Is the Carter(Doomsday) hypothesis valid?

[DAL59]

I say no because you could apply it to anything.  If you start working for a new car manufacturer, and see a car with the registry number 10, you don't assume the factory will only produce a dozen or so new cars.  

Note: This is going to be like the airplane treadmill thread, isn't it.   

[Mad Rocket Scientist]

Isn't that just because you have assumptions about how many cars car manufacturers will make? Since there's other car companies, it's reasonable to assume that this isn't one of the few cars that came from a car manufacturer that only made 20 cars. It's not really possible to compare that to civilizations, since we've never seen any other than our own.

It's like the thought experiment where you wake up in a room with no windows and no doors and no memory of how you got there, and so assume you're in a large city, since that's where people are statistically likely to be. It's highly likely to be inaccurate, but it's the only thing you have to go on.

[mikegarrison]

It seems silly to me. I think the problem has to do with an inappropriate statistical model. Humans are not randomly distributed over time.

 

[Bill Phil]

It assumes a finite number of total humans distributed over time.

Even more so, it isn't really a doomsday hypothesis, just an interesting little thing.

[mikegarrison]

13 minutes ago, Bill Phil said:

It assumes a finite number of total humans distributed over time.

Even more so, it isn't really a doomsday hypothesis, just an interesting little thing.

Well, in fact it seems unlikely that there will be an infinite number of humans. That means that humans will come to an end sometime. Thus it's not surprising that the analysis would predict this -- it's built in to the input assumptions.

What's more questionable is the rather specific prediction of just how many humans will come after us.

[Superfluous J]

By not all dying, we've proven that anybody who had this thought any time in the past 10,000 years or so would have been wrong.

So if the hypothesis has been wrong for 10,000 years, statistically it's pretty unlikely to be correct now. And the longer we go without it happening, the less likely it will be to be true.

I just saved Humanity for all time.

[Bill Phil]

46 minutes ago, mikegarrison said:

Well, in fact it seems unlikely that there will be an infinite number of humans. That means that humans will come to an end sometime. Thus it's not surprising that the analysis would predict this -- it's built in to the input assumptions.

What's more questionable is the rather specific prediction of just how many humans will come after us.

There won't be an infinite number of humans, but we're still making some big assumptions that make the hypothesis almost irrelevant.

Eventually we die off or evolve. And both may very well happen...

[YNM]

We don't know. We will never know until it's here, looming before our eyes.

 

I do have my beliefs that have a concept of doomsday.

 

3 hours ago, 5thHorseman said:

So if the hypothesis has been wrong for 10,000 years, statistically it's pretty unlikely to be correct now. And the longer we go without it happening, the less likely it will be to be true.

Yeah, like many natural disasters...

 

We just need the sun to go bang I guess.

[magnemoe]

12 minutes ago, YNM said:

We don't know. We will never know until it's here, looming before our eyes.

I do have my beliefs that have a concept of doomsday.

 Yeah, like many natural disasters...

We just need the sun to go bang I guess.

Sun only goes bang once and its an very long time to. 

Earthquakes comes in groups until ground stabilizes to an new level then build up again. 
Crime follow the same pattern because of thieves opperate in an area before moving on for violent crimes you tend to have copycats. 
Accidents follow the other pattern after an accident like event the chance of an repeat goes down until people get less safety oriented again. 

 

[YNM]

16 minutes ago, magnemoe said:

Sun only goes bang once and its an very long time to. 

Earthquakes comes in groups until ground stabilizes to an new level then build up again. 
Crime follow the same pattern because of thieves opperate in an area before moving on for violent crimes you tend to have copycats. 
Accidents follow the other pattern after an accident like event the chance of an repeat goes down until people get less safety oriented again. 

Hence you will only know it once it's right before your eyes.

Within my beliefs, there are signs. But we get complacent and ignore them, very often.

So it'll only be surprising because we always live in the now, not the past or the future. Until that "now" is the time itself we'll never know.

But one thing's sure : as time passes, it gets closer.

 

It's like the extinction of near-extinct species : while there are a few pairs, you'll say "ah, sometime else"; only when you have passed a few decades after the last sighting, you'll realize they're gone forever.

Edited May 22, 2018 by YNM

[Gargamel]

@DAL59, hows about a link that describes this? 

[kerbiloid]

Does this  hypothesis work if take 200 000 years ago, when there were 10 000 or so living humans?
Should the humanity die 150 000 years ago?

[NSEP]

2 hours ago, Gargamel said:

@DAL59, hows about a link that describes this? 

https://rationalwiki.org/wiki/Doomsday_argument

[K^2]

33 minutes ago, kerbiloid said:

Does this  hypothesis work if take 200 000 years ago, when there were 10 000 or so living humans?
Should the humanity die 150 000 years ago?

But you're not living 200k years ago. You are living now. And the probability that you'd be living 200k years ago is vanishingly small, as vast majority of human population has lived or live much after that point. Of course, the argument can be wrong for us as it was for the few unfortunate who got to live back then. That's the whole point of a statistical argument. It's only right most of the time. Or, in this case, for majority of humans, whenever they happen to live. We can cross our fingers and hope that we're still in the tiny fraction of all humans that will live millennia from now across the stars, and that we just pulled a short straw to live in this era. There are two alternatives, however, that are far more likely. Statistically, either population will continue to grow, and then it's likely that we'll face massive extinction in very near future. Or population is going to level off, and we're about to enter millennia of stagnation.

Of course, you hardly need to understand statistics to say that these two are the most likely outcomes.

[kerbiloid]

2 minutes ago, K^2 said:

Of course, you hardly need to understand statistics to say that these two are the most likely outcomes.

https://www.stat.ubc.ca/~rollin/teach/643w04/lec/node46.html

[magnemoe]

34 minutes ago, kerbiloid said:

Does this  hypothesis work if take 200 000 years ago, when there were 10 000 or so living humans?
Should the humanity die 150 000 years ago?

This, do not see why you could not go back to any point in time and get the same result.

Now 3 relevant counters:
Human population is likely to peak in the 2050-280 period, based on population trends. We however don't know the trends for over 100 years.
in the multi thousand year perspective here its likely that humans will have self supplied colonies other places in space. 
Finally will people thousands of year from now even count as humans, generic engineering is the obvious one. 
 

[Racescort666]

Consider this: humans like all life on earth and presumably elsewhere is constantly evolving. So at what point did Homo heidelbergensis start becoming Homo sapiens? Then at what point does Homo sapiens start becoming something else? Have all of the humans died off after this? At what point does the population stop becoming human and start becoming something else? It's not really a doomsday scenario in this case just an estimate of evolutionary progress.

Also worth noting: correlation =/= causation. 

[YNM]

51 minutes ago, Racescort666 said:

correlation != causation.

But events are events. The correlation still exists whether it's really the cause or not.

2 hours ago, magnemoe said:

This, do not see why you could not go back to any point in time and get the same result.

True, back then it seems as pointless as it is now.

But again, just because the winds wasn't strong back then it couldn't be stronger right now.

2 hours ago, K^2 said:

That's the whole point of a statistical argument. It's only right most of the time.

^^ this ^^

The answer is : we never know, until it is immensely obvious otherwise.

[Superfluous J]

10 hours ago, YNM said:

Yeah, like many natural disasters...

I'm not big on watching random YouTube videos so I'll just assume it's some nutjob saying certain natural disasters haven't happened recently so won't happen at all or something like that.

I'm not saying there's not a problem. I'm saying the hypothesis has a flaw.

There's better statistical evidence that we're living inside a simulation.

[kerbiloid]

11 minutes ago, 5thHorseman said:

There's better statistical evidence that we're living inside a simulation.

And simulator looks like this

Spoiler

 

[magnemoe]

6 hours ago, Racescort666 said:

Consider this: humans like all life on earth and presumably elsewhere is constantly evolving. So at what point did Homo heidelbergensis start becoming Homo sapiens? Then at what point does Homo sapiens start becoming something else? Have all of the humans died off after this? At what point does the population stop becoming human and start becoming something else? It's not really a doomsday scenario in this case just an estimate of evolutionary progress.

Also worth noting: correlation =/= causation. 

Yes, at some time back they stopped being that we define as humans, same in the future but ignore evolution, genetic engineering is much faster and after an number of interaction they would not be able to breed true with us.

However this is probably overkill as in close in air support by a deathstar. The way we think has changed a lot simply because of culture and education, simple example:  anybody before the doomsday hypothesis was formulated could not ask the question 
 

[DAL59]

6 hours ago, Racescort666 said:

At what point does the population stop becoming human and start becoming something else? It's not really a doomsday scenario in this case just an estimate of evolutionary progress.

It doesn't matter what the start point is.

[wumpus]

Last I heard, the chance of any species to go extinct was pretty much a constant.  Unfortunately, I can't begin to remember where I learned that little bit (the author of my lost source seemed to agree that it was far bigger than the evolutionary biology community thought at the time).  If this is true, it pretty much kills the Carter "logic".  Trying to sort historical extinction from human caused extinction is pretty difficult, although there is little reason to believe that humans are wiping out newer species than old ones (grasses are pretty new and doing quite well in the anthrosene).

[kerbiloid]

This is the bottleneck of this "hypothesis".

Quote

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.

XX+XXI ~= 120 years ~= 5 generations. Population 2.5→7.5 bln, average 5 bln, so ~25 bln humans living in XX..XXI. i.e. ~1/3 of all humans ever lived.
What are chances that a random human indeed is born "roughly in the middle", not 100 000 years ago?
I'm afraid that inaccuracy is wider that total range, so a prediction power of this hypothesis doesn't significally differ from pure random.

I.e., this hypothesis is valid with invalid probability.

Edited May 22, 2018 by kerbiloid

[YNM]

7 hours ago, 5thHorseman said:

I'll just assume it's some nutjob saying certain natural disasters haven't happened recently so won't happen at all or something like that.

I'm not saying there's not a problem. I'm saying the hypothesis has a flaw.

Fine. It was about Hawkes Process.

But what's the flaw ? Is it a flaw within the finiteness of anything ?

(also, I think I kind of misunderstood that just because something has a probability of something with certain return period doesn't mean they have to occur at such intervals. But it is true to have them consecutively would still be statistically less likely.)