Does average Joe and Jane understand prediction markets?

[Sveinung Arnesen]

No. Not from my experience. I think some of the reason why prediction
markets so haven´t been as successful as expected inside companies is
that the average person does not fully get how it works.

I conducted an experiment with prediction markets within political
parties, where I recruited as traders active party members who were
taking part in the election campaigning. I made sure I had had person-
to-person contact with everyone before the market opened, explaining
how it worked and what the purpose of the experiment was. My
impression was that most of the participants did not quite understand
how prediction markets worked before I talked to them, and some did
still not understand it after I had talked with them.

Prediction market organizers need to increase the effort to make
markets understandable, and there is still some room for improvement
on the visual design and user interface. Close follow-up with
participants is also preferable, if that is practically possible.

Those are my views, at least. Anyone else with thoughts around this
subject?

[David Pennock]

I favor using simpler language that people do understand:

http://blog.oddhead.com/2009/03/17/kiss-prediction-market-lingo-goodbye/

Using automated market makers helps too. Having two prices (bid and ask)
for every prediction can be confusing.

I'm biased but I believe predictalot was about the simplest UI. But even
that is not simple enough. Still more work to be done in this direction.
Many of the vendors get this including crowdcast, nf, and inkling.

thanks,
Dave

[Sveinung Arnesen]

Indeed. I agree very much with the blog post. In principle I also agree that using betting language would make it more understandable for many. However, when talking with skeptics and other people new to prediction markets, I have a hard time convincing them that prediction markets are different from entertainment betting. Using the betting language does not help in this matter.

I think we should move more towards asking "do you agree with the prediction?" and if not follow up with "how strongly do you disagree?" After having indicated his/her opinion, the participant can be shown the trading consequences of this opinion. The slider shown in the blog post is a good idea. We used something similar in our experiment, together with a real time update (Ajax) of what the new prediction would be, as well as how much it would cost the trader to move the slider to the point he had indicated.

-----Opprinnelig melding-----
Fra: prediction-market...@googlegroups.com på vegne av David Pennock
Sendt: ti 28.09.2010 17:17
Til: prediction-market...@googlegroups.com
Emne: Re: [PM Industry-18] Does average Joe and Jane understand prediction markets?

http://blog.oddhead.com/2009/03/17/kiss-prediction-market-lingo-goodbye/

thanks,
Dave

--
You received this message because you are subscribed to the Google
Group: "Prediction Markets Industry".
To post to this group, send email to
prediction-market...@googlegroups.com
To unsubscribe from this group, send email to
prediction-markets-open...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/prediction-markets-open-discussion?hl=en

[Anthony Towns]

On Wed, Sep 29, 2010 at 18:55, Sveinung Arnesen
<Sveinung...@isp.uib.no> wrote:
> I think we should move more towards asking "do you agree with the prediction?" and if
> not follow up with "how strongly do you disagree?"

"How strongly" is a really hard question to relate to there, I think.
You can assign values (like "strongly" means ">75% chance" or "willing
to lose 1000 credits if I'm wrong"), but personally, I don't think it
ends up quite making the sort of intuitive sense that would be nice.

Betting has the advantage that the consequences are clear -- you
either win some money for being right, or you lose some money for
being wrong, and at any given time there's a fairly small range of
choices for that in a market.

Betting definitely has the problem of having negative associations
though -- it's just for entertainment, it's an addiction, it hurts
players' families, it's a zero-sum game, etc. And really how familiar
are betting slips to people, anyway? Enough that they "work", I guess,
and more than things like "short selling" maybe, but...

I've been pondering "insurance" terminology recently -- a lot of
people are familiar with it (car insurance, home and contents, and
life insurance), and while there's still negative associations
(insurance companies ripping you off and not paying out, insurance
fraud), buying insurance in case your office burns down doesn't have
anywhere near the same connotations as "you're gambling on your office
getting torched???" does, eg.

For things that "matter" (what will the share price of X be, who will
win an election, will terrorists blow up the world), getting insurance
against an outcome probably makes reasonable sense and the idea of
saying "$5000 worth of insurance cover in case of a Party B victory
will cost you $3000" seems fairly feasible.

Although, while "$1M worth of cover in case of a terrorist attack on
Metropolis for $20,000" seems plausible, I'm not sure "$1M worth of
cover in case there isn't a terrorist attack on Gotham" would sound as
good as "I bet $980,000 there won't be a terrorist attack in Gotham."

(As far as Lex Luthor or the Joker buying anti-crime insurance goes,
the usual insurance fraud principles come into play -- if you turn out
to have been responsible for the thing you insured against, you don't
get the cash)

I don't know if that would work for company-internal prediction
markets -- "uh, I'd like to buy insurance just in case this product
doesn't meet management's projections" might not go down terribly
well. Though if management was confident enough to bet their annual
bonuses in the other direction, maybe it would...

Cheers,
aj

--
Anthony Towns <a...@erisian.com.au>

[David Pennock]

In principle I would say you should give up and admit that prediction
markets *are* the same as entertainment betting.

But I understand that in practice that may not fly.

So you might use slightly different language, like invest, stake, or
risk instead of "bet". "I stake 10 points that Obama will win" or "I
risk 10 points..."

I agree with Anthony that asking qualitative questions will be hard to
translate into quantitative results. I also agree that insurance
language is more palatable and works well in certain circumstances when
one side of the event is something naturally "bad" that one would like
to avoid. Interesting idea.

thanks,
Dave

-------- Original Message --------
Subject: Re: [PM Industry-19] Does average Joe and Jane understand
prediction markets?

Date: Wed, 29 Sep 2010 22:03:35 +1000
From: Anthony Towns <a...@erisian.com.au>
To: Sveinung Arnesen <Sveinung...@isp.uib.no>
CC: David Pennock <d...@nnock.com>,
prediction-market...@googlegroups.com

On Wed, Sep 29, 2010 at 18:55, Sveinung Arnesen
<Sveinung...@isp.uib.no> wrote:

> I think we should move more towards asking "do you agree with the
prediction?" and if
> not follow up with "how strongly do you disagree?"

"How strongly" is a really hard question to relate to there, I think.

Cheers,
aj

--
Anthony Towns <a...@erisian.com.au>

[Sveinung Arnesen]

Good points. However, if a prediction market is nothing else than entertainment betting, then there is no use for the term "prediction market". There has to be a difference, and I remember that there previously has been a discussion on the definition of prediction market. Without a clear result, if I remember correctly. But in general one could say that it is betting on matters of public utility. Agree?

In some situations I imagine that recruiting participants with only personal economic incentives seems a bit odd. Consider an influenza market, for example, with a set up where only doctors, nurses, microbiologists, etc participate. I don´t think the well paid doctors who take part do it to earn €25. They do it to improve an early warning system, and for them the whole betting setting is a bit tacky. The same would go for a terrorist market, I presume.

In my opinion, we need to focus more on the public utility part in order to legitimate prediction markets, and talk less about "putting the money where the mouth is".

Cheers,
Sveinung

-----Opprinnelig melding-----
Fra: prediction-market...@googlegroups.com på vegne av David Pennock

Sendt: on 29.09.2010 21:14
Til: prediction-market...@googlegroups.com
Emne: Re: SV: [PM Industry-21] Does average Joe and Jane understand prediction markets?

thanks,
Dave

Cheers,
aj

--
Anthony Towns <a...@erisian.com.au>

--

[David Pennock]

Fair enough: I agree there is a difference in goals. In gambling and
finance, trading *is* the goal. With prediction markets, information is
the goal and trading is simply a means to the end. I also agree that
there shouldn't be religion about pure monetary incentives. If
information can be obtained without it, then that should not be ruled
out. Indeed evidence shows that many wisdom of crowds approaches work
surprisingly well, from play money games to polling and averaging.

http://messymatters.com/2010/01/14/prediction-without-markets/

Still, though other methods often come close, real-money markets still
perform best when accuracy and credibility are paramount, and are the
only institution I know of that actually consistently provide real-time,
public, quantitative predictions.

thanks,
Dave