Re: poly: Idea Futures, some questions

Robin wrote:

> The generic situation is this: Someone has offered a reward for
> an idea, and competing teams can form to attempt to gain this reward.
> The reward is given to the first team to submit a valid entry.
> This model covers ad agencies, patent applicats, idea
> futures speculators, and probably much more.
>
> Each team should be concerned that, just before they complete their
> entry, one if their members will secretly sell their results to
> another team. Many approaches are available to deal with this
> potential problems:
> 1) Teams can monitor their members more carefully.

Even if team members were under close surveilance 24h a day, it would
still be difficult to prevent them from covertly communicating one
bit of information to an accomplice. They could have agreed in
advance that if the team member crosses a public place with his coat
under his left arm, that means "sell", and right arm means "buy".

> 2) Large costs can be imposed on entrants, inducing few entrants.
> Investigation of theft can then focus on a few suspects.

For example, only accredited financial institutions would be allowed
to own idea futures. That might work but it would severely limit the
charm of IFs.

> 3) These costs can be early and sunk. Then a theiving team would
> have to plan and count on stealing an idea at the last minute.

Or it could just opportunistically steal them if it happened to get a
chance to.

> 4) Entries could be required to include a history of their efforts.
> Theives would then have to make up a plausible history.
> 5) Entries could be required to included detailed "reasoning".
> Theiving entrants may have trouble generating different reasons.
>
> I think that Nick B. sees patents as including something like 4&5.
> But I think that any features of patents that seem like this are
> incidental. Patents require some detail in their applications in
> order to police patent violations, not to deter patent idea theft.
> I don't believe much patent theft is deterred due to the need for
> patent application detail, so I'm not very worried about similar
> theft with idea futures.

I'm not sure what you mean by "application detail" here. The ideas
behind a patent are typically complex and it would take many bits to
express them whatever were the regulations for patent applications.
Thus the people who wrote the patent laws might not have worried
about idea theft, because whatever exact regulations they
laid down, any workable patent system would have to require a
patent application to contain quite a lot of information --simply in
order to express the idea being patented --, and that would be more
than enough to deter patent idea theft.

Now, here are the replies to my objections that I would give if I
were you:

When many researchers have to work together in a team, then
individual researchers may steal information from the team and
investing in idea futures. However,

(1) There are some questions which don't require large teams.

(2) With questions that require large teams, we can use traditional
methods of funding research. We can still use idea futures to benefit
from the other advantages, such as a consensus estimate of the
probabilities of important issues.

(3) It is even an advantage that team members will leak information!
It results in more information quicker becoming available to society
and getting reflected in the price of the IFs.

(4) Even with questions that require large teams, part of the
research could still be funded with IF. Part of the salary to the
researchers could consist in the right to profitable insider trading.

(5) In some cases one could keep researchers isolated from the the
rest of the world during their workdays. At the end of the workday,
the management looks at the results of the day's work and buys or
sells any idea futures they may want before they let the researchers
leave the building. That way, there won't be any fresh information
that the researchers could leak.

_____________________________________________________
Nick Bostrom
Department of Philosophy, Logic and Scientific Method
London School of Economics
n.bostrom@lse.ac.uk
http://www.hedweb.com/nickb
Received on Tue Jun 30 22:22:34 1998