Biology Reference
In-Depth Information
the number of traders increases the accuracy of the predictions. While the
minimum number of traders needed for an accurate prediction is still an
open research question, the empirical data from prediction markets designed
to forecast elections suggests that the number of participants need not be as
large as the number required to obtain comparable accuracy with surveys.
For example, election prediction markets with 200 active traders routinely
yield predictions with smaller errors than those from opinion polls involving
ten times as many respondents. And predictions markets designed to predict
sales for a company have performed well with as few as a dozen traders.
In general, prediction markets with at least 20 or 30 traders can yield good
results, but accuracy tends to improve as the number of traders increases.
Finally, there must be incentives to trade; transactions made for the pur-
pose of maximizing profits are the means by which traders reveal their pri-
vate information to the marketplace. However, the use of money involves
two problems. First, current gambling laws and financial market regulations
prohibit the operation of prediction markets in which traders put their own
money at risk. And, second, use of money to motivate traders might have
negative connotations in some health care settings.
A “funny money” market solves both of those problems; it was the solu-
tion adopted for the novel influenza A (H1N1) prediction markets described
below. Specifically, trading takes place in an artificial currency, some amount
of which is endowed to each trader at the beginning of the market. After the
market closes, the balances remaining are worthless, but during the course
of the market, the account balances serve as a measure of the trading and
prediction success of individual traders. Those balances can, for example, be
posted, either anonymously or with the traders identified, and the standing
on this “leader board” might itself provide motivation for active and care-
ful market participation. The efficacy of funny-money markets has not yet
been completely resolved. Some experimental evidence, however, shows that
accurate results can be achieved, though the variance around predictions
seems to be greater than with markets in which traders earn real profits and
losses (Servan et al. 2004).
A second solution is to operate the markets with real money, but the money
used is provided by the market managers rather than by the participants.
As with funny-money markets, traders are given a trading account endowed
with some amount of an artificial currency, but at the end of the market, the
balance in the account—the original endowment plus trading profits and less
trading losses—is converted to U.S. dollars and paid to the trader. Traders thus
realize real monetary rewards in exchange for their participation, much like
respondents are sometimes paid for their completion of a survey instrument.
The difference, however, is that the size of the reward depends on the suc-
cess of the participant, both in accurate assessment of the likelihood of event
outcomes and in successful market actions. To minimize any adverse reac-
tions, the payments at the end of the market can be in the form of educational
grants to traders for use in paying for reference topics, journal subscriptions,
Search WWH ::




Custom Search