Prediction Markets, Collective Intelligence, Innovation and Growth
Prediction markets produce dynamic, objective probabilistic predictions on the outcomes of future events by aggregating information that traders bring when they agree on prices during their transactions.
12 thoughts on “Prediction markets produce dynamic, objective probabilistic predictions on the outcomes of future events by aggregating information that traders bring when they agree on prices during their transactions.”
It is strangely very difficult to come up with a definition of “prediction market” that covers even the existing varieties. Your definition would exclude HSX that site doesn’t produce event probabilities. It could also exclude LMSR trading depending on one’s interpretation of traders “agreeing on prices” (certainly, LMSR traders don’t agree to each other’s prices quite as much as CDA traders do). It would exclude Betfair unless you consider their bettors a “traders”.
When we discussed the PMIA’s purpose in London, one idea was to create a “Prediction Market” label that we could use to differentiate our products and our approach from the growing number of “wisdom of crowds” barbarians (eg, blubet, guessnow, etc.) who are shamelessly exploiting the “prediction market” brand even while their offerings have little resemblance to such a thing. Unfortunately, coming with a proper definition seems to be a real challenge.
The reason I publish this definition… is precisely… to get some comments… as Mike Giberson was kind enough to provide, last time. (I probably not fully integrated his wisdom, naughty boy that I am.)
– BetFair traders are included in my definition. BetFair odds are easily translated into percentages ( 1 / digital odds ).
– HSX event derivatives are included, provided that an interpretation of their odds is given (by them).
– MSR is included… if you call their actions “trading”… I do… but for that point, it’s not me who decides, but the others… What do they think about MSR? Is that trading, for them? [Since Robin Hanson defines MSR as a mix between CDA and Scoring Rules, then I do call MSR “trading”. Not sure, though.]
If you want to have a discussion, then publish your proposal on a popular forum (MO or else), and let’s see what the others say…
No, HSX box-office predictions are forecasts, not event probabilities, so if you restrict the definition to event probabilities, you exclude most of HSX. Same for “vote-share” markets at IEM.
If you insist on the term “traders”, then you exclude Betfair, because most people on Betfair would not think of themselves as “trading” anything. They’re just betting.
– A trader who trades only one time is like a bettor, indeed. But he/she is a trader, still. (BetFair traders can trade a lot, if they wish.)
– BetFair is a prediction exchange (”betting exchange” as they say in Britain) just like NewsFutures, Betdaq, InTrade-TradeSports, MatchBook, and the late HedgeStreet.
– You’re a better expert than me as for HSX. The prices of their stocks may be interpreted and translated into probabilistic predictions —am I correct?
HSX’s index-based markets are only probabilistic in interpretation, e.g. if you find that historical errors for similar markets are such and such and then establish a confidence band around the current price, in a sense you have a probabilistic prediction, but there is nothing intrinsically probabilistic in a single price quote. (A study on that sort of interpretation would be interesting.)
As Emile has pointed out, many to most questions that corporations want answered are better framed as index markets rather than binaries.
“Aren’t most firms using binaries for their internal prediction markets?”
It depends what you’re trying to predict. If you’re trying to predict an event probability, a binary contract will do just fine. But if you’re trying to forecast a variable, like a sales volume (or a box office result a la HSX), then using a set of binaries quickly become unwieldy. The so-called “Arrow-Debreu design” where you associate a binary contract with each of several potential ranges for the target variable is what H-P used in their classic printer sales experiment. But for traders that means trading several contracts for a single forecast, as if trading one contract wasn’t complex enough already. If you’re trying to forecast a half dozen variables and each of them is represented by, say, five contracts, then traders have to consider 30 contracts! It quickly gets impractical… Unless you really care about producing probability distributions for your variables, it’s just not worth it. H-P abandoned that approach long ago.
The bottom line is that if you want to use a prediction market to make forecast, you’re better off using a single contract whose payoff is a continuous function of the target variable’s outcome, like an HSX movie-stock, or an IEM vote-share contract. Alternatively, you can also choose a design that does away with trading entirely, like H-P’s BRAIN or NewsFutures’ Comptetitive Forecasting.
Not sure if Niall’s comment was in response to me but on second thought it would be more interesting to see how accuracy relates to price volatility while the market is open.
ESS: “certainly, LMSR traders don’t agree to each other’s prices quite as much as CDA traders do.”
In a manner of speaking, I guess you could say that when LMSR traders agree on prices they don’t trade, and when they disagree, they do trade. But this doesn’t differ too much, in terms of the market point of view, from CDA.
While CDA traders explicitly agree on a price to trade at, in some sense the agreement on a price to trade reflects an underlying disagreement. The buyer thinks the price is lower than the event probability and the seller thinks the price is higher than the event probability. Once everyone agrees, trading stops.
So, too, you could say that when CDA traders agree they don’t trade, and when they disagree, they do trade.
[…] Our good doctor Emile Servan-Schreiber of NewsFutures (some time ago): […] we could use to differentiate our products and our approach from the growing number of “wisdom of crowds” barbarians (eg, blubet, guessnow, etc.) who are shamelessly exploiting the “prediction market” brand even while their offerings have little resemblance to such a thing. […] […]
From bombastic to verbose.
It is strangely very difficult to come up with a definition of “prediction market” that covers even the existing varieties. Your definition would exclude HSX that site doesn’t produce event probabilities. It could also exclude LMSR trading depending on one’s interpretation of traders “agreeing on prices” (certainly, LMSR traders don’t agree to each other’s prices quite as much as CDA traders do). It would exclude Betfair unless you consider their bettors a “traders”.
When we discussed the PMIA’s purpose in London, one idea was to create a “Prediction Market” label that we could use to differentiate our products and our approach from the growing number of “wisdom of crowds” barbarians (eg, blubet, guessnow, etc.) who are shamelessly exploiting the “prediction market” brand even while their offerings have little resemblance to such a thing. Unfortunately, coming with a proper definition seems to be a real challenge.
Shall we have this discussion here?
The reason I publish this definition… is precisely… to get some comments… as Mike Giberson was kind enough to provide, last time. (I probably not fully integrated his wisdom, naughty boy that I am.)
http://www.midasoracle.org/200…..-exchange/
– BetFair traders are included in my definition. BetFair odds are easily translated into percentages ( 1 / digital odds ).
– HSX event derivatives are included, provided that an interpretation of their odds is given (by them).
– MSR is included… if you call their actions “trading”… I do… but for that point, it’s not me who decides, but the others… What do they think about MSR? Is that trading, for them? [Since Robin Hanson defines MSR as a mix between CDA and Scoring Rules, then I do call MSR “trading”. Not sure, though.]
If you want to have a discussion, then publish your proposal on a popular forum (MO or else), and let’s see what the others say…
Robin Hanson’s previous attempt:
http://www.midasoracle.org/200…..finitions/
No, HSX box-office predictions are forecasts, not event probabilities, so if you restrict the definition to event probabilities, you exclude most of HSX. Same for “vote-share” markets at IEM.
If you insist on the term “traders”, then you exclude Betfair, because most people on Betfair would not think of themselves as “trading” anything. They’re just betting.
– A trader who trades only one time is like a bettor, indeed. But he/she is a trader, still. (BetFair traders can trade a lot, if they wish.)
– BetFair is a prediction exchange (”betting exchange” as they say in Britain) just like NewsFutures, Betdaq, InTrade-TradeSports, MatchBook, and the late HedgeStreet.
– You’re a better expert than me as for HSX. The prices of their stocks may be interpreted and translated into probabilistic predictions —am I correct?
HSX’s index-based markets are only probabilistic in interpretation, e.g. if you find that historical errors for similar markets are such and such and then establish a confidence band around the current price, in a sense you have a probabilistic prediction, but there is nothing intrinsically probabilistic in a single price quote. (A study on that sort of interpretation would be interesting.)
As Emile has pointed out, many to most questions that corporations want answered are better framed as index markets rather than binaries.
Thanks for your HSX remark.
Aren’t most firms using binaries for their internal prediction markets??
“We find similar accuracy in another play-money market called the Hollywood Stock Exchange
(http://www.hsx.com/). Prices of securities in Oscar, Emmy, and Grammy awards correlate well with
actual award outcome frequencies, and prices of movie stocks accurately predict real box office results.”
From Pennock et al ; “The Real Power of Artificial Markets.”
“Aren’t most firms using binaries for their internal prediction markets?”
It depends what you’re trying to predict. If you’re trying to predict an event probability, a binary contract will do just fine. But if you’re trying to forecast a variable, like a sales volume (or a box office result a la HSX), then using a set of binaries quickly become unwieldy. The so-called “Arrow-Debreu design” where you associate a binary contract with each of several potential ranges for the target variable is what H-P used in their classic printer sales experiment. But for traders that means trading several contracts for a single forecast, as if trading one contract wasn’t complex enough already. If you’re trying to forecast a half dozen variables and each of them is represented by, say, five contracts, then traders have to consider 30 contracts! It quickly gets impractical… Unless you really care about producing probability distributions for your variables, it’s just not worth it. H-P abandoned that approach long ago.
The bottom line is that if you want to use a prediction market to make forecast, you’re better off using a single contract whose payoff is a continuous function of the target variable’s outcome, like an HSX movie-stock, or an IEM vote-share contract. Alternatively, you can also choose a design that does away with trading entirely, like H-P’s BRAIN or NewsFutures’ Comptetitive Forecasting.
Not sure if Niall’s comment was in response to me but on second thought it would be more interesting to see how accuracy relates to price volatility while the market is open.
ESS: “certainly, LMSR traders don’t agree to each other’s prices quite as much as CDA traders do.”
In a manner of speaking, I guess you could say that when LMSR traders agree on prices they don’t trade, and when they disagree, they do trade. But this doesn’t differ too much, in terms of the market point of view, from CDA.
While CDA traders explicitly agree on a price to trade at, in some sense the agreement on a price to trade reflects an underlying disagreement. The buyer thinks the price is lower than the event probability and the seller thinks the price is higher than the event probability. Once everyone agrees, trading stops.
So, too, you could say that when CDA traders agree they don’t trade, and when they disagree, they do trade.
[…] Our good doctor Emile Servan-Schreiber of NewsFutures (some time ago): […] we could use to differentiate our products and our approach from the growing number of “wisdom of crowds” barbarians (eg, blubet, guessnow, etc.) who are shamelessly exploiting the “prediction market” brand even while their offerings have little resemblance to such a thing. […] […]