Re: Doom2k

From: Fred Chen <>
Date: Mon, 13 Dec 1999 01:50:30 -0800

Jacques M. Mallah wrote:

> On Tue, 7 Dec 1999 wrote:
> > Suppose there are two possibilities: you live in a universe where there
> > will be 100 billion people total, or in a universe where there will be
> > 100 trillion people total, and a priori you think there is a 50-50 chance
> > which one is the case. You check your birth order and find that you are
> > about number 50 billion.
> >
> > Now, that would be pretty likely if you were in the 100-billion universe,
> > but it would be very unlikely if you were in the 100-trillion universe.
> > Hence by Bayesian reasoning you find you are more likely to be in the
> > 100-billion universe, and therefore the human race is likely to end
> > relatively soon. This is the Doomsday argument.
> >
> > However introducing the all-universe model and the self-selection
> > assumption (that you are a random individual from among all individuals in
> > all universes) then a priori the chances that you are in the 100-trillion
> > universe are ten times greater than that you are in the 100-billion
> > universe. This exactly counters the shift which you made in the Doomsday
> > argument, based on your birth order, which made you think you were more
> > likely to be in the 100-billion universe.
> The Doomsday argument still works. The uncertainty is not which
> "universe" you're in; as you say, if both universes exist and you know
> that, there's no Doomsday argument. But the thing is, you don't know
> that. Suppose there are N "universes" that all exist. Some X of them
> have 10^11 people, (N-X) have 10^14, but you don't know what fraction X/N
> is.

Generally, and conservatively, this is true. For AUH, however, N>>1
and population size is allowed to vary freely among the universes, so we
would expect many, many more universes with populations maximally
exceeding the current population, i.e., as large as possible within constraints
of cosmic evolution.

> If your number is 5*10^10, this suggests X/N is large: Doomsday.

X/N is not known, as you said. What I understand you to mean is that
for N=1, 10^11seems far more likely than 10^14, according to the
Doomsday argument. But the most likely number of people should be 5*10^10,
in this case.

(Now, the current world population growth rate is ~5 people every 2 seconds.
So each moment we find ourselves in a less and less likely position in the
of the human race.)

> Of
> course, if you could calculate X/N from first principles, there would be
> no argument. The one-world case is just N=1; again, if you could
> calculate whether X=0 or X=1 in this case, there would be no argument.


> - - - - - - -
> Jacques Mallah (
> Graduate Student / Many Worlder / Devil's Advocate
> "I know what no one else knows" - 'Runaway Train', Soul Asylum
> My URL:
Received on Mon Dec 13 1999 - 01:53:03 PST

This archive was generated by hypermail 2.3.0 : Fri Feb 16 2018 - 13:20:06 PST