Comments on: Categorizing prediction markets

By: How To Buy Stocks

How To Buy Stocks — Sat, 17 Jan 2009 19:52:00 +0000

Thanks for this.

By: tiberiu84

tiberiu84 — Mon, 15 Dec 2008 09:27:36 +0000

You also try offering promotional mugs. Many people seem to find that as a nice gesture.

By: kevindoylejones

kevindoylejones — Sun, 07 Sep 2008 03:38:58 +0000

interesting effort

By: jedc_mercury

jedc_mercury — Wed, 03 Sep 2008 22:33:50 +0000

Thanks for that, Alex. I think I remember that, but do you have a link?

That's some good food for thought. I hope to post an updated version soon.

By: Alex

Alex — Wed, 03 Sep 2008 20:05:56 +0000

Hey Jed,

When I did my heuristic reviews of prediction markets almost two years ago, I used the following categorizations:

Market Purpose: (e.g. fun, gambling, create industry specific predictions, etc.)
Market Control: (i.e. who creates the markets: users, market makers, the platform providers, etc.)
Targeted End User: (e.g. gambler, farmer, general user, doctor, etc.)
Currency: (real or play)
Underlying Event Types: (e.g. events, stock market, etc)
Contract Types: (e.g. binary, etc.)
Trading Mechanism:
Size of Market: (i.e. volume)
Number of Contracts: (or, rather, number of markets)

By: jedc_mercury

jedc_mercury — Tue, 26 Aug 2008 20:28:42 +0000

I'm curious; I completely forgot to address the method of trading in the criteria above. I think this is probably a fairly critical distinction in prediction marketplaces. Sites that use a standard Continuous Double Auction (CDA) certainly operate differently than those using Robin Hanson's MSR or David Pennock's DPM. Thoughts?

By: jedc_mercury

jedc_mercury — Tue, 26 Aug 2008 20:26:16 +0000

Hi, Jose. Thanks for commenting.

I do think that Currency and Rewards should be kept seperate. But perhaps there should be a more granular scale for Rewards, from:
Intangible Rewards -> "Gift"-type Rewards -> Cash Rewards

Though I still believe that marketplaces should be classified by the maximum reward a trader can earn, not every person is really even trying to be in competition for that top slot.

On Background Data, I know HSX has some algorithms that they use to generate their forecasts for the studios that essentially "correct" the publically traded figures. But no market uses any demographic data to change the publically-traded figures. However, if you're a good trader you should have earned more capital, and thus have more influence on the price through shear volume of shares/contracts purchased.

By: josejose50

josejose50 — Tue, 26 Aug 2008 12:35:00 +0000

On currency or rewards, they should be listed seperately, because the behavior of the players in the market will vary and differ based on whether there are tangible rewards or not (this is mainly for play money, I cant think of a situation with real money currency for intangible rewards). It makes sense that the people in Currency – Play Money, Rewards – Tangible are going to be the least risk averse, since they have nothing to lose and everything to gain, with the opposite being Currency – Real Money, Rewards – Real Money, which should lead to the most risk-averse behavior. Not sure which leads to “better” prediction markets, since differing levels of credibility could be applied based on the Currency/Rewards setup.

One thing about Background Data, I'm curious to see if any prediction market uses the background/demographic data to “weigh” trades/predictions, to help either control for insider knowledge or to promote insider knowledge. For example, if I'm trading in stock for Iron Man 2 on HSX, it would be helpful to know industry insiders are trading on the stock (similar to the SEC filings that are required by Wall Street).

By the way, like the ICROP acronym, easy to remember!

By: jedc_mercury

jedc_mercury — Mon, 25 Aug 2008 15:59:58 +0000

Hi, Nigel.

1. Completely agree. I know I participate on HSX and other sites without thinking about prizes, though I'm not sure how widespread that is. To be honest, I was thinking here of Predictify. They offer small cash rewards and play-money winnings. But when I go on Predictify, all I do is trade on the markets where there is a potential cash reward, and ignore the rest. I think this is an important dimension, perhaps there are more "steps" than I indicated here.

2. That's what I was thought based on what I've seen on Hubdub so far. Again, you hit on the criterion where there's certainly some grey areas. Perhaps the first two examples I mentioned are actually the same thing? (They are both different than the "background data" intention, in my opinion.)

And I just added you to the blog roll. Sorry; it's been a while since I updated it!

By: Nigel Eccles

Nigel Eccles — Mon, 25 Aug 2008 15:21:41 +0000

Interesting categorization. A couple of points:
1. Prediction markets with tangible rewards rely on both tangible and intangible rewards. The vast majority of people who play games with prizes generally aren't doing list the prizes as a important motivation.
2. On objective I would say we are actually aiming for all three however initially we are focusing on community.

BTW Any chance of a link on the blog roll?