercise in mathematical reasoning (but without advanced mathematics) it falls fairly squarely in the category of philosophy. --Wragge (talk) 22:30, 21 November 2011 (UTC)

Non-human sentient life

Sorry if this is independent research, but I'm someone who thought up this principle and just now thought of an important extension/rebuttal. If other life evolves into sentience elsewhere, at any time, in any galaxy, in any universe, then we could re-frame the question to include any life-form capable of thinking of the question.

How many life-forms are capable of thinking of the question? Well, most people who subscribe to the weak anthropic principle believe that infinitesimal chances of life are made up for by infinitely many times and place, so that even if it is improbable that we should ever see aliens or invent philosophizing AI or whatever, most people would still agree that infinitely many sentient life-forms must, in some sense, evolve. Or in any case, if life is a chance event, then it is very unlikely that that chance should be met once and only once.

This changes the nature of the argument from one of a specific die roll, to one of large numbers. And puts a probable-collar, not only on our particular future with nuclear wars or whatever, but on the statistical distribution of technological/reproductive 'success' for all life anywhere. For if we suppose that one sentient life-form in a thousand goes off and colonizes peacefully, reproducing a quadrillion times, then we should expect to be one of those guys and not a doomed Earthling.

It's not impossible that civilized life should have a major-habit of dying out, but at the same time it's a very strong statement, much stronger than the standard DA, and therefore implicitly a critique...

I guess this is a lot of unsourced stuff, I'll see if I can find an article or something stating the same. --Crinttae

I agree that this could be at least mentioned in this article, but it probably belongs mostly in the Fermi_paradox article. Also, if there are so many rocks, why don't we exist as rocks? I'd say it's because rocks do not think the same way we do, not directly because they aren't sentient, because sentience is subjective, with no number tied to it. Vibrations of particles in a rock due to thermal energy interact with each other forming a complicated system, so a sufficiently large rock's processes are more sophisticated than any human. The reason probability didn't make us rocks is not because rocks are in any way dumb, but because they think in such a different way. So as long as the universe colonizing life-form thinks as differently to us as rocks do, we still can't say they're unlikely to exist. In fact, if such life-forms can spread their way of thinking to other life-forms, we can also explain why we don't see them with the same reasoning that explains why we haven't seen rocks which converts, everything seeing it, into more rocks (that reasoning is that we can't coexist, e.g. we don't live on an Earth orbiting a black hole because we can't). 173.180.202.22 (talk) 03:50, 3 June 2012 (UTC)

Rocks don't think. Complication<>sophistication. My argument was just for aliens that think similarly enough to be asking similar questions. Though I agree that something might think radically different. And, for that matter, we ourselves might start thinking radically different with some technological adaptation in the near future.--Crinttae

The word "extinction"

I also disagree with other parts of this article, but an obvious problem is the use of the word "extinction". While this article claims that our civilization stabilizing at 10 billion with a 80 year life span has a 95% chance of collapsing before another 9120 years, this does not imply that whatever causes the collapse will reduce the number of living humans to precisely zero. The word extinction strongly implies that there will be no single human surviving, yet the argument doesn't state this is true. Less than (60 billion * 5%) will live even if the "extinction" 9120 years later leaves a few remaining, as long as population stays low. 173.180.202.22 (talk) 04:59, 3 June 2012 (UTC)

Simple analogy

I think many readers will benefit if a simple analogy is added.

I like the following analogy of the doomsday argument. Suppose there are two boxes, identical from outside, put on the table. One box contain 10 white balls and 1 black ball, and another contain 1000 white balls and 1 black ball. You do not see the balls inside, and you can not understand how many of them are inside by weighting the box. You can take one ball randomly out of the box.

Then you randomly chose one of the boxes and get one ball out of it. And it is white. You try the same box second time and you get a black ball. The question is which box did you chose: with 1000 white balls or with 10 white balls? The answer is obvious that if you get the black ball from you second attempt, then it is probably the box with 10 white balls.

Do you think something like this can be incorporated into the article? I also remember that some science fiction book has this analogy, but I do not remember name/author for citation. —The preceding unsigned comment was added by MxM (talk • contribs) 20:07, 26 April 2007 (UTC).

Also with this analogy, a simple counter-argument is possible. What if you have 1,000,000 boxes that have 1000 white balls and one black ball in each, and only one box that has 10 white balls and 1 black ball. When you peek the box at random and your second ball is black, is it still more probable that you have 10-white-ball box in your hands? The answer is, of cause, no. MxM (talk • contribs) 20:07, 5 May 2007 (UTC).

The book you're talking about is Stephen Baxter's "Manifold: Time".

And by god, this is one silly argument. May look convincing, until you realize it is completely empty of any content, and is nothing but an amusing mathematical exercise.

Its main fault, of course, is the fact it takes the identity of the asker into account, giving inconsistent results when raised in different times. Another glaring flaw is the dependence on your arbitrary determination of how many people have lived so far (the boundary between ape and human); and failing to account for the fact this could all be easily circumvented, by the very same logic, by saying that around the year 10,000 we will have evolved into a form beyond human - the argument becomes null. okedem 20:40, 21 August 2007 (UTC)

But this are the same objections as in case of the main argument. If this analogy makes it more clear than this is good analogy. I feel like it is beneficial for people who do not want to go through math, yet want to understand the logic behind the argument even without special training. MxM (talk) 18:22, 5 October 2012 (UTC)

Named Wiki Editors

MxM , Crinttae and I (Mark v1.0) are the only named Wiki editors recently (one year) interested in the "Doomsday" article. Anyone interested in changing Doomsday please state so here, with a registered name.--Mark v1.0 (talk) 20:19, 10 January 2013 (UTC)

Found other named editors Rracecarr Gregbard Khazar2 R'n'B , who made recent changes without posting on this TALK page.--Mark v1.0 (talk) 20:25, 10 January 2013 (UTC)

Circular reasoning

This article assumes that human population is finite in time, then tries to prove it. There is no valid reason for that assumption (if you don't see why, just replace humans with numbers for correct, abstract appoach). Something else that has been mentioned multiple times: if you apply logic presented by DA to every human since the beginning of our race, you'll end up with conclusion that we're descendants of lucky 5%, who are descendants of 5%, who are descendants of 5%... See where I'm getting with this? According to DA odds that we continue to exist are less than fraction of a percent, if applied to first 10 humans. Article should be deleted or clearly labelled as another example of circular reasoning, right now it gives impression of valid logic. 87.205.128.27 (talk) 12:12, 10 November 2012 (UTC)

If you look at humans as just another animal on the planet like the animal shark, sharks will not end as long as they have food and can breed. I agree this Doomsday article should be deleted or labelled as circular reasoning. Who is going to do the deletion? Someone has to have the power and will. This Talk page is also HUGE. I will have to check which posts are the most recent to form a proper consensus of editors.--Mark v1.0 (talk) 20:12, 10 January 2013 (UTC)

Whether it is valid in the real world or not, the Doomsday argument is a well-established problem that is illustrative of the different schools of interpretation of applied statistics. Compare it to Schroedinger's Cat, which is a silly question with an easy solution; again it illustrates the importance of proper interpretation of mathematical law.

So if anything should be deleted, it's most of this Talk page. SamuelRiv (talk) 14:13, 24 February 2013 (UTC)

If it were up to me (and WP convention allowed it) I would copy&paste that paragraph, make it bold, and put it on top of this talk page. Minvogt (talk) 23:26, 12 March 2013 (UTC)

Past gives no information for future?

From the article: "the DA says that if the absolute number of humans born gives no information on the number that will be, we can predict the species’ total number of births after discovering that 60 billion people have ever been born: with 50% confidence it is 120 billion people".

But if the absolute number of humans born gives no information on the number that will be, then it cannot be possible to double the number born so far, because that would be assuming that the number born so far does give useable information. --Henrygb 29 June 2005 15:45 (UTC)

This is where the Copernian Argument comes in -- We cannot assume that we live in a special time or place -- therefore we assume we are in the middle. Brad (talk) 21:47, 11 July 2013 (UTC)

The article is misleading, and the argument itself is absurd

This whole argument is absurd, such thinking would lead the 10th human to conclude that with 95% there will be only 200 humans, the 11th human to conclude that with 95% certainty there will be only 220 humans, and so on. In general Nth human would claim that with x% accuracy there will be only (100*N)/(100-x) humans, for example the 10th human could claim with 99.99% accuracy there won't be more then 1000/0.01=100 000 humans.

Obviously such argumentation is absurd, the problem lies in switching from the case where N is defined to the case where N is the random variable, in the first case there is only a finite number of cases to consider and so probability is normalized to N, the second case on the other hand has *infinite* number of possible outcomes and so the probability has to be normalized accordingly.

There is exactly 0 probability that you are living withing the last 95% (or last whatever percent) of humans given the assumptions stated in the article. It's actually very simple, no matter what your position is the number of cases for which you are within the last 95% is finite and the number of cases for which you are not is infinite, the probability that an event which can only be realized in a finite number of ways out of an infinite set of possible outcomes is 0 according to probability theory. (In general the probability that a particular position is arbitrarily close to the end of an infinite sequence is 0.)

Of course this does not mean anything for human population as this whole problem is way to simplified and irrelevant. The article should however be clarified to point out clearly that the argumentation used is completely wrong and the "prediction" has no value whatsoever apart from being a mathematical curiosity perhaps. Enemyunknown (talk) 15:32, 24 February 2009 (UTC)

Without saying that I agree with the argument's conclusion, you are correct that the 10th human would conclude that there will be only 200 humans overall. This is not the major flaw that you think it is however: in fact the entire premise of this argument rests on the claim that you are extremely unlikely to be the 10th human, given that there will exist billions of them. The 10th human could indeed draw this conclusion and he would be wrong, however 99.999% of humans exist in a period of time in which there have been far more than 10 humans. The argument goes that, since the majority of humans who will ever exist can correctly claim to be roughly in the middle of the human timeline, that majority of people can make an educated guess towards where the end of humanity lies. Charles Baynham (talk) 23:45, 2 October 2013 (UTC)

Charles pointed out quite well the flaw in the Enemyunkown comment's earliest section, the second part however, that there is a 0 probability that we are living with in the last 95%, has its issues as well. For instance, given that n/N is uniformly distributed over (0,1], then the probability of any one given value for n/N is equal to 1/N. And such the probability that you are in the position f≤(n/N) is equal to (1/N)+(1/N)+(1/N)+... n times. Or, n/N.

From there:

P(f>0.05) = 1-P(f≤0.05) = 1-0.05 = 0.95. Therefore, there is a 95% chance that we are in the last 95% of humans.

There is also a problem in the reasoning that there is an infinite amount of cases for which I (for instance) am not in the last 95% of humans, namely that the writer draws his argument from the fact that there are infinitely many values that N can take, which, while true, is irrelevant. We know N to be finite, because humans WILL go extinct eventually, whether due to failure to escape Earth before the Sun goes red giant or as a consequence of the heat death of the universe or whatever, and as such no matter what N actually is, there is only a finite number of cases for which I am in the last 95% of humans.

Enemyunknown's argument itself falls to pieces when you consider it a little more closely:

The writer states that P(n/N ≥ 0.05) = 0 because the number cases for which this is true is finite and that the number of cases for which it is not is infinite. This is true; assuming that n = 1x10^11 people, this puts an upper limit on N at the value 2x10^12, because for the case where n/N = 0.05:

We have established that P(n/N) = n/N

therefore:

1x10^11/N = 0.05

1x10^11/0.05 = N

N = 2x10^12

and the lower limit n=1x10^11 for n/N = 1

Conversely the upper limit for P(n/N < 0.05) is given by:

1x10^11/N = P

1x10^11/P = N

So P→0, N→∞

And as Enemyunknown implies:

1x10^11/∞ = 0.

The problem with this is that using that reasoning, P(n/N ≥ f) = 0 for ANY f on the interval (0,1] which implies that N is infinite, which we know it is not. A contradiction.

Finally, the DA is an argument that gets updated as more data is accrued, as per Charles' explanation.

Archive the talk page

I want to archive this Talk Page, both because of its length and its oft-repeated sections objecting to the article's topic and not the article itself. WP suggests getting consensus before doing so, so let's give 10 days to object. The traffic is suitably low that archiving can be done manually, though there's no real downside to using a bot. SamuelRiv (talk) 02:36, 13 March 2013 (UTC)

Sounds like a good idea for the oldest material, keep anything that is particularly useful. Of course, the lead could be clearer. Finally watching article so at some point I'll come back and can figure out just what the argument is. Math not being my strong point. User:Carolmooredc 00:07, 12 October 2013 (UTC)

Clarity in the first bullet of Gott's formulation

P(N) is the probability prior to discovering n, the total number of humans who have yet been born.

This sentence seems to be missing a word or two or a comma or the right word. I'm not familiar enough with the topic to edit it myself, but I can write these alternatives based on the words in that sentence: 1. P(N) is the probability prior to discovering n, the total number of humans who have NOT yet been born. 2. P(N) is the probability prior to discovering n, the total number of humans who have yet to be born. Hope that helps. Rickcolosimo (talk) 22:05, 19 November 2013 (UTC)

Argument uses completely arbitrary start point to count N

The argument -- not the general theory but the specific calculations on when Humans might go extinct -- relies on the notion that 60 billion humans have been born so far. But this is an utterly arbitrary number based on current definitions of the word "species." Why does it make a difference that the person 61-billion-births-ago is not counted as "human" when it comes to calculating our future? Rather, if we're interested in calculating how long our descendants and their descendants will survive, we should count starting from our ancestors. Our ancestors appeared on Earth sometime around 3.5 billion years ago. Counting the faster reproductive times of single-celled organisms, there have been trillions upon trillions of living things born so far. So why not take that as the number, and then do the rest of the math? It would show that we have 95% confidence of trillions upon trillions of descendants. — Sam 63.138.152.196 (talk) 15:28, 6 November 2013 (UTC)

Not to mention the obvious and fatal flaw that it assumes that we somehow know how long the species will exist. I mean, if the species has existed for 20000 years, then following the argument that there will be 20000 more years of human population sounds absurd or that population growth will definitely go negative even with a chance that more resources will come with expansion to other planets(the overpopulation argument's not even remotely certain). The reason being that the distribution over time is not a bell curve or the like. Far from it in fact. It seems to totally ignore exponential growth or (and most fatal) changing definitions of what's human. Rationally and intuitively, I can't possibly see how this makes much sense as an argument. Yes, every species has to go extinct (assuming entropy[ignoring evolutionary drift] problem isn't solvable), but to somehow magically know that we're at the exact 50% time (or population) mark or even near it is pretty wishful thinking. This is a moving goalpost, unless you either know the timeframe or the population curve is pretty simple(i.e. bell curve). Worse yet, it ALWAYS appears that you're the definition of human. People from 20000 years in the future or past might not even consider each other the same species, if they somehow met. 2601:1:9280:155:412E:EB34:E60A:B742 (talk) 08:40, 17 May 2014 (UTC)

Infinite Expectation

I added a section on Infinite Expectation as a Rebuttal, although perhaps it should be a remark, since it doesn't directly attack the central claim of the DA. Please feel free to clean up my /math bits. -- Marcus 11 July

Hi Marcus :) I think this argument doesn't work because the distribution of f in [0,1] is actually discrete. For instance, if there are 10 people, f can only be 1/10, 2/10, 3/10, etc. Thus no ln(0). The integral is just an approximation for large N, but it's never exact. Would it be ok to remove the section or at least add a note about this? In any case, if the argument hasn't been made in the literature, it doesn't meet Wikipedia's policies... Cheers! Brian Tomasik (talk) 02:48, 13 June 2014 (UTC)

I just don't get it

I'm not quite smart enough to understand most of the article, if I'm honest, but I've never understood why anyone even comes close to seriously considering the Doomsday argument. The simple fact of the matter is that, when considering "you" or the people alive around you, you/they are not special. The article says:

Simply put, it says that supposing that all humans are born in a random order, chances are that any one human is born roughly in the middle.

Which is reasonable enough (even excusing the "born in random order" part, since really you are a product of time in which you are born). But surely the issue is that when running through the argument, you don't (and can't) pick one human at random. You consider yourself, or those alive around you, which is surely just an accident of whenever you happened to be born.

In other words, someone had to be born at your position in the grand list of all humans ever, whether that's at the 1% point or the 5% or the 99% point. It's like shuffling a deck of cards and declaring the resulting order to be miraculous because it's so unlikely (what's the formal name for that?).

I don't see anything covering in this angle in the article, and of course I wouldn't expect anyone to add it based on my vague ramblings here, but am I missing something? Is it covered in one of the other rebuttals, but I just can't see it? — Preceding unsigned comment added by 83.137.249.20 (talk) 11:39, 30 March 2015 (UTC)

layman's description

A lot of people on this talk page seem to be missing the point. Perhaps we could find a simple example somewhere? To me it's roughly like this: The sign to a fast food restaurant says "serving seven hundred billion customers." You go up to the counter and say, "wow seven hundred billion, really? You guys must be great." The clerk replies, "well haven't served them all yet, but we plan to." You say "oh, how many have you served so far?" The clerk says, "six." You say "six hundred billion?" "No, just six." Now at this point, you have to wonder, are you very lucky or is the clerk wrong? The DA is basically this thought process applied to population. There are obvious issues with formalizing such an argument, and it has limited use. But it's not clearly 'garbage', 'science', 'pseudoscience', etc. it's a meaningful and interesting proposition.

Epistomology doesn't have to be meaningful or interesting to pull people into debating it ;)

To me it seems like they're missing a step towards causation - the fact that you are only the sixth customer does not in itself warrant some prediction that they won't reach seven hundred billion others. What warrants that is their financial backing, their ability to make business decisions to accomplish this, etc.

For your restaurant, they could in fact have them lined up waiting to be served, and the probability of that isn't easy or even meaningful to determine - your "luck" has no part in the outcome of reality, and reality doesn't care what you consider 'lucky'. There are enormous amounts of extremely unlikely events that happen that don't relate at all to human experience, good or bad, and we don't call ourselves lucky because of them. They just are. You could be the first or the last human on Earth, or a particular rock on a particular moon of Saturn. Either could be considered 'lucky,' but either could also be considered meaningless. Given the size of the universe, it's extremely 'lucky' that we are presently here, where Earth is, instead of two millimeters or three feet or seven billion AU away from where we are now. etc. Hopefully you see what I mean.

There isn't even any 'future' population to select randomly from. The probability that you were born in this era is 100%, because this is the era we are currently in, and you are alive. The proposition makes the mistake of assuming the future or the past is something we can select from statistically, but a closer definition of reality makes it an empty set. There is no doomsday, just like there is no 18th century to go out and count.

Eagleon (talk) 22:37, 26 September 2015 (UTC)

Possible Addition?

Anyone want to add a short section on the "New Doomsday Argument"? Seems like the basic idea is that just as other animals are unaware of risks that we've identified, there could be risks that even more intelligent beings could recognize that we can't (cognitive closure). Furthermore, these risks could range from unlikely to extremely probable, given the observation selection effect. Seems like something worth talking about. http://ieet.org/index.php/IEET/more/torres20160222 — Preceding unsigned comment added by 24.211.151.68 (talk) 20:19, 14 April 2016 (UTC)

Done. Carnaptime (talk) 01:32, 26 April 2016 (UTC)

What if...

What if we are all... in first 5%?

That's covered in the article, Doomsday argument#Rebuttals. Gap9551 (talk) 21:00, 27 June 2016 (UTC)

Seems a bit synthesized (August 2016 synth/primary/morerefs tagging)

There are long multi-paragraph sections that summarise a primary source. These seem to me in great danger of synthesis/OR. There needs to be a lot more third-party references, directly to Bostrom's argument in particular rather than just the mostly-tangential ones present, showing that these long sections should even be here - David Gerard (talk) 19:54, 16 August 2016 (UTC)

External links modified

Hello fellow Wikipedians,

I have just modified 2 external links on Doomsday argument. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:

When you have finished reviewing my changes, please set the checked parameter below to true or failed to let others know (documentation at {{Sourcecheck}}).

Y An editor has reviewed this edit and fixed any errors that were found.

If you have discovered URLs which were erroneously considered dead by the bot, you can report them with this tool.
If you found an error with any archives or the URLs themselves, you can fix them with this tool.

Cheers.—InternetArchiveBot (Report bug) 19:35, 15 December 2016 (UTC)

95%?!

"If Leslie's Figure[5] is used, then 60 billion humans have been born so far, so it can be estimated that there is a 95% chance that the total number of humans N will be less than 20 × 60 billion = 1.2 trillion. Assuming that the world population stabilizes at 10 billion and a life expectancy of 80 years, it can be estimated that the remaining 1,140 billion humans will be born in 9,120 years."

Why does this example have any meaning? This is not "95% chance", this is 95th percentile. You can also say that there is a "less than 99.99....9% chance" that the total number of humans will be less than a quadrillion, and then extrapolate that we'll survive for a whopping ten million years (or less). A better idea of an estimate would be to put us right now at the 50th percentile; i.e. there are six generations left to live, or doomsday will on average arrive around AD 2500. 93.139.58.82 (talk) 22:16, 26 June 2016 (UTC)

95% is commonly used for Statistical significance, in the sense that something that has less than 5% chance of happening is considered too unlikely to be due to chance. But ultimately it is an arbitrary number. The argument would be the same if you'd use 90%, 99.9% or even 50%. But something with 50% chance happening is not weird. Gap9551 (talk) 20:59, 27 June 2016 (UTC)

It's important to highlight that the choice of percent chance does affect the argument in some sense. To be specific, it affects our degree of confidence in our upper bound. Morganrconnolly (talk) 21:23, 20 May 2017 (UTC)

This Whole Thing Breaks: Future Tech Breakthroughs

Unknowable breakthroughs in physics applications in the future, at the very VERY least, make this laughable at best. To say nothing of colonizing mars. This entire argument is a house of cards that may seem shiny to some. — Preceding unsigned comment added by Sinsearach (talk • contribs) 17:12, 16 July 2017 (UTC)

My Rebuttal

The rebuttals here are of MUCH better quality that those in the actual artical. The ones in the artical are barely worth reading, by comparison, because they don't deal with the essential problem.

In my opinion, it boils down to a single paragraph:

"Let us further assume that our fractional position f is uniformly distributed on (0,1] even after we learn of our absolute position n. This is equivalent to the assumption that we have no prior information about the total number of humans, N."

I don't think the second sentence actually follows from the first. The first sentence is not presumable if we have no information about N; if we have no information about N, f will have only a few possibilities near 1 (N=n+1, N=n+2, N=n+3...) while it will have a huge number of possibilities near 0, because this is what f approaches as N approaches infinity.

Thus f is nowhere near uniformly distributed; it has an infinite number of possible values that are below any given fraction, and only a finite number of possibilities above that fraction. This reflects the fact that N has an infinite number of possible values above any number one might select, and only a finite number of possibilities below.

It's logically correct. The apparent problem is the result of assuming the Copernican principle alone, and ignoring all other information we might have about the human race. Note also that since the only input to the problem with units of population is the number of humans already born, we expect any calculation of the future humans born to be proportional to that input.121.2.74.60 (talk) 05:25, 19 November 2017 (UTC)

Caves' Rebuttal

I'm having a hard time following the article's explanation for why Caves' Rebuttal is flawed, but it seems immediately obvious to me that his choice of 50 year is a glaring flaw, because the median age today is actually 26. One would be most likely to stumble into a birthday of someone in their mid-late 20s, not 50. Am I right that the explanation in the article is basically just explaining that in a roundabout way, and if so, couldn't it be simplified to just say that? Or is this just me not understanding something? Thanks. — Preceding unsigned comment added by 223.136.202.6 (talk) 00:48, 10 December 2019 (UTC)

Last Section

The last section of the article (5.2) simply ends with an equation. This seems like "bad style" to have the final section finish without a completed statement regarding the relevance of the equation, and also to have the complete article end without a concluding, or closing statement. SquashEngineer (talk) 14:52, 9 March 2020 (UTC)

The Doomsday argument as a tricky problem

I was removing junk from the article and found that this section is redundant to previous ones, but good enough to preserve on this talk page. –LaundryPizza03 (d c̄) 02:48, 26 March 2020 (UTC)

Extended content

Sometimes, the Doomsday Argument is presented as a probability problem using Bayes’ formula.^[1]

Hypotheses

Two hypotheses are in competition:

The theory A says that humanity will disappear in 2150,

and the theory B says that it will be much later.

Under assumption A, a tenth of humanity was alive in the year 2000, and humanity has included 50 billion individuals.

Under assumption B, one thousandth of humanity was alive in the year 2000, and humanity has included 5 trillion individuals.

The first theory seems less likely, and its a priori probability is set at 1%, while the probability of the second is logically set to 99%.

Now consider an event E, for example: "a person is part of the 5 billion people alive in the year 2000". One may ask "What is the most likely hypothesis, if you take into account this event?" and apply Bayes' formula:

\mathbb {P} (A\mid E)={\frac {\mathbb {P} (E\mid A)\cdot \mathbb {P} (A)}{\mathbb {P} (E)}}

According to the above figures:

\mathbb {P} (E\mid A)=10\%\ ,\ \mathbb {P} (E\mid B)=0.10\%

Now with :

\mathbb {P} (A)={\frac {1}{100}}\ ,\ \mathbb {P} (B)={\frac {99}{100}}

We get :

\mathbb {P} (E)=\mathbb {P} (E\cap A)+\mathbb {P} (E\cap B)=\mathbb {P} (E\mid A)\cdot \mathbb {P} (A)+\mathbb {P} (E\mid B)\cdot \mathbb {P} (B)={\frac {19.9}{10000}}

Finally the probabilities have changed dramatically:

\mathbb {P} (A\mid E)={\frac {10}{19.9}}=50.25\%

\mathbb {P} (B\mid E)={\frac {9.9}{19.9}}=49.75\%

Because an individual was chosen randomly, the probability of the end of the world has significantly increased.

Attempted Refutations

A potential refutation was provided in July 2003:^[2] Jean-Paul Delahaye showed that Bayes' formula introduces "probabilistic anamorphosis", and demonstrated that Bayes' formula is prone to misleading errors made in good faith by its users. In 2011,^[3] Philippe Gay showed that many similar problems can lead to these mistakes: each change of a weighted average by a simple one leads to odd results.

In 2010,^[4] Philippe Gay and Édouard Thomas described a slightly different understanding: the formula must take into account the number of humans involved in each case. These explanations show the same algebra:

\mathbb {P} (B\mid E)={\frac {0.1\%\times 5\cdot 10^{12}\times 99\%}{0.1\%\times 5\cdot 10^{12}\times 99\%+10\%\times 50\cdot 10^{9}\times 1\%}}={\frac {99\%}{99\%+1\%}}=99\%=\mathbb {P} (B)

Using a similar method, we get:

\mathbb {P} (A\mid E)={\frac {1\%}{99\%+1\%}}=1\%=\mathbb {P} (A)

Copernican Principle/Mediocrity Principle actually suggests several contradictory things

The "number of individuals" metric seems arbitrary to me. A purer reading of the copernican principle would suggest that since anatomically modern humans have been around for about 300000 years then they will remain for another 300000 years. Additionally, the principle could also be used to argue that the human race will last the average duration of a primate species, OR of a keystone species OR of an apex predator species 2600:1002:B013:926B:217C:4E5E:D733:1305 (talk) 20:40, 3 April 2021 (UTC)

^ "Logique, informatique et paradoxes" Jean-Paul Delahaye, Belin, pages 30-32
^ "La Belle au bois dormant, la fin du monde et les extraterrestres", Jean-Paul Delahaye, Belin, Pour la science, juillet 2003, pages 30-32, "http://www.lifl.fr/~delahaye/pls/107.pdf" (PDF). {{cite web}}: External link in |title= (help)
^ "L’Argument de l’Apocalypse… selon la Répression des Fraudes|collection" Philippe Gay, Image des mathématiques (CNRS), août 2011, http://images.math.cnrs.fr/L-Argument-de-l-Apocalypse-selon.html
^ "Détournements de Bayes" Philippe Gay and Édouard Thomas, Tangente, septembre-octobre 2010, n°136

[1] "Logique, informatique et paradoxes" Jean-Paul Delahaye, Belin, pages 30-32

[2] "La Belle au bois dormant, la fin du monde et les extraterrestres", Jean-Paul Delahaye, Belin, Pour la science, juillet 2003, pages 30-32, "http://www.lifl.fr/~delahaye/pls/107.pdf" (PDF). {{cite web}}: External link in |title= (help)

[3] "L’Argument de l’Apocalypse… selon la Répression des Fraudes|collection" Philippe Gay, Image des mathématiques (CNRS), août 2011, http://images.math.cnrs.fr/L-Argument-de-l-Apocalypse-selon.html

[4] "Détournements de Bayes" Philippe Gay and Édouard Thomas, Tangente, septembre-octobre 2010, n°136

[1]

[2]

[3]

[4]