See the article by Tomasz.Wegrzanowski@gmail.com   (Spell that 5 times
   really fast!)
   [1]http://t-a-w.blogspot.com/2013/11/if-assassination-markets-were-real
   .html
       I'd praise this guy's take on assassination markets, but since 90%
   of it is virtually identical to my own from January 1995 (before I'd
   published Part 1 of AP), that would pretty much be me patting myself on
   the back.
       Even so, there are some things on which we disagree:  For example,
   TAW (Tomasz Wegrzanowski) says he believes that today's 'assassination
   markets' are probably scams.  (I know only of Sanjuro's "Assassination
   Market", and David's DPL, or Digital Prediction Lottery; Technically,
   DPL doesn't label itself as an 'assassination market', but for purposes
   of my analysis I will label it one.)  It is hypothetically possible
   that either or both of these are some government honey-pot, although I
   don't think so.  I have previously pointed out that I see little reason
   for any government to, now, promote a system similar to my AP idea,
   particularly when I wasn't promoting AP itself since my release from
   the Federal governments' 'gated communities' in March 2012.  Indeed, to
   do so would have been about as much a mistake as 2003's FutureMAP
   system.
   [2]http://en.wikipedia.org/wiki/Information_Awareness_Office#Futures_Ma
   rkets_Applied_to_Prediction_.28FutureMAP.29      (And I'm not implying
   FutureMAP was a bad idea; rather I think that to float the idea as they
   did led to so much negative reaction that they had to withdraw it,
   amazingly within one day too!)
        But I also don't believe that either system is likely to be a
   money- (Bitcoin-) seeking scam, either.  The reason is based on Game
   theory.  Imagine somebody starts up an AP-type operation.  Imagine
   further that it is acting is what I'd call a 'healthy' fashion,
   actively collecting at least dozens of hundreds of individual donations
   per day.   (This would be strongly impeded if the system collected
   minimum  donations of 1.0 BTC, like Sanjuro's system started out doing
   , or it was somehow not sufficiently easy to use, or it did not provide
   the various proofs of operational security sufficient to induce
   confidence in potential donors.)  I think if it was 'working', such a
   system would quickly develop hundreds of target-names, albeit some with
   tiny reward amounts,  and would certainly collect enough donations
   within, say, three months to raise the bounties to a level sufficient
   to result in at least one, uh, 'hit'.  Prior to that, many potential
   donors would still hold back, either because they weren't confident
   that somebody would be rewarded, or they would wait to see if something
   'happened'.
        But suppose that one 'hit' occurred.  Suppose the system
   demonstrated that some unknown-named person had correctly predicted the
   date of death of the person listed. (This should be unavoidable:  It
   should not be possible for the system to conceal the existence of a
   correct donation; in any case, the public and all potential future
   donors would insist on the exposure of all relevant donations naming
   that particular target.) Would the operator 'defect' or 'cooperate'
   (both are terms used in Game Theory.)  Would the operator of that
   system actually pay the stated  amount, or would he refuse?  (I am
   assuming that the system was at least set up to have the technical
   ability to prove, to the public, that the reward had been paid to the
   correct predictor.)
       To me, it is obvious that if the operator indeed pays the reward,
   and this fact can be proven to the public without revealing the true
   name of the correct predictor, for every Bitcoin paid out as a reward,
   this confidence-building measure will result in at least tens, and
   probably many hundreds of Bitcoins of further donations from other
   themselves naming other targets.  In other words, at that point it
   would become totally illogical for the system operator  to 'defect', to
   refuse to pay a (relatively) small donation, because by doing that he
   would be foregoing further donations that would be collectively
   hundreds of times larger.  (Nobody would donate to a system which
   provably refused to pay on a correction 'prediction', or failed to
   sufficiently disprove the non-payment of the reward.)
         This same analysis would apply after the second 'hit', and the
   third, and the fourth, etc.  At no point, I think, would the
   'assassination market' system operator find it financially more
   advantageous to refuse to pay the reward, than to pay it and remain in
   operation and collect far more donations.  He will always be better off
   by paying the reward.
       There is, however, one exception that I can think of to this
   analysis.  A person (or persons, or a government) might conceivably run
   an 'AM'-type organization, for a while, specifically for the purpose of
   discrediting the entire concept of AP.  That person might run the
   system, for a while, even paying off one, or a dozen or two,
   predictions, and collect a substantial reward in the process,
   eventually intending to run away, thereby leading to 'fear,
   uncertainty, and doubt' over the question of whether a credible
   'AP'-type system could exist.  The big problem, for that person, is
   that by doing so he would be inadvertently demonstrating the proper
   (for awhile) functioning, and effect of, that system.  Once
   demonstrated to work, for that period of time, it would become obvious
   that such a system could, indeed, function 'properly' if the person or
   people involved are motivated to do so.  And, there would be nothing
   they could do about stopping credible, reliable people from initiating
   and running an honest AP-type organization.  Eventually, reputable
   systems would win out, because they would build up the longevity
   leading confidence as to their offers.
              Jim Bell

References

   1. http:///
   2. http:///