NewsFutures has (always had) an interesting feature in its user interface that I want to highlight. I want to explain the difference so I can rely on it in my explanation of other market formats. Whenever they create a new (binary) claim, they write a description of each outcome phrased in positive terms. When they offer a claim on the San Jose Sharks playing the Los Angeles Kings (tonight at 7:30), the two positions are “Kings over Sharks” and “Sharks over Kings”. Whenever you are looking at the market screen that allows you to buy claims, you only see one side of the market (“Sharks over Kings”, for concreteness). If you don’t have a position in the claim, you can only buy the claim. If you have positive assets (from this viewpoint), then you are also offered the option of selling some. But there’s always a button that lets you look at the claim from the opposite point of view. Once you click the button, you see things from the viewpoint of “Kings over Sharks”. If you don’t have any assets, or all your assets are in “Kings over Sharks”, You can still only buy.

When you switch viewpoints, the prices in the order book invert. if the prices were X$65 offered and X$93 asked when looking at “Sharks over Kings”, then it will be X$7 offered, X$35 asked when looking at “Kings over Sharks”. I really like this way of presenting the prices, but it shouldn’t apply only to binary outcomes. It’s useful to be able to think of the prices inverting in this way when you talk about multi-outcome claims as well.

The most common way to present a single binary outcome is to phrase the question in a positive sense, and have a single scale for quoting odds. This means that people betting against the proposition have to subtract from 100 to figure out how much they’re spending or earning.

NewsFutures’ interface simplifies this interaction for the user. There’s a secondary benefit, in that it’s particularly helpful to the seller, and people are more reluctant to bet against outcomes than in favor of them.

Tom Bell’s paper on how to interpret, skirt or redefine the laws to ensure that real money prediction markets are clearly legal to operate covers a lot of ground, and is a substantial addition to his previous paper on the subject. This paper, though I can’t comment on whether the legal arguments will hold up, makes a clearer and more explicit case that prediction markets fall outside the jurisdiction of most US law. The paper covers

  • What a prediction market could do,
  • how prediction markets compare to other markets,
  • what the legal obstacles are,
  • why legality in the US matters,
  • arguments that prediction markets aren’t covered by gambling law, the CFTC, or the SEC, and
  • an explanation of his draft of an act to explicitly make Prediction Exchanges legal.

In order to draw a clearer line around the kinds of markets he’s interested in, Tom presents his proposal in the context of a critique of the effectiveness of patents and copyrights in promoting Science and the Useful Arts. This gives him a platform from which to argue in favor of claims on open questions in science, technology and public policy, and he focuses on markets that deal exclusively in those subjects as worthy of extra attention and especially deserving of protection from onerous regulations. (Those are the Prediction Exchanges the draft act would legalize.) In choosing this narrow focus, he omits one of the goals originally identified for prediction markets: PMs can also enable hedging exposure to risk in areas covered by relevant claims. But that requires substantial liquidity, which may be forever out of reach on these kinds of claims, so perhaps it’s better to try to take advantage of the kind of legality he promotes than to look for a more encompassing definition that might be harder to defend.

Markets v. Benefits
Entertain Hedge
Prediction Market primary secondary tertiary tertiary? N.A.
Futures Market secondary tertiary N.A. primary N.A.
Securities Market secondary tertiary? N.A. tertiary primary
Gambling Market N.A. N.A. primary N.A. N.A.

In order to clarify how prediction markets differ from other organized markets, Tom presents the table on the right. He points out that the most important goals and achievements of prediction markets are at best secondary goals for the others, and their primary goals are relatively unimportant for prediction markets. A second table shows that prediction markets work differently, but doesn’t have as strong a conclusion. Each of the features of PMs is either present in some of the other kinds of markets, or at least is already a visible variation

Market Type v. Market Feature
Risk of Loss
Greater Than
Prediction Market yes yes usually no no
Futures Market yes no yes usually yes
Securities Market yes usually no usually sometimes
Gambling Market no yes yes no sometimes

among them.

As Bell says, The gambling and financial laws that menace prediction exchanges do so almost accidentally. They serve several different purposes, and each is written fairly narrowly, as the various authors weren’t thinking beyond prohibiting or regulating particular activities that they thought they could clearly distinguish. That’s as it should be: laws should be tailored to the circumstances, and should leave non-infringing behavior to proceed in peace. Unfortunately, innovations sometime tread near old paths, and the old laws don’t clearly say whether they cover ideas which hadn’t been conceived when they were written. As Bell continues, The policies behind those restriction do not fit prediction markets at all. This leaves a gap which entrepreneurs (or more specifically, their backers) are wary of entering, for fear that all their work will go for nothing. Underwriters prefer the uncertainties of their ventures to be in areas they have more chance of affecting.

The heart of the paper is the argument that existing laws and oversight agencies don’t apply to Prediction Markets. IANAL, so I can’t comment on whether the arguments would hold up in court, but they are clearly a good start on a brief that should at least give a D.A. reason to pause before attacking a (narrowly science-based) prediction markets as gambling. There aren’t as many restraints on the CFTC or the SEC, so it’s more plausible that they would search for alternative interpretations that would allow them to regulate. But the plain law says that the CFTC is limited to regulating contracts for future delivery of commodities. The only future consideration traded in prediction claims is the underlying currency, and the asset itself (which will pay out depending on some future conditions) is transferred at the time of the transaction.

There’s also an explicit exception for instruments which are “predominantly a security”. This seems to fit prediction markets very closely. I would add a single caveat: one of the requirements of the definition of “predominantly a security”, is that the instrument not be subject to “mark-to-market” margining requirements. As I understand that term, it’s not much of a stretch to apply it to the way that TradeSports handles short sales. As I explained in my introduction to Prediction Markets, TradeSports reserves a varying amount of your short holdings against the possibility of losing the bet. They have rules about frozen funds, margin requirements, and explicitly says they may make Margin Calls. I said before that the short selling model doesn’t have any different effect on trading than the other ways of betting against a particular claim. This seems to be a difference; if there are regulations that cover trading in which there might be margin calls, then the short selling model is affected and the others are not.

Escaping the jurisdiction of the SEC seems to come down to a point highlighted in the first table: After all the technicalities are covered, the SEC regulates instruments that are designed to allow people to raise capital in order to form an enterprise. There are steps you can take to make it less likely that prediction market activities will seep over the line and appear too similar to securities. They need not be onerous: a prominent disclaimer about SEC oversight ought to suffice.

In case all the legal analysis isn’t enough to get business started, Tom proposes legislation that could be passed either by the federal government or (to lesser effect) by individual states in order to create a protected status for somewhat limited prediction markets. The limitations would be enough to keep the more lucrative markets from opening up shop within the US, but could be sufficiently broad to allow a variety of markets in science, technology and social policy. Given the number of new markets opening up recently, that may be enough to encourage some experimentation in real money claims. Given those same new startups, we may not have to wait long to find out whether the exemptions Professor Bell has identified are sufficient on their own.

Google Alerted me to a press
from new startup Inkling, which aims to start up another
marketplace for predictions. They are just getting started, but
intend to offer a variety of claims, and allow the users to suggest
new questions. So far, they have sports, entertainment, and
technology (Apple rumors and web traffic comparisons).

They have a couple of different approaches to generating predictions:

  • the market closes on a preannounced judging date (usually the day
    before an expected event.) The correct outcome pays out at $100.

  • Rolling close: In their Apple Rumors and Big 10 Basketball coach exit
    markets, they list a number of choices, and whenever one comes true,
    they cash it out at $100.

In the first case, the bid prices shouldn’t exceed $100, and you
should sell the basket whenever they do. If the offered prices are
below $100, you can safely buy a basket. (Since the markets are
separate, the price swing needs to be significant so you aren’t caught
by price movements due to others trading while you make multiple
transactions. Even though the site seems slow (due to heavy interest,
presumably), I didn’t have trouble selling baskets at advantageous

They are using an automated market maker. In order to simplify the
user experience, the interface asks whether you think the current
probability (and price) for an issue is too high or too low, then
picking a number of shares to trade from a menu showing what price
you’ll pay.

The bid prices for the markets I looked at were significantly above
$100 bid in exclusive markets. When you sell, the interface implies
that you are selling short (and talks about how much you will get from
the sale), but after the transaction, your balance is lower. I assume
they’re reserving the value of the negative coupons. The software
doesn’t seem to notice that I bought a basket, and so have an assured

I can’t figure out their accounting. The starting account is 5000
inkles. As you see in the image, I’ve earned almost 4500 inkles
selling assets short (having started from 5000), and they seem to be
holding 4500 against that value, so I still have almost 5000 inkles to

I couldn’t login with Firefox. I had to fall back on Safari. Their
FAQ says firefox is supported.

Chris Masse posted another
thorough review
of the state of the
Prediction Markets field in the guise of an annual awards list.
Zocalo and my name came up a few times, sometimes positively,
sometimes not.

His summary from his rss feed gives the highlights. There are plenty
of pointers to interesting reports, articles, blogs, etc. I won’t repeat all the links here. That’s what Chris (doesn’t) get paid for.

  1. Blog Of The Year 2005: Jason Ruspini
  2. Book Of The Year 2005: James Surowiecki, The Wisdom of Crowds – Why the Many Are Smarter Than the Few and How Collective Wisdom Shapes Business, Economies, Societies and Nations
  3. Consultant Of The Year 2005: NewsFutures
  4. Exchange Of The Year 2005: BetFair
  5. Market Of The Year 2005: The Harriet Miers confirmation prediction market @ InTrade/TradeSports
  6. Media Of The Year 2005: Fortune magazine, with this story, Making a
    Market in (Almost) Anything.

  7. News Of The Year 2005: The CFTC spanked TEN (TradeSports / InTrade), hard.
  8. Opinion Of The Year 2005: Good News: CFTC Busts Intrade. – by Tom Bell – 2005-10-10
  9. Paper Of The Year 2005: Justin Wolfers and Eric Zitzewitz’s 2005 Interpreting Prediction Market Prices as Probabilities
  10. Presentation Of The Year 2005: John Ledyard Information Markets
  11. Resource Of The Year 2005: / tag / predictionmarkets
  12. Scholar Of The Year 2005: David Pennock
  13. Software Of The Year 2005: Consensus Point’s Idea Futures

I’m looking forward to more growth of the field, and lots of
interesting work this year.

Donald Luskin
noticed a stock market anomaly
, and when he looked for
explanations, he found a possibility on TradeSports. On January 20th,
the stock market had its biggest one-day drop in three years, but
recovered completely over the next few days. Luskin wondered what had
caused the short term spike, and didn’t find any reasonable
explanations in the financial press. When he looked on TradeSports,
he found that the markets on whether the Republicans will keep their
majority in the House of Represenatives this fall also had a one-day

Luskin suggests that the increased chances of a change in control of
the congress, whatever the cause of that, was probably enough to spook
the market. Since the GOP House contracts recovered to their former
levels within a few days, he concludes that that scare receded, and
that would explain the market’s short term slump and recovery.

Unfortunately for his thesis, Luskin left something out. TradeSports
has separate markets for the Republican’s control of the House and the
Senate, and the two markets moved strongly in opposite directions.
Now it may be that you can explain this away by noting that a change
from 69% to 63% in the chances for the Republicans to retain the House
is much more significant than a change from 79% to 89% in their
chances for retaining control of the Senate, since the former says
there is doubt that they will retain control of both, while the latter
says it’s still likely they’ll continue to control the Senate.

While Luskin only handwaved about the relationship between the stock
market and politicians’ expected behavior, a paper by

Wolfers and Zitzewitz
found significant correlations between the
financial markets and TradeSports’ markets on the chances for war and
for Saddam’s capture. Seems like it shouldn’t be too hard to do a
similar study now that there are explicit markets on control of the

Koleman Strumpf presented a very interesting paper at the Prediction Market Summit last Friday, so I finally read his article on Manipulating Political Stock Market. It bears on the experiment I am working on with some folks at GMU, since both deal with Manipulating markets. Koleman’s paper looks at three examples, two studies of naturally occurring manipulation attempts, and one controlled study of manipulation on the Iowa Electronic Markets. The most interesting is his study of what actually happened in the Presidential election markets that ran in New York and other major American cities between 1880 and 1944.

The newspapers reported daily prices on election wagers, and also often printed names and amounts bet. From these reports, we can tell that many prominent people in the stock market, business, and politics were visibly betting on the outcomes of political races. The betting volume was quite high: on an inflation adjusted basis, $158 Million in 2000 dollars were wagered on the 1916 race.

Koleman based his study on reports in the papers of charges of manipulation of the exchanges, and compared the prices before and after the manipulations were purported to take place. The study shows that while there may be a short term (1-day) effect from the alleged manipulations, the effects seem to wash out within 2 days.

Koleman goes on to study manipulation episodes on TradeSports, and a controlled study of manipulation on IEM. In both these cases as well, the effects of manipulation are transitory. The IEM manipulations were substantial but not sustained. It’s conceivable that in a market that doesn’t restrict deposits (IEM has a maximum of $500) a partisan could spend a more substantial sum over a longer period and sway the odds. But these studies don’t make it look likely to work.

We shouldn’t be surprised that there was betting on elections 120 years ago. Pope Gregory XIV banned betting on papal elections in 1591. He wouldn’t have bothered if betting hadn’t been rampant. Anyone know of any records of wide-scale election betting before that?

HedgeStreet issued a
press release
last week touting a new interface
and improved platform. The UI is somewhat simpler, but I can’t tell
that it is a big improvement. They mention two changes:

  • Rather than paired yes and no contracts on each question, they now
    have a single instrument for any claim that you can take either
    position on. They describe it as buying or “selling short”, but I
    think their terminology is confusing.
  • Their interface offers better charting, streamlined order entry, and
    centralised information about the holdings and account history.

Their description of betting against a position as “short selling”
seems confusing to me. After reading all their help screens, it
appears that when you sell short, you are investing money and
acquiring an asset that may pay out a positive amount, just like when
you buy. HedgeStreet wants to present it as simple by saying there is
only one asset, and you can buy it or sell it short. The problem with
their description is that it isn’t the same as short selling. Their
model is actually more similar to the “buying complementary assets”
model I described in my introduction to basic Prediction Market
formats. Brokers in the stock market can’t sequester the entire
amount that might be required to repay a short position, since the
underlying asset can grow without bound (see Google for an example.)
HedgeStreet’s contracts pay out at either $0 or $10, so HedgeStreet
can (and does) collect the entire price up front. Short sellers pay
$10 less than the stated price for an asset with a positive payout.
Buyers pay the stated price, and also get an asset with a positive
payout. The short selling terminology will be confusing to people who
understand short selling, and will be a handicap to anyone else who
tries to understand short selling later.

They could patch up this description by saying that rather than
having a margin requirement, they simply collect the maximum loss up
front, and repay it when the question is settled, but at that point,
they’d be better off describing it as buying the opposite position
rather than explaining how they stretched the analogy to short

HedgeStreet’s new interface is certainly simpler than their previous
version, which had paired markets in each outcome. When they divided
possible crude oil inventories into 4 bands (x < A, A < x < B, B < x <
C, and C < x), they had 8 distinct markets for one outcome. With
their new software, they will have only four markets, which is an
improvement, but as I said in my talk at the Prediction Market Summit
in San Francisco, and will repeat in New York, you can do better by
linking the markets together. I haven’t yet written up the
explanation for that point; it will appear on this blog when I do.

It also appears that HedgeStreet is now saying that all trades will be
made directly between their customers. I haven’t found their old
documentation (the Wayback
is slow-to-nonresponsive this afternoon) so I can’t verify
what they used to say, but they seemed to have automated market
makers in some cases with the old platform. I don’t see how it’s an
advantage to either HedgeStreet or their customers to drop that
possibility. In illiquid markets, an automated market maker can, at
limited cost, increasing trading possiblities substantially.

We released a

new version
of Zocalo in December. (There have already been about 50 downloads, even without an announcement.) It is available on SourceForge. It includes
some new configurable features for prediction market experiments: the
main one is that experiments can be run in which traders keep their
earnings across rounds within a session. (This allows experiments in
which there is progressive revelation of information over the session
about the underlying value.) In order to better support longer-lived
sessions, we fixed a bug that allowed data to overrun the right edge
of the trade history chart.

In parallel, we have started working on support for long-lived
markets. This is the main step needed to support internal markets
within a business or organization. In order to handle that, we’ll
need persistent data for trader’s accounts and book orders, as well as
an ability to display the price history to traders without requiring
that they be continuously connected to the market (this is currently
necessary for the experiments that Zocalo supports). I’m pleased to
say that I have integrated both
(open source Object-Relational Mapping software) and
JFreeChart (open source
Chart drawing) into Zocalo. Neither is called to the full extent
necessary for a usable long-term market, but both are being used
thoroughly enough to show us that the integration works and they are
being invoked appropriately. The next step is to ensure that all
aspects of users, accounts, and orders are persistent. I’m hopeful
that I’ll be able to demonstrate this functionality by the end of the
month. (This functionality is not yet in the released versions.
We’ll publish it as soon as its usable and stable.)

Smarket is an interesting new market that lets you bet on whether the rank of books and other products on Amazon will go up or down. This is another addition to a type of market I don’t think there’s a good theory for yet; markets with open-ended pricing. There are several other examples including the Hollywood Stock Exchange (HSX), ProTrade, and Yahoo!’s Tech Buzz Game. These markets all try to follow the real-world rise and fall of some abstract value (box office returns, on-field athletic performance, and search rank, respectively), and allow traders to bet on them.

So far, all the markets of this type use only play money, and that’s part of what I’m talking about when I say there’s no theory. One of the innovations that makes standard Prediction Markets (of the kind I described last week) workable is that a risk-neutral market operator can sponsor a market without taking a position in the trading; that is, without taking on risk. The traders buy and sell from one another, and the operator merely facilitates the trades. Even if you add an automated market maker, you can constrain the potential costs, while still facilitating unlimited trading.

Traditional bookies learned to handicap the events they were selling bets on or they went out of business quickly. Part of the development of insurance was inventing the actuarial techniques that allowed insurers to reliably estimate how their exposure would change as they sold various policies. Modern stock markets only allow short selling when the investors back their trading with assets that the broker can rely on if prices go the wrong way. Real money prediction market operators will need the same ability to constrain losses.

With the open-ended markets, the value to be paid out to stocks that gained isn’t limited, and isn’t related in a simple way to the price paid by the purchasers. I would expect a theory that showed how the market’s exposure is limited to start from the observation that one asset can only rise in value when another falls. This is true of Smarkets and TechBuzz because they’re based on relative rankings. It’s not obvious whether it’s true of HSX (Box Offices returns are only weakly limited) or ProTrade (the formulas for ranking athletes haven’t been published.)

It’s also not clear how to read the market results as predictions. In a standard Prediction Market, the traders’ incentive is to bid the price up or down until it matches the subjective probability, so prices should equal probabilities. David Pennock says the optimum target on TechBuzz is the square root of the search share. I think a separate analysis will have to be done of each institution, and if the rule isn’t easy to understand, it will be hard to claim that traders are producing prices that follow an optimum rule.

There also isn’t yet any theory regarding the feedback from underlying value to market value. Most of these markets seem to reflect changes in the underlying value by paying dividends to holders of the assets. Sometimes they (also? instead?) change the market maker’s current price for the asset. I expect this to make it difficult to come up with closed-form expressions of how the market operator’s exposure changes when assets are revalued. Remember that there’s no guarantee that similar amounts of money are invested in each market, or each side of any particular issue. And this kind of distinction is where the money pump was discovered in the TechBuzz Dynamic Parimutuel Mechanism. You can’t do this with real money if you don’t know what your exposure is.

These markets are definitely interesting. They allow Surowiecki’s Wisdom of Crowds to give us a hint about the rising and falling fortunes of things we care about. I welcome efforts to turn them into markets that can make predictions, or allow insurance, hedging, and the other side effects we expect from standard Prediction Markets.

Hat tip to Dave Pennock.

[This post was accidently embargoed due to a bug in our handling of WordPress drafts. It was sent out on the feed in October, but never appeared on the blog page.]

David Porter
of the Interdisciplinary
Center for Economic Science
at George
Mason University
wrote a very complimentary note to
Allan Schiffman,
CommerceNet’s Director. Professor Porter described Zocalo as the
backbone for the experiments that we are collaborating on with
Robin Hanson, Ryan Oprea, and Dorina Tila. He also said that he
knows of a half dozen US researchers who are planning to use Zocalo in
their research, and mentioned that he is hosting several foreign
researchers at George Mason who will also be introduced to Zocalo.
Over the summer, Porter spoke at GMU’s prestigious Graduate Student
, and demonstrated Zocalo to the attendees.

This isn’t the first time that Prof. Porter has talked about the researchers
he knows about who are using or planning to use Zocalo, but it’s the
most specific list I’ve seen. If you are planning to use Zocalo, or
strongly considering it, it would be useful if you got in touch with
me. I have been adding support recently for more configuration
options to the code, and the more I know about how people are hoping
to use the system, the better I’ll be able to trade off and prioritize
implementation choices.