Re: Ratchets against devolutionary speciation

Earle Jones <earle.jones@xxxxxxxxxxx> wrote or quoted:
> > > "r norman" <NotMyRealEmail@xxxxxxxxxxxx> wrote in message

> > > > The real problem with devolution is probabilistic. Given that
> > > > mutations are random (sorry, John Edser, you are likely to grumble
> > > > here), then evolution can be modeled (here, too) as a random walk
> > > > process with a very large number of dimensions. And these, simply
> > > > put, do not return to the origin.
> > > >
> > > > In other words, put aside adaptation and selection for the moment.
> > > > The probability of a mutation producing a change is relatively high
> > > > (but not one). After a bunch of these occur, the probability that a
> > > > series of mutations will exactly undo what was changed is so low as to
> > > > be zero.
>
> *
> In fact, the average distance from the origin is proportional to
> sqrt(n), where n is the number of steps.

In high-dimensional spaces, the chances of getting back where you started
anytime soon via a random walk diminish very rapidly with the number of
steps taken so far - just as was claimed.
--
__________
|im |yler http://timtyler.org/ tim@xxxxxxxxxxx Remove lock to reply.

.