IDEA FUTURES …

                               IDEA FUTURES
                      Encouraging an Honest Consensus
                                            by Robin Hanson
                               Visiting Researcher, The Foresight Institute
                              P.O. Box 61058, Palo Alto, CA 94306 USA
                               hanson@charon.arc.nasa.gov 415-651-7483

                                                  These consensuses may be uncertain and
INTRODUCTION                                      temporary, and individuals may prefer to
      Are you fascinated by some basic            combine the results of several institutions,
questions about science, technology, and          but basic need for such estimates remains.
our future? Questions like: Is cryonics                For example, popular media chose
technically feasible? When will nano-             what they consider to be ranges of reason-
assemblers be feasible and how quickly will       able and noteworthy opinions on notewor-
resulting changes come? Does a larger             thy issues for presentation. Peer-reviewed
population help or hinder the world envi-         academic journals and societies offer more
ronment and economy? Will uploading be            detailed, though less accessible, consen-
possible, and if so when? When can I live         suses about which opinions are reasonable
in space? Where will I be able to live free       and in favor. Government agencies often
from tyranny? When will A.I.s be bucking          try to form and act on such consensuses, as
for my job? Is there intelligent life beyond      the U.S. E.P.A. does for pesticide health
earth? If you are like most Extropy readers,      risks. And there are many other such con-
such question matter to you.                      sensus institutions, such as opinion polls.
      Now how do we, as a society, go                  How well do these institutions work?
about answering such questions? People            How many of us are confident that, when a
who have an appropriate background, and           technical controversy arises, a widely visi-
who are interested enough in a particular         ble consensus will quickly emerge repre-
question, can research that subject in depth      senting society's honest best estimate on the
themselves, and come to a considered              issue, reflecting the relevant insights of the
opinion. And people who happen to know,           best relevant experts? Or that those with
respect, and trust such a person can simply       foresight will eventually be rewarded for
take those opinions as their own, avoiding        advocating positions which later become ac-
all the hard work. But what is everyone           cepted?
else to do, people whose actions often im-             People who have little contact with an
plicitly depend on such questions?                existing social institution, or who have a
      In practice, people usually defer to        position of power within it, may feel things
larger social institutions on most questions,     are basically okay. But those "in the
institutions which combine and evaluate           trenches" typically voice more skepticism.
contributions from many specialists, and          Your opinion on the trustworthiness of
which offer apparent institutional consensus      newspapers probably changed for the worse
estimates on many different questions.            the last time you read about an event in
                                                  which you were personally involved. Since
subjects like cryonics and nanotechnology              ideas evolve, many people try to reform
have often been unfairly treated by most               existing social institutions, and a few try to
current institutions, I expect at least a few          invent new ones [Li, Kan,Han88]. In this
Extropy readers to be dissatisfied with such           paper I suggest a new social institution,
institutions.                                          called "idea futures", which can create a
     Skeptics about current institutions are           visible expert consensus with clear
not typically focused on methods, often the            incentives for honest contribution.
center of philosophical discussions, but on
incentives. Skeptics see too many rewards              CONCEPT
for bias, and too few for honesty and care,                  Idea Futures is intended to aid the
and so distrust official statements. People            evolution of a wide range of ideas, from
often promote beliefs which serve their self-          public policy to the nature of the universe,
interest, and try to appear more confident,            and in particular should be able to help us
original, and knowledgeable than they are.             predict and understand our future. The ba-
People don't correct for standard human bi-            sic concept is to combine two phenomena,
ases [Kah], such as wishful thinking, over-            convergence and markets, and so make "a
confidence, and belief fixation. They mas-             futures market in ideas". Let me explain.
sage evidence, suppress criticism, and just                  Disagreement is rarely as fundamental
plain lie. Fashion, eloquence, and politics            as it seems. In the long run, beliefs often
often dominate expert consensus. Rewards               converge. For example, in science the
often go to those whose ideas are popular              steady accumulation of evidence eventually
now, rather than those who are later proved            settles most debates. We take the advice of
correct [Tul]. Paid advocates distort the              experts, indicating that we think we would
consensus we perceive by using raw media               come to believe what they believe, if only
exposure, bribes, and by exploiting human              we were to study what they have studied.
biases [Cia]. An honest consensus of rele-             Randomly selected juries usually reach a
vant experts is often lost from public view.           unanimous verdict, even more often than
     Many existing social institutions, such           seems rational. Theory [Se] and experiment
as investigative reporting, due process,               [Li] indicate that the people's beliefs should
public debates, and peer review, claim to              and do converge. In sum, we generally
address these problems. But there is room              trust in a convergence of human judgment.
for improvement. To improve the way
Figure 1 Some Controversial Claims
· By 2030, the greenhouse effect and other causes will have raised the average world sea levels by 1 meter.
· Cold fusion of deuterium in palladium can produce over 10 watts/cc. net power at STP.
· By the time we have surveyed our galaxy in the infrared to the 25th magnitude, there will not be any
  evidence of another technological civilization in our galaxy.
· By the time world GNP is four times the 1990 level, 1000 people will have physically lived in space over
  90% of the time for the previous seven years.
· By the year 2000, over 20,000 people in the U.S. will commute to work in Vertical Take-Off and Landing
  (VTOL) aircraft they park in their garage and drive down the street.
· The rest mass of the electron neutrino is greater than .01eV in ordinary space.
· If labor saving device X were widely used in industry Y, industry employment five years later would be
  less.
· Death-bed confessions or other evidence will eventually show that person X was murdered.
If people wait long enough for evidence to      about a future consensus, then futures mar-
surface and then take enough effort to study    kets might be able to create an immediate
and debate a specific enough claim, they        and honest consensus in ideas. If the mar-
often come to agree. When the people are        ket price for a "$1 if person on Mars"
reasonable, knowledgeable, and detached         coupon were 23¢, then that would typically
enough, and when they avoid subjects like       represent a consensus that there was about a
religion, they usually agree. When such a       23% chance of this happening. Anyone
group is diverse and independent enough,        could express their opinion on the subject
we believe we would probably also agree.        by trading coupons, or could just read the
      Markets are a way to create a consen-     "market odds" to see the best present esti-
sus about the value of an ownable item,         mate. This market consensus would com-
i.e., the "price". Futures markets are a way    pare favorably to other methods of forming
to create an immediate consensus about fu-      perceived consensus, such as by advertis-
ture consensus. For example, a market in        ing, opinion polls, or elite committees. An
corn creates a price in corn, so that most      idea futures consensus could be simultane-
buyers pay about that price. A futures mar-     ously open, egalitarian, universal, expert,
ket in corn creates a futures price, which is   honest, self-consistent, and cheap.
an immediate estimate of what the actual             A mature idea futures market could of-
price of corn will be in, say, nine months.     fer coupons on many claims about the fu-
Traders have clear incentives to make hon-      ture of technology and society. The con-
est contributions to the consensus; you "put    sensus prices would give describe a consis-
your money where your mouth is". A              tent set of probabilities for various possible
trader who believes the future price will be    future events, and conditional probabilities
higher than the market indicates buys, and      for some events given others. Investors
in so doing raises the consensus price.         there should be as diverse as investors are
Those who are right make money from             elsewhere, with a mix of short-term and
those who are wrong.                            long-term focuses, large investment houses
      Of course markets have limitations.       and daring do-it-yourself individuals, each
Ideally, items to value should be of wide       contributing their specialized knowledge
interest, exclusively ownable, cheaply          about an issue or the connection between
transferable, and have many identical           two issues to the total consensus.
copies. How can we apply this to ideas?              Like cryonics, idea futures is another
By creating coupons whose value depends         way to take advantage now of the fact that
on whether an idea is validated. For exam-      the future should be rich with power and
ple, a coupon which says "Exchangeable          knowledge. We create good incentives now
for $1 should a person land on Mars by          by letting the future settle our bets.
2020" is a direct tie between an idea, people        To make the whole idea more vivid let
on Mars, and money, a well-known unit of        us consider an simple (fictional) scenario.
incentive. Such coupons can be thought of
metaphorically as futures, and more literally   SCENARIO
as bets, a metaphor often used to describe            Pat Thgisni was not a model student.
both investments and science.                   A knack for making experiments work is
      If convergence creates a future consen-   probably what got him into graduate school
sus in ideas, and if futures markets can cre-   -- it certainly wasn't his grades. Worse, he
ate an immediate and honest consensus
was unkempt and had a disturbing habit of        science bets are illegal in Nevada} on his
bending people's ears with one harebrained       way back to Toledo. He paid $100 to have
idea after another. Definitely not one of the    a reputable judging group decide if it was
rising stars of the University of Toledo         precise enough for them to judge in 2013
Physics Department.                              (which it was), $20 to the Bank of Vegas so
      In his second year, 1992, Pat hit upon     they would issue coupons on it, and another
his best idea yet, "superscattering". If a       $20 to have a computer market set up.
neutrino could scatter off all the nuclei in a   Finally, he funded an automated market
crystal at once, the interaction could be a      maker with $200 in seed capital, and set the
billion billion times more powerful, perhaps     initial market odds at 30%. Back at the uni-
allowing neutrino telescopes [Web]. Pat          versity, Pat set his computer up to track the
showed his calculations to Prof. Ezra            market, and then spread the word, causing
Puccuts, a local and renowned neutrino ex-       an epidemic of giggles. One of Prof.
pert, though rusty on scattering. Prof.          Puccuts smirking students agreed to put up
Puccuts explained to Pat that a similar idea     $20 against him, and a half dozen other stu-
had occurred to him, but he had found it         dents joined in, mostly at $2 each.
conflicted with an accepted formula. Such a            It worked like this. While Pat bought
negative conclusion wasn't worth a publi-        coupons which said "$1 if superscattering"
cation.                                          from the market maker for around 30¢, the
      Pat persisted, however, bringing out       other students bought "$1 if no superscat-
his pages of calculations. After ten minutes     tering" from the market maker for around
of going through the first page, and finding     70¢. Whoever was right in the end would
three glaring, though irrelevant, math er-       make money on the deal, receiving the $1
rors, Prof. Puccuts lost his patience. "I do     plus 7% interest per year, compounded.
not have the time to correct your math for       Every time Pat bought a coupon the market
you", he declared, and shut the door. Over       odds went up, and every time the other stu-
the next few months Pat redid his calcula-       dents bought the odds went down. The
tions several times, but Prof. Puccuts was       market maker got these coupons from the
not interested and other professors referred     bank, who would sell the pair of coupons
Pat back to him. Pat submitted his work for      "$1 if superscattering" and "$1 if no super-
publication anyway, and then waited; he did      scattering" to anyone for $1 a pair. The
not have the $100K he figured it would cost      bank made money on transaction charges,
to do an experiment.                             and risked nothing because exactly one of
      That Christmas in L.A., Pat told the       the pair will be worth $1 (plus interest) in
story to his family. His brother Al, a sports    the end. The bank also bought such pairs
fanatic, suggested that Pat dare them to         back for $1, allowing people to sell
make a bet. Before Pat could object, Al de-      coupons back to the market maker. Many
scribed how idea futures were revolutioniz-      students took advantage of this feature in
ing the oil industry, and were a new way         the next few weeks, as some professors
for the little guy to contribute to the world    made it clear they were not amused.
of ideas. After a few more drinks, Pat saw             In 1995, Pat earned an early Ph.D.
the light.                                       Like most students, he could not find an
      Pat wrote up a precise statement of his    academic position and went to work in in-
claim, and then stopped by the idea futures      dustry. A year later he finally published his
mart in Las Vegas{footnote: Actually,            superscattering article in a vanity journal.
Over the years, Pat had tried to follow the     push Nevada's fledgling idea futures indus-
literature to see if anyone else had the same   try, and made an issue of the fact that the
idea, but without success. Meanwhile bets       markets had apparently discovered a num-
slowly trickled in, with the odds hovering      ber of potential new technologies. For ex-
around 15%. In 1997 the market told Pat of      ample, she said, if there was a 15% chance
another bettor, in Peru, who made a number      of superscattering, why wasn't this possi-
of publicly declared purchases, or "public      bility being pursued in the labs?
bets", of superscattering coupons. The                Wheels turned, but Prof. Puccuts'
Peruvian had also published on the idea, but    technical explanations did not satisfy
in an obscure Peruvian journal. Someone         Senator Toshgib. Federal funding agencies
else created a market on whether there          wanted to avoid a confrontation, but also
would be a compact neutrino telescope,          wanted to protect their turf from every sena-
which became popular with amateur astron-       tor's whim. So they prodded the adminis-
omy clubs. Certain traders even specialized     tration of the University of Toledo, to qui-
in keeping these two markets in rough cor-      etly make a few bets with university over-
respondence. By 1998, the total value of all    head funds. This infusion of capital over-
coupons out on superscattering, the "market     whelmed what Pat and a few others could
capital", reached $8K.                          scrape up. The odds fell to 2%, and the
      That year an aide to Nevada Senator       issue was dropped.
Sue Toshgib, member of the Senate                     But the $100K now in the game raised
Committee on High Tech, noticed the high        the interest of a few experienced speculators
odds for neutrino telescopes, and traced it to  looking for an angle. They hired a few grad
Pat's claim. Sue saw an opportunity to          students to try the superscattering calcula-
                                                               tions, and the odds crept up to
             flow of capital              total capital        6% over the next year. About
  $100K                                     in market          the time the university realized
                                                               there might be no limit to the
                                                               capital required to keep the
   $10K
                                                               odds down, Prof. Puccuts
                                                               published a paper showing
   $1K
             "no superscattering"                              why superscattering was im-
                 coupon values                                 possible. The hired grad stu-
   $100                                                        dents were intimidated, and
                                                               the odds fell to 1%. Six
    1995           2000           2005            2010         months later a student from
                                                               across the hall over heard
   $100                            "superscattering"           Prof. Puccuts mention that an
                                    coupon values              equation in his paper was an
   $1K                                                         ad-hoc approximation. So he
                                                               re-tried the calculation, and
   $10K                                                        got a rather different result.
                                                               He told his old professor,
   $100K                                                       Prof. Yikkul, and they jointly
                                                               wrote a paper. As rumors
              market odds          PAT'S TIMELINE
spread, the market odds shot to 20%.            horse races to suggest betting will work
     In 2001 the first experiment was           there also, but what about everything else?
started. The odds fluctuated under the in-           There are a number of parameters
fluence of some false rumors, and some ex-      which indicate when a claim will be more
perimenter's friends made extra income by       difficult to handle, including the time and
trading before the experimental results were    expense required to resolve a claim, the
revealed. In 2005 the market odds were at       probability it will never be resolved, the
70%, and by 2008 the issue seemed pretty        strength of emotions on the issue, the lop-
much settled, with the market odds at 98%.      sidedness of the odds, and the scarcity of
Pat, who had doggedly stuck with super-         interested traders. The procedures de-
scattering, sold his coupons to reap a total    scribed below are intended to allow idea
profit of $700K on his $20K investment          futures to handle as many claims as possi-
(which was all his spouse would let him         ble. I hope to make the case for wide appli-
risk). This profit came from selling "$1 if     cability plausible enough to inspire interest
superscattering" coupons for 98¢, after         and experiments.
buying them for as little as 1¢, and from the
7% interest the Bank of Vegas had agreed to     PROCEDURES
pay on the money it held. Just before the             In idea futures markets people would
coupons were to be judged in 2013, the last     exchange coupons like those attached to the
traders sold their coupons to avoid paying a    beginning of this paper (Anti-gambling laws
judging fee. The market was closed down         require that these be void, unfortunately.)
and the judges were never needed.               Each coupon is issued by a bank, and spec-
     Pat was still not offered an academic      ifies a judging organization who will decide
position, as Prof. Yikkul became the cele-      the issue "beyond a reasonable doubt".
brated discoverer of superscattering. So Pat    Coupons have the basic form "$X if A",
started a new market, to be judged by a de-     where A is a claim and $X is a "conditional
tailed historical study in thirty years, on     value". A claim includes a sentence, such
who was the first discoverer of superscat-      as those in Figure 1, and any clarifications
tering. Certain universities vowed to let       on word meanings. The claim part of a
professors defend their own ideas. Prof.        coupon also specifies who will judge the
Puccuts, who never bet any of his own           claim, a judging date, and any declarations
money, still has tenure.                        about the decision criteria or process to be
SCOPE                                           used. There can be many coupons on the
                                                same claim, each to be judged by different
     Mechanisms like idea futures have been     judges on different judging dates.
used for a long time to create consensus              Coupons also specify a total judging
about corn prices, stock dividends, life ex-    fee and a maximum percentage judging fee.
pectancy, marine accident risks, horse          The judging fee is obtained by reducing the
races, and football games. So clearly it can    face value of each coupon in the market on
work for some topics. But the vision of-        the judging date by whatever percentage is
fered here is to make much wider use of         necessary given the total market capital. If
such mechanisms. Some areas of science          this would violate the maximum percentage
and technology seem similar enough to           judging fee, then the banks must try to cre-
                                                ate enough market capital by gambling the
existing capital in an "audit lottery". If the          funding to decide a question, judges may be
coupons win the lottery, enough capital is              allowed to postpone judging to a new date
created to support judging, and coupon face             with a new, perhaps larger judging fee.
values are increased. If not, coupons are               They could in the process offer some clari-
worthless and judging is not needed.                    fication of the question, and would use up
     The judging fee creates an incentive to            some portion of the max percentage judging
"settle out of court" by selling before the             fee to pay for their trouble. Markets can
judging date, as happened in Pat's super-               also be set up so that if no decision can be
scattering market. Audit lotteries preserve             agreed on, judges can declare "all bets are
incentives for honest evaluation even when              off".
an issue is of only limited interest and very                 The conditional value part of a coupon,
expensive to judge [Pol], such as whether               the $X, specifies a standard investment in-
your daughter would make a good doctor, if              strument, such as a stock, bond, or mutual
only someone would pay her way through                  fund, and gives specifics like amount, date
medical school. Pat specified that an audit             of purchase, interest rate, etc. There should
lottery be used, if necessary, to fund the              be a liquid market in such instruments, so
historical study on who discovered super-               that it is always clear what the equivalent
scattering.                                             cash value is. In the superscattering exam-
     Judge's verdicts should be "beyond a               ple, bonds issued by the Bank of Vegas
reasonable doubt" and are expressed as per-             paying 7% interest were used. There can be
centages to be paid off to each side. Judges            coupons with different investment instru-
have additional incentives to be careful if             ments for each claim and judge combination
they agree to spend some fraction of their              described above. By building on standard
judging fee to keep the market price of an              investments, an idea futures investor can
"appeals" coupons near that of their verdict.           expect a better rate of return than he could
Appeals coupons are on the same claim, but              get with any standard investment alone.
judged much later by independent judges                       Banks are long-lived financial institu-
with a much larger budget.                              tions trusted to properly report judging fees.
     If there is not yet enough evidence or             A bank's main function is to split and join
Figure 2. Idea Futures Home Version 1) Choose a claim like "I will win this hand of stud poker". 2) Get
a pencil, and some chips. Let red chips be "$10 if claim", blue chips be "$10 if not claim", and something
else be money. 3) Give each person $100. 4) At any time players may buy or sell pairs of red and blue
chips for $10. 5) Place the pencil across the bar below between 5 and 6 on the CLAIM side. This means
the market maker will sell one red chip for $6 or buy it for $5. 6) Whoever yells first, such as "buy red for
6", can trade one chip at the offered price. The pencil then immediately moves in that direction, such as to
6-7. Repeat till market settles. 7) Reveal new information, like the next card, and repeat step 6. 8) If the
claim was right in the end, reds are worth $10, blue $0. If wrong, blue is $10, red $0. Have fun!
                                     7
                                     6
                                     5

                                     3
                                     2
                         9

                                     8




                                     4




                                                                                   1



                                                                                                       CLAIM
CLAIM
 NOT




                                    2
                                    3

                                    5
                                    6
                                    7
                         1




                                    4




                                    8

                                                                                  9
coupons. For example, the claim "True"          trusted to be consistent. Therefore arbitrage
can be split into "A" and "not A". Imagine      activity should keep the total social consen-
giving a bank one share of General Motors       sus roughly self-consistent.
(GM) stock. The bank would see this as a             Other market innovations, such as op-
"1 GM share stock if True" coupon and ex-       tions, baskets, and hedges, allow investors
change it for the coupon pair "1 GM share if    to specialize in details they think they know
A" and "1 GM share if not A". The bank          about and ignore other issues. Options al-
would hold on to the pair and if A wins,        low bets on price volatility, independent of
give the 1 GM share to any holder of a "1       the way the price moves. Baskets allow
GM share if A" coupon. While the bank           one to ignore differences; one can buy a
held the 1 GM share betting stakes, the         basket of all types of coupons on a certain
wealth it represents would, we hope, be put     claim, and ignore differences in judges, in-
to productive use by GM. An "A" coupon          vestment instruments, etc. Hedges allow
can be further split into "A and B" and "A      bets on price differences, such as when in-
and not B". Using certain combinations of       vestors kept the odds on superscattering and
such coupons, one can bet on the condi-         compact neutrino telescopes in rough corre-
tional probability of "B given A" and be in-    spondence. For example, one could correct
sensitive to the verdict on A.                  for the human bias of overconfidence by
     Each type of coupon must have at least     betting that on average the odds are not as
one public market for trading coupons.          extreme as the market odds.
Preferably, such markets will be continuous          While Pat had to risk a substantial por-
bid/ask markets allowing anyone to post or      tion of his wealth on one question, a more
take offers via computer. A single com-         typical scenario would include larger private
puter could implement thousands of low-         research labs whose salaried employees di-
activity markets.                               rect investment in many questions.
     To increase liquidity and reduce price          Idea futures markets could be inte-
fluctuations and spreads, anyone can fund       grated with one or more publishing media
automated market makers [Bl], computer          or "registries". People could make "public
programs always available for trading. A        bets", where they buy a coupon for a claim,
simple market maker algorithm exists which      write a statement of support, and commit to
can function indefinitely and not be cheated    having registries reveal both of these at a
by clever combinations of traders (see Math     pre-specified date. Track records could be
Appendix). The degree of price smoothing        compiled from such information and used
it provides depends on the amount of sacri-     as reputation scores. People with high
ficial cash it starts with. This provides a     scores could become investment advisors,
way to subsidize a market, as does offering     making a public bet with each piece of ad-
to pay part of the judging fee. Pat funded      vice. A advanced publishing medium
such a market maker to promote trading in       [Han88] would allow anyone to post evi-
his market.                                     dence and arguments and link them to the
     If the odds on different claims are in-    disputed claim.
consistent, i.e., do not adhere to the stan-
dard axioms of probability, then arbi-          ADVANTAGES
trageurs can make money by buying or                Idea futures offers many possible ad-
selling "Dutch books" [De]. This profit         vantages. A visible consensus would im-
comes at no risk if the final verdicts can be
mediately form on a wide range of hotly de-      change the market odds runs the risk of the
bated issues. This consensus would be            word getting out and the market ending
relatively universal, expert, honest, self-      right back where it started. Finally, hedge
consistent, and cheap. Such a consistent         bettors can correct for standard biases in
consensus might allow society as a whole to      individual judgment.
approach the level of rational consistency            Individuals would have clear monetary
that is now only expected of individuals.        incentives to be honest and careful in con-
      The market consensus could carry so-       tributing to the market consensus. If the
cial weight, serving as a coordination point     odds you believe are different enough from
for thousands of independent conversa-           the market odds, you believe you will on
tions. In each discussion, the market odds       average make money, even more than with
on an issue could be assumed as the default      a standard investment like a stock index
unless specific arguments were presented to      fund. And compared to stocks, idea future
the contrary. Dissenters could be given the      bets are precise and modular. In stock bets
time-honored challenge to "put up or shut        one must usually bet on a combination of
up". In the same tradition, those willing to     ideas, such as the company's product, mar-
put themselves on the line would be given        keting strategy, production techniques, etc.
due respect and attention. I have observed       In idea futures you can bet on exactly the is-
that the challenge of a bet makes people no-     sue you think you know something about.
ticeably more cautious about what exactly             It might be fun! Imagine a page in the
they are claiming.                               newspaper like the stock page, showing this
      As debates become settled, they would      week's odds on controversial issues.
leave a trail of agreed upon statements.         Imagine coming home from an enlightening
These could be used to counter bogus             discussion to change your investments.
statements, often made by those ignorant of      Imagine reading something you disagreed
solid expert consensus. Visionaries like Pat     with, and stopping for a minute to make a
would have a new way to try to convince          bet against it. The knowledge you created
others of a revolutionary claim; they could      while reading would be directly useful to
throw all available capital into bets. If this   society and yourself, instead of thrown
were enough to change the market odds,           away as is usual now.
they could point to these odds in arguments.          Non-scientists could have a direct,
If not, they would at least expect to make a     even if small, influence and personal stake
healthy profit, and gain social credit for be-   in science to heighten their interest, like the
ing serious. True cranks would end up            amateur astronomy clubs in the story of Pat.
subsidizing leveler heads.                       Amateur trading would induce scientific re-
      The social weight of consensus could       search by traders seeking an edge, subsidize
help damp many presently distorting biases.      professionals who can better predict, and
It would be harder for popular media to          might even fund research by judges.
create consensus by sheer repetition of a        Savings would be encouraged and research
claim; they would have to convince those         would be directed more at issues of general
willing to bet. A sincere public relations       interest. Capital and hence intellectual effort
campaign could make a public bet, but an         would flow to markets where there is broad
insincere one would know they were               interest, strong disagreement, and relevant
throwing money away. And an insincere            data obtainable for a modest effort or a short
attempt to throw enough money away to            wait.
     Idea futures markets create informa-        specific concerns of Extropy readers in fu-
tion, combining what individuals know. A         ture issues.
market with more capital will probably have
better information, as people will see there     RELATED WORK
is more to win by figuring out the answer.             In Bayesian decision theory, an agent's
By subsidizing a market you can pay to           degree of belief in A is often defined to be
create information, though you won't get         the price they would be willing to pay for a
exclusive access. This might be a better         "$1 if A" coupon [DeF]. Idea futures just
way for government to fund scientific re-        applies this definition to a society as a
search, instead of the usual grant-giving ap-    whole to find our consensus degree of be-
proach [Han90]. In fact, governments             lief. In the presence of a market, agents ap-
might use odds from subsidized markets as        pear to agree [Kad].
factual input for government decisions. We             As was mentioned before, markets
could all have our say about whether pro-        similar to idea futures in commodities, fi-
jected usage would justify a new mass            nance, insurance, and sports betting.
transit system, or whether a death was sus-      Science and technology bets are frequently
picious enough to justify an autopsy.            made between individuals [Hal,Ti,WSJ], as
Schools might even admit students based on       they have been through history [Deb]. The
the market odds of candidates getting a high     idea of betting on a wide range of legislative
G.P.A. if they attend.                           and technological issues is raised in a recent
     Idea futures can also provide insur-        science fiction novel [Bru,Ve], and scat-
ance. A risky business venture based on a        tered proposals [Fa,Ho,Lea,Ze] have been
new technology might bet against that tech-      made to formalize bets in science.
nology to reduce total risk. Idea futures can          Business schools widely use such
be a foundation for reputation systems,          markets to teach M.B.A. students about
providing another way to encourage experts       markets [Fo]. In economic theory, the
to give honest advice, and allow other ex-       coupons I have been describing are called
perts to disagree. Idea futures offers all       "contingent assets", and are often used as a
these benefits without requiring any coer-       foundation in analyzing financial invest-
cion or taxation. Unlike patents, it requires    ments [ShW] and the effect of uncertainty
no international enforcement or litigation       [La]. Ideally, there would be a "complete
about the origins of an idea.                    market", with assets contingent on every
CRITICISMS                                       possible state of the world. In reality mar-
                                                 kets are not complete, and various sorts of
     By now you probably have in mind at         "market failure" are traced to this fact.
least one objection to idea futures, and will          Incompleteness is usually [Hir] ex-
not be entirely comfortable with it until this   plained as due to judging difficulties, finite
objection is addressed. Longer papers on         transaction costs, and market thinness. In
this subject [Han90] consist largely of de-      fact, these authors are often unaware that
tailed responses to such objections. Space       such markets are almost universally prohib-
limitations preclude such detail here, so fig-   ited by anti-gambling laws, as joint-stock
ure 3 just gives a list of some issues ad-       companies, life insurance, and commodity
dressed in those papers. The editor of           futures [Ros] were prohibited before special
Extropy willing, I could respond to the          interests managed to obtain exemptions.
Though unevenly enforced, such laws pro-              up markets, letting prices fluctuate, and
hibit public science bets between strangers           playing market-maker. Because of this,
in all of the U.S. and in most of the world.          they are usually unwilling to offer bets on
     Only Great Britain, to my knowledge,             claims where they do not know how to es-
allows such bets, and then only for the last          timate the odds, and few bookies have ad-
three decades. English bookmakers per-                vanced science educations. As a result,
ceive little demand for science bets, and so          they mainly take safe bets, siding with the
take them mainly to induce popular articles           scientific establishment against "crazy" out-
mentioning the going odds on unusual sub-             side theories, which doesn't help the image
jects [ShG]. This publicity brings in new             problem betting has in many quarters. One
clients, who may then switch to the "real"            cannot even subscribe to a publication list-
betting on sports. Because of this, bookies           ing the going prices on science questions.
prefer small bets on subjects "in good taste"         It should be possible to improve on this.
that anyone can understand, like UFOs,
Yetis, and Moon landings. They avoid                  AN APPEAL
subjects that seem too esoteric for the gen-               Idea futures is mainly just a curiosity to
eral public, like the recent "cold fusion"            most people, even those convinced of its
claims, and subjects that won't very clearly          feasibility and desirability, . It would re-
resolve themselves, as a judging industry             quire substantial effort to implement, and in
has not yet evolved.                                  some sense is an trivially obvious idea,
     Bookmakers traditionally prefer to set           given the appropriate theoretical back-
prices and stick to them, rather than setting
Figure 3 A Few Concerns about Idea Futures
Isn't Gambling Illegal? Isn't Betting a Useless Zero-Sum Game? Does Anyone Ever Bet This Way? What
About Compulsive Gambling? Is There Enough Interest in Science Questions? Will These Markets Be
Too Thin? Doesn't Betting Only Work For Clear Cut Questions Like Horse Races? How Often Do Beliefs
Really Converge? What If Beliefs Never Converge? What Do Convergent Beliefs Have To Do With
Truth? What About Badly Worded Claims? Can't Wrong Ideas Still Be Useful? What If The Fine Print
Differs From The Summary? What About Sucker Bets? Don't Science Questions Resolve Too Slowly?
Why Should I Trust The Judges? Won't Judging Cost Too Much? Won't Wealthy People Have Too Much
Influence? Won't The Market Be Dominated By Fools? Won't Advertising Manipulate Opinion? Aren't
Markets Full Of Cheats And Thieves? What About Insider Trading? What About "Moral Hazard"? What
About Incentives To Start False Rumors? What About Incentives To Keep Information Secret? Won't An
Apparent Consensus Create A Crowd Mentality? Will The New Incentives Slow Or Stop Convergence?
Won't Different Claim Wordings, Judges, And Base Assets Confuse The Consensus? Won't The
Consensus Reflect Risk Preferences As Well As Beliefs? Won't Betting Challenges Discourage Creativity?
What's The Point Of A "Consensus" That People Disagree With? Isn't It Better For People To Argue Out
Their Own Disputes? Won't This Have The Same Problems As Patents? Wouldn't Anonymous Trading
Screw Up Reputation Statistics? If This Is So Great, Why Hasn't It Happened Already? Won't greed sully
the pure pursuit of ideas? Does a few dollars of compensation in the end really help a rejected visionary?
Doesn't this presume there is some absolute truth? Won't convergence be culturally relative? Isn't
consistency unhealthy? Doesn't organized crime take over anything having to do with gambling? What
about libel and national security? What about "Nuclear war will destroy 90% of the world by 2020"?
Won't different claim wordings, judges, and investment instruments fragment the market? Why should
verdicts be consistent with one another? Won't judges be reluctant to contradict the market? What if the
probabilities get very small? Why not do without judges?
ground. I think the only people who might
actually be willing to work to make it hap-     MATH APPENDIX
pen are people who are particularly unhappy     VARIABLES:
with current methods of forming and com-        P(A) = Market probability of A
municating scientific consensus, and how        J(A) = Judge's verdict probability of A
those methods have treated issues dear to       $X = An investment with a current market value of
them. People perhaps like Extropy readers,         X dollars.
sympathetic to markets and subjects, like       C = Total value of distributed coupons on a claim
cryonics and uploading, which current           F = Total budget available for judging
consensus institutions deal poorly with. I      f = Maximum percentage judging fee
fear it will require more effort than I alone   IDENTITIES: $X = "$X if True"
                                                "($X if A) if B" = "$X if (A and B)"
can muster to make it real. It may well be      EXCHANGES: (These remain valid if change $X
that if you don't do it, no one will; what do      to "$X if A", or multiply all $ by a constant)
you say?                                        Split/Join: $1  "$1 if B", "$1 if not B"
     There are many options for pursuing        Trading on A: $P(A)  "$1 if A"
idea futures. I have worked to gain the at-     On A given B: $P(A given B)  "$P if not B",
tention of "science policy" academics              "$1 if B and A"
[Han92], and idea futures will soon be a        Cash in with Judges: "$1 if A" -> $J(A) "$1 if not
known, if oddball, suggested alternative           A" -> $(1 - J(A))
mechanism for science funding. I have           DUTCH BOOK EXAMPLE: If P(A) + P(not A) <
mostly developed a board game and to a             1, then can buy "$1 if A" and "$1 if not A" for
lesser extent an email reputation game.            less than $1, sell the pair to the bank for $1, and
                                                   make a profit.
Legal research is probably the most impor-      JUDGING FEES:
tant task, but it is on hold for a lack of      1) If f*C >= F, pay $F to judges, reduce coupon
funds.                                             values $X -> $X*(1-(F/C))
                                                2) If f*C = 0, no judging happens
CONCLUSION                                      3) If 0 < f*C < F, Take C and play a lottery: With
      I have argued that futures markets in        probability C*f/F, increase value of coupons $X
ideas could help the evolution of ideas by         -> $X*F/(C*f) and do 1) Otherwise $X -> $0
                                                   and do 2)
creating a visible consensus of relevant ex-    MARKET MAKER ALGORITHM: (See Figure 2)
perts, and better incentives for honesty and    Choose a function M(i) from integers to [0,1] such
care when making contributions. Idea fu-           that M(i) > M(i + 1), M(0) = 1/2.
tures might offer these and many other ben-     Choose a transaction quantity Q.
efits cheaply and without coercion. Though      Market starts at j = 0.
some problems remain, it seems worth fur-       Offer "$Q if A" -> $(Q*M(j)) and if taken j -> j+1
ther study.                                        Offer "$Q if A"  j-1
Can you think of a question where 1) you        If M(i) = 1/(1 + exp(i/k)), total loss