March 17 2008 / by Alvis
Category: Social Media Year: 2008 Month: Mar Rating: 7
Prediction market expert Chris Masse over at Midas Oracle has observed
that although Google has perhaps the largest in-house prediction
market going, open exchanges such as Reality Markets get much
higher engagement from their users.
Masse reports that Google’s internal prediction exchange
consists of about 1500 traders
who made 80,000 trades over the last two-and-a-half years,
whereas 
public prediction exchange Reality Markets boasts 500
more-involved users who made 180,000 trades in the last
year-and-a-half. This averages out to approximately 24 yearly
predictions per user for Google, and 240 yearly predictions per
user for Reality Markets.
What does this 10-to-1 ratio in favor of public exchanges
reveal? For one, that Google employees are very busy people.
Second, that people who really want to participate in prediction
markets will be way more active. Third, that there’s a body of
people who can contribute to corporate prediction markets that are
structurally disenfranchised.
If we take into account a core lesson about social dynamics
taught by James
Surowiecki in his oft-quoted Wisdom of the Crowds,
namely that a diverse set of people (who understand enough about
the problem being considered) will tend to make better predictions
as a whole than a homogeneous group, then one would assume that
Google (and other big boys) should also look to tap out-of-house
predictors to boost the accuracy of their internal markets. While
the prospect of sharing internal information with the broader
public is not appealing, isn’t it possible to find middle
ground?
What if a company like Google were able to convince a smaller
amount of these more active predictors, say 50 or so, to come in
and help forecast their more sensitive corporate future? (This is
totally do-able w/ the cash and reputation Google possesses.) Might
not the overall results turn out significantly better with a group
of more-independent predictors contributing new perspectives and
context at a much higher rate? If so, it sure would be worth the
investment.
There’s got to be some way to bridge the gap between private and
public prediction markets, thereby making the most of both
worlds.
Comment Thread ()
