[extropy-chat] Re: Formulating a bet

[Brett Paatsch]

Robin Hanson wrote:

> At 10:32 AM 3/24/2005, Brett Paatsch wrote:
>>>What if I were to define "cryonics can work" as something like, a
>>>person frozen with today's technology is successfully revived within,
>>>say, 100 years, with substantially identical memories and personality.
>>
>>In a real money version of idea futures I'd be interested in taking you
>>up on your bet but I'd have the following issues with it which I wonder
>>if we can address together (also with Robin if he is interested).
>>
>>1) 100 years is too long for me to have to wait to get a financial return
>>on investment. ...
>>2) Robin said "I'd want to be clear if uploading counts as revival, but
>>otherwise, sure I'd accept many third party judgements."
>>... I would have a problem with [judges] being required to pre-accept
>>that uploading is possible. ...
>>3) ... There is not a lot of distance between 0 in 100 and 1 in 100
>>for a payoff to be achieved by either party. ...
>
> You can bet on anything you can get judges to decide, which is pretty
> much anything.  The question is how well that judge decision correlates
> with the real dispute you are having.

Well if you've signed up for cryonics, it seems implicit that you and I do
disagree on the identity aspect for one thing.  I think that we'd have to
tighten the wording on Hal's bet a bit to see if we'd disagree on the
probability, given identity, and given uploading counts as revival.

For instance what's "substantially" identical memories and personality?

Hal probably knows what he means, but I'm not sure you or I know
what he means well enough yet to come up with a probability and so
to see where we'd stand in relation to each other. Perhaps we would
not disagree other than on the identity question.

Care to define any terms in Hal's bet as you like and then assign a
probability and we'll see if our disagreement is in an area other than
identity?

How do you read "substantially" identical memories and personality
for instance?

> ...................................................... Perhaps you can be 
> clever, but if
> the correlation won't happen for 100 years, then you just have to wait
> that long.

Yes, if.

Hal said. "say, 100 years". So I see that 100 years is somewhat arbitrary
If without doing violence to the rest we could bring that timeframe forward
(and Hal could alter his odds according) to when one of us might expect in
the normal course to still be alive, then that would be slightly more 
interesting
for me. Yet I can imagine that Hal would reasonably envisage success as far
more likely to come later than sooner. Given 150 years I wonder what odds
Hal would put? (Hal?)

> With combinatorial markets there is less of a problem of a "small
> distance" at the extremes.  You can "reuse" your bets to bet on other
> topics via conjunctions of this topic and those other topics.

I noticed that PAMs was about combinatorial markets but I haven't
yet read your relevant paper on it.

> If you and someone else are using words in different ways, you could
> bet on the "right" way to use those words.  But I'd rather bet on the
> concepts, so I'd rather try to clarify meaning ambiguities up front.

I understand that you see my preferred bet approach as conflating
two (at least) separate aspects.  Where practicable, and where the
objective is to eliminate disputes and find shared truth it makes sense
to get the ambiguous wording stuff out of the way.

I'm not sure my concern re identity amounts to mere wording concerns
though.

Regards,
Brett Paatsch