Subscribe to this feed
Harvard Business School Case Study: IdeaWarz and Prediction Markets
Last week MJ, Jasmine, and myself were invited by Professors Andrew McAfee, Karim Lakhani, and Peter Coles of the Harvard Business School, to do a case study on the merits of using prediction markets for IdeaWarz.
Prediction markets are something we have considered in the past, though for a variety of reasons were never completely comfortable with.
Here is what Harvard’s best and brightest had to say on the subject.
What are prediction markets?
A prediction market is a vehicle for placing bets on the outcome of certain events. For example, people use prediction markets to bet on who is going to win the next Super Bowl.
While prediction markets like Tradesports.com do brisk business betting on sporting events, they have also recently become hot in business circles. They’re being used by companies like HP and Google internally to determine things like product release dates and how many Gmail subscribers are likely to sign-up in a given month.
The case for prediction markets
At beginning the discussion, Professor McAfee polled the class and asked who thought prediction markets made sense for Cambrian House’s IdeaWarz.
Here is how the students voted:
| A. Loved prediction markets | (14) |
| B. Liked prediction markets | (36) |
| C. Disliked (favored current voting system) | (19) |
| D. Hated prediction markets | (3) |
Just the day before, Bo Cowgill of Google, gave a good presentation on how they were using prediction markets internally to help determine new product launch dates. It was no surprise that after hearing Google’s success with prediction markets internally, the students thought IdeaWarz would benefit from them as well.
Most students liked the fact that prediction markets caused people to make their best predictions. If they guessed right, they would be rewarded. Wrong and they would be left with nothing. That feeling of really having some skin in the game, much like the stock market, meant prediction markets should lead to better results than simple voting.
Others liked the increased granularity prediction markets provided. Instead of half the ideas being eliminated each round of the tournament, good ideas could hang around until the end.
Overall, students thought prediction markets would give an accurate depiction of which ideas were best in the community.
The case against prediction markets
Shifting gears, Professor McAfee asked those students who weren’t fans to explain why.
Top of the list was increased complexity. Prediction markets are more complicated than simple American Idol style voting. People need to be knowledgeable on terms like: bid, ask, quantity, and volume.
Another difficultly would be in defining the trigger points for the market itself. What would be the event that decides whether the contract becomes true or not? We could ask community members to buy contracts based on ideas that return the highest profit - but at the end of the month, with out any confirming data, this is just a guess. It is neither true nor false.
Another student commented that in the absence of an independent event, the market would become self-fulfilling. People buying the contracts on the market would themselves decide the outcome. For a prediction market to work, the event deciding the outcome needs to be independent of the voting. Otherwise all you have is a crowd rewarding itself for acting as a herd. It was this last comment that I found most insightful, and the biggest knock against using prediction markets for IdeaWarz.
Near the end of the discussion, Professor McAfee polled the class one last time to see if their thoughts had changed:
| A. Loved prediction markets | (5) |
| B. Liked prediction markets | (15) |
| C. Disliked (favored current voting system) | (29) |
| D. Hated prediction markets | (15) |
Not surprisingly, many who previously liked prediction markets were now not so sure.
Critical thinking
What I found most valuable about this experience was having an independent group apply critical thinking to the questions and issues we faced. MJ and I had talked a lot about prediction markets over the year, but were not able to fully articulate our unease until these students provided us with striking insight in areas we missed before. It was an honor and a privilege to have access such a collection brains.
We would sincerely like to thank the Harvard Business School and their faculty, Professor McAffe, and Lakhani for hosting us for the week.
This is one crowd we would love to tap again.
April 9th, 2007 at 10:35 pm
JR- first congrats on getting to tap that audience- very cool. and don’t take this the wrong way but I would think Ideawarz becomes significantly less important with the new decentralized style of CH. Whatever the method for tapping WOC for insight (prediction markets or voting) the emphasis for crowdsourcing should shift more towards the marketplace concept for finding development muscle for projects within the community. Maybe use voting/prediction markets for deciding which projects to reward with additional funding but this game is 90% about execution at this point and it’s mildly disconcerting to see focus put towards WOC before the marketplace is working…
again, not trying to naysay here - more just looking for confirmation that you guys agree on what the focus is. or if not, perhaps share the 6,12,24-month vision a little better so I can understand the roadmap.
thanks!
Sean
April 10th, 2007 at 2:11 am
[…] HOW COME I AM THE LAST TO KNOW THIS??? Cambrian House: Just the day before, Bo Cowgill of Google, gave a good presentation on how they were using prediction markets internally to help determine new product launch dates. […]
April 10th, 2007 at 2:30 am
[…] Are prediction markets too complex for the Joe Six-Packs? Cambrian House: People need to be knowledgeable on terms like: bid, ask, quantity, and volume. […]
April 10th, 2007 at 7:26 am
Hey scrollin - no worries. We are still completely focused on finding the development, marketing, copywriter, graphic design muscle necessary to turn peoples ideas into reality.
You are right - it is all about execution and we haven’t lost sight of that.
The Harvard gig was in the works for a long time and we first were in contact with them late last year when we still were focused on your previously business model.
Still, it was very interesting to hear some other peoples thoughts on the merits of a prediction market. I know that was something many of us in the community debated for some time). Harvard really helped identify the main challenges we would face if we went down that route.
Thx for the comment.
April 25th, 2007 at 8:42 am
Hi JR and MJ,
Is there a video (or voice) recording of the Harvard session? I would love to watch or listen to the full session to follow the full discussions.
Can you please email me to let me know. Much thanks.
Kempton
April 25th, 2007 at 9:43 am
Hi Kempton,
We don’t have any video or audio material was can share on the sessions unfortunately. This is HBS property. I am not sure, but I don’t see why HBS wouldn’t resell case study material on their web site. Seems like that would be something people would be interested in.
Cheers - JR
May 3rd, 2007 at 2:12 pm
Hi! I enjoyed this article. I started http://www.memes.org and am beginning to understand your argument against PMs — they are too complex and are easily replaced by Polls and Votes and the data can be as good and nuanced.
May 28th, 2007 at 7:21 am
[…] Chris Abraham: May 3rd, 2007 at 2:12 pm Hi! I enjoyed this article. I started http://www.memes.org and am beginning to understand your argument against PMs — they are too complex and are easily replaced by Polls and Votes and the data can be as good and nuanced. […]
June 22nd, 2007 at 1:57 am
You said: “People need to be knowledgeable on terms like: bid, ask, quantity, and volume.”
I built some decent prediction markets at National Gazette (http://nationalgazette.org/). I agreed that “Chicago Mercantile Exchange-like” terminology and trading features would be too complex for the general public. So, we built our trades to always be buying or selling quantities of “yes” or “no” contracts. Bids, asks, and volumes disappear.
Stripped down to their bare essentials, I think prediction markets are just as easy to understand as polls, but more fun for everyone involved because you can build a warchest (or lose your shirt… it feels like sitting at the poker table sometimes), all while excercising your knowledge about the problem domain and contributing to a collective, crowdsourced probability on a future event. Very cool stuff.
July 11th, 2007 at 10:39 am
[…] “Pour qu’un marché prévisionnel fonctionne” ajoute la Harvard Business School , “l’évènement dont il est question doit être indépendant du vote. Faute de quoi on a une foule se récompensant quand elle agit comme un troupeau.” […]
October 7th, 2007 at 3:58 pm
LasikPlus Vision Centers
LasikPlus Vision Centers
March 9th, 2008 at 12:15 am
$anchor$basketball Betting,final Four,final Four Betting,final Four Gambling,final Four Sports Book,final Four Sportsbook,march Madness,march Madness Betting,march Madness Gambling,march Madness Sports Book,march Madness Sportsbook,ncaa,ncaa Betting,…
$anchor$basketball Betting,final Four,final Four Betting,final Four Gambling,final Four Sports Book,final Four Sportsbook,march Madness,march Madness Betting,march Madness Gambling,march Madness Sports Book,march Madness Sportsbook,ncaa,ncaa Betting,nc…
November 16th, 2011 at 9:38 pm
Listen to Pro, good tips to help you get results.
January 19th, 2012 at 3:50 am
Many thanks this terrific post, Now i am glad I observed this website on yahoo.