Prediction markets at Google

According to the report, “Using Prediction Markets to Track Information Flows: Evidence From Google,”
which was presented Friday at the American Economic Association meeting
in New Orleans, the strongest correlation in betting was found among
people who sat very close to one another, trumping even friendship or
other close social ties.

This is tangible evidence, the authors
argue, that information is shared most easily and effectively among
office neighbors, even at an Internet company where instant messaging
and e-mail are generally preferred to face-to-face discussion.

It is an argument, the authors say, for giving greater importance to
“microgeography,” or how people interact in the workplace. The finding
that information moved fastest among people who were the closest
together is also an endorsement of the company’s “third rule for
managing knowledge workers: Pack Them In,” the authors say.

And Adam Smith is validated once again:

The other crucial finding of the report was that there was a
detectible “optimism bias” among Google employees. That is, results
that were good for the company tended to be overpriced, particularly
for “subjects under the control of Google employees, such as, would a
project be completed on time or would a particular office be opened.”

This
optimism was most evident among new employees, the report found, and it
was bound to show up on days when Google stock had climbed.

Here is the story.  Of course this is very important work.  Thanks to Chris Masse for the pointer.  Elsewhere in prediction markets land, InTrade has now started conditional prediction markets, which consider oil prices, interest rates, and U.S. troops in Iraq, conditional on who becomes President.

Comments

This seems to be to be relevant to Ed Glaeser's recent paper with Giacomo A.M. Ponsetto about how the decreasing costs of transportation has decreased the comparative advantages of manufacturing cities but increased that of information and idea based cities, which Arnold Kling links to here.

Just because information travels faster, that doesn't necessarily mean it's correct.

As one who appreciates contrarianism, I see a lot of flaws in going to an extreme based on conclusions of this.

The speed of information flow might be good when it comes to facts, but dangerous when it comes to opinions.

Once prediction markets become more mature, APIs will be available for programs to correct some of these errors if they are indeed pervasive. i.e., "If Google stock rises X%, sell Y% of all contracts".

Sadly, Intrade doesn't look like it plans to mature too much. I hope Inkling will start offering real-money markets.

This is important work for organizational sociology, but not for prediction markets, as this does little to help us find and field high value markets.

G, if you're implying that Intrade doesn't have an API, you're wrong.
I have code at http://www.bayesianinvestor.com/amm/intrade_amm.py which uses their API to run an automated market maker that provides liquidity for the conditional prediction markets that Tyler mentioned.
Intrade apparently hasn't put their API documentation on the web, but I suspect they'll send it to anyone who asks.

Peter,

Don't you have to pay a significant fee to use their API? The last time I looked at it I believe that was the case.

It is enlightening!

If u love somebody, send him work for Google, if u hate somebody, send him work for Google.

Comments for this post are closed