[extropy-chat] Re: Formulating a bet

[Brett Paatsch]

Hal Finney wrote:

> What I had in mind, regarding the "substantially the same person"
> test for a bet on cryonics success, was something like a modified
> Turing test. We would pose questions and get responses from the
> person both before and after the suspension.  These could then be
> randomized and someone who did not know which were which
> would try to distinguish the post-suspension responses from the
> pre-suspension ones.  If they were not able to do so, then we'd
> say that the personality and memories of the suspendee were
> substantially intact.

Yep, that would work. Or could be made to. Easier with biologicals
before and after perhaps but it could also work with biologicals and
uploads.

If you are still interested in formulating a bet, and I am, let's see if we
can agree on some broad stroke stuff first and then look at detail
later.

I am only interested in this so long as it is oriented towards a design
for a real bet. I don't mean that you and I should actually bet money,
though we might. I mean that we approach the design of the bet in such
a serious minded way that it could readily be turned into something that
could be ported to a FX or some similar (perhaps a real money equivalent)
or perhaps even sold as a bolt-on service to a betting exchange like
Betfair operating legally out of some jurisdiction like the UK. I think it
would be good, (i.e.. fun, interesting, ground-breaking) to have a for real
money market on cryonics somewhere.  If we develop the idea together
we might share IP rights in the bet or we might just give it away and
perhaps by doing it in public we can establish prior art ;-).

If you want to proceed, then, we can leave until last, the question of
identity, perhaps the bet can work as a trailblazing tool, (i.e..can still 
be
interesting to a wider market as a bet without it), and the identity bit,
is probably handleable as a design tweak in any case. i.e.. It might be
an important component for a market of wider interest but its not
important right now for us to get the other aspects of bet design that
don't depend on it out of the way.

One thing I'd like you to do, if you are willing, is to come up with an
estimate for the probability you'd place on your original wording
given 150 years rather than 100. The reason for this is that that would
give us a third data point (the first data point I'd take as today (2005)
p=0, the second, 2105, p=0.01, so the third data point would be 2155,
p = 0.?? ) and this would allow us to draw a curve of how you currently
see the probability varying with time.

The purpose of this curve (rather than a straight line which is all two
datapoints give us) is to give us a potential tool for coming forward
on the 100 years as well, whilst still allowing us to think about the
100 year point if that is where it is easiest for us to work. Sometimes
it may not be.

Being able to move our design frame of reference back and forward
along a time curve (two time curves as I'll give you mine as well if you
want to proceed and when I work it out) will help us in design and
shouldn't alter anything else important about the bet.

Regards,
Brett Paatsch