Emile Servan-Schreiber (of NewsFutures):

From: Emile Servan-Schreiber
Date: Mon, 15 Jun 2009 22:45:43 +0200
Subject: Re: GE Imagination Market License

That’s pretty funny, considering that the creator of the GE Imagination Market herself concludes in several published papers that “using markets to rank ideas may be no better than other methods of idea ranking,” (1) and that “a survey of market participants yielded mixed results regarding the market’s effectiveness at ranking ideas.” (2) Elsewhere, on [General Electric]‘s own blog, one of those participants details at length the conceptual and practical incoherence of such self-referential “preference markets”. (3)

(1) http://portal.acm.org/citation.cfm?id=1265501
(2) http://www.ingentaconnect.com/content/ubpl/jpm/2009/00000003/00000001/art00004
(3) http://www.grcblog.com/?p=378

Previously: General Electric’s internal betting exchange: The Imagination Market

“We use them ["them" = the enterprise prediction markets] as another point in the decision-making process, alongside asking experts and other business leaders,” said Christina LaComb, a computer scientist in the R&D lab at GE.

OK. You’re using enterprise prediction markets; you’ve gotten your name in the newspapers; you’re “cool”.

And our good friend David Perry‘s gotten your money.

But do your event derivative traders have the same access as you do to “experts” and other “business leaders”?

Or do you leave them in the dark? In that case, your enterprise prediction markets would be clueless, useless, and worthless.

Indeed, it appears to me that [prediction markets] are growing not from corporate or government use, but mostly organically from within academia, stock-futures circles and political-junkie communities. I’m reading the interesting variety of writers and prediction-marketeers at Midas Oracle, which brings together widely ranging posts from faculty members at Harvard and other universities, daytraders, and even a few “amateurs.”

-

Lewis Sheperd notes in his post that a number of for-profit companies (like Google and General Electric) are using private prediction markets (a.k.a. enterprise prediction markets). Non-for-profit organizations (like governmental agencies) would do great, too, using the same forecasting tool —an “information aggregation mechanism” (IAM), more exactly.

Robin Hanson, instead of boring us with philosophy, go evangelizing that newbie.

-

UPDATE: Yes, he is willing to learn. See his comment.

-

Brought to you by Koleman Strumpf (circa November 2007):

-

Henry Berg, Microsoft <slides>
Discussant: Robin Hanson (George Mason Department of Economics) <slides>
-
Christina Ann LaComb, GE (The Imagination Market; abstract is free, text is gated) <slides>
Discussant: Marco Ottaviani (Kellogg School of Management, Management and Strategy) <slides>
-
Dawn Keller, Best Buy (Best Buy’s TAGTRADE Market) <slides>
-
Bo Cowgill, Google (Putting Crowd Wisdom to Work) <slides>
-
Jim Lavoie, Co-Founder and CEO, Rite-Solutions <slides>
-
David Perry, Co-Founder and President, Consensus Point <slides>
-
Mat Fogarty, Founder and CEO, Xpree Inc <slides>

-

Tom W. Bell, Chapman University School of Law <slides>

-

1. David Perry (Consensus Point)

Perry shared many lessons learned, and chose to focus people’s attention on the pitfalls of PMs (in addition to their power).

1. Implementation can take a long time. It’s a change management process.
2. Asking the wrong questions can be a killer, but often the right questions emerge after trading begins.
3. Poor incentive structures is the #1 killer, according to Perry.
4. Don’t suppress the reputational incentives associated with PMs.
5. Be explicit and transparent about how traders will be compensated.

2. Dawn Keller (Best Buy)

Keller began her remarks by explaining how Best Buy management wants to become better at listening. There is an authentic passion for tapping into the best information available among the vast network of Best Buy personnel.

PM offer a powerful tool in this regard. PM fits with the Best Buy mantra of: “Have fun while being the best.”

Keller framed her remarks around how management has tried to educate the Best Buy workforce about PM and the evolution of PM implementation from a committed band of volunteers to senior management buy-in. Best Buy has opted for a “shot gun” approach rather than a “rifle shot” approach. Best Buy tends to run many parallel markets at once to see what sticks rather than trying a few targeted pilots. Consistent with many other speakers, Best Buy runs many non-business related markets to invoke interest and fun, noting that participation is key and must be continually cultivated.

3. Christina LaComb (General Electric)

LaComb set the stage for her comments by noting GE managements desire to implement reliable intelligence aggregation processes. PM offers one such technology.

GE management, to date, has had little interest in the forecasting applications of PM. Rather, they [want] methods for grappling with new technologies and other factors associated with disruptive change. So, in the GE setting PM takes on the nature of “preference markets” rather than “prediction markets.”

LaComb went on to describe lessons learned in the implementation of GE preference markets. Much of this information is available in a published case study.

LaComb stressed that: “Validation is a huge challenge.” To date, the validation most relevant in the GE environment has been “gut checks” from senior management.

Can prediction markets “grapple” with disruptive change, really? I view this forecasting tool as conformist, whereas inventions/innovations require a dose of anti-conformism.

Dr. John Nafeh – Director

John Nafeh is the Founder of HedgeStreet. He combined his expertise in decision analysis and risk management, his experience guiding Internet-related start-ups, and his vision of an Internet-based mass market for risk-hedging financial instruments [*] to create HedgeStreet. As Managing Director of Pareto Partners, a venture capital fund, Dr. Nafeh has demonstrated extensive expertise in start-up financing, strategic planning, marketing, and mergers and acquisitions. He was involved with Atari and Apple Computer during their start-up phases, founded PC and database software companies in the mid-1980s, and earlier held management positions with Ford Aerospace and General Electric. Along with a BS in electrical engineering, Dr. Nafeh holds an MS in management science and engineering and a PhD in decision and risk analysis from Stanford University.

[*] a flawed vision?

Information markets [= prediction markets] are typically used as prediction tools, aggregating opinions about the likelihood of future events, or as preference indicators, identifying participants’ product preferences. However, the basic information market concept is more widely applicable. In our experiment, we utilized information markets [= prediction markets] within the domains of idea generation and group decisioning. Participants were allowed to propose ideas regarding potential technology research areas; these ideas were represented as securities on a virtual financial market. Participants were able to trade shares of technology ideas over the course of 3 weeks, resulting in the market identifying the “best” idea as the highest priced security. Our findings suggest that information markets [= prediction markets] for idea generation result in more ideas and more participants than traditional idea generation techniques; however, using markets to rank ideas may be no better than other methods of idea ranking. Additional benefits include providing immediate feedback, allowing visibility of all ideas to all contributors, and being a fun mechanism for consensus building.

UPDATE: Is General Electric Imagination Market a joke?

Consensus Point:

- General Electric

- Best Buy

- Motorola

- Qualcomm

- Siemens

- Nokia

NewsFutures:

- Arcelor – Mittal

- Corning

- Dentsu

- IHG

- Lilly

- MasterFoods

- Pfizer

- Siemens

- Thomson

- Yahoo!

Inkling Markets:

- Abbott

- Chrysler

HSX:

- Chiron – Novartis

+ Google + MicroSoft + etc.

Would prediction markets help? Nokia U.S.A. is a client of Consensus Point.

Previous: OFFICIAL: Consensus Point clients are Best Buy, General Electric, Nokia, and Samsung. + Email Interview: Ken Kittlitz

“Jack [Welch, ex-CEO of General Electric] was essentially saying a graduate business degree was a waste of time.”

That’s a misunderstanding of the comment. What Welch was saying was that the most valuable thing the students were paying for was the contacts they made, not the content of the coursework. The contacts are hard to make any other way, and many people consider the investment worthwhile. Even if you manage to meet the same people without having the shared background, it won’t do you as much good, since you won’t get the same reaction from them.

Chris Hibbert is one of the world’s best prediction market designers (Yahoo! research scientist David Pennock outsources to him the most complex problems) and the software architect of Zocalo.

Chris Hibbert has published a bunch of very good technical explainers of the many prediction market designs:

• PM Intro: Basic Formats – [simple double auctions] – by Chris Hibbert – 2005-12-30
• PMs with Open-Ended Prices – [markets with open-ended prices] – by Chris Hibbert – 2006-01-05
• Looking at Both Sides – [the symmetry of complementary purchases] – by Chris Hibbert – 2006-04-17
• Market Design: Book and Market Maker – [how to integrate an order book with an automated market maker] – by Chris Hibbert – 2006-04-28
• Increasing Liquidity in Multi-Outcome Claims – [the mechanics of multi-outcome markets] – by Chris Hibbert – 2006-07-19
• Continuous Outcomes: Bands, Ladders, and Scaled Claims – [predicting the value of a continuous variable] – by Chris Hibbert – 2006-09-20
• Integrating Book Orders and Market Makers – (mirror on MO) – by Chris Hibbert – 2006-09-20
• Conditional and Combinatorial Betting – (mirror on MO) – by Chris Hibbert – 2007-03-06
• Market Makers for Multi-Outcome Markets – (mirror on MO) – by Chris Hibbert – 2007-09-10