Daily Archives: November 20, 2006

Overcoming Bias dot com = Robin Hanson’s group blog on truth discovery and decision rationality

Posted on by
Reply

A Web-based “forum”, rather, he says:

To me “forum” connotes that we [CFM: the blogging scholars] are primarily talking to each other, though we don’t mind if others join in to comment or listen.

“Blog” to me connotes that we are primarily writing for other people, and we are just sharing the load of putting together something for those readers.

—

URL: Overcoming Bias dot com

Overcoming Bias

—

Addendum: Robin Hanson has just posted a comment…

Chris, the picture is a famous painting of Ulysses bound to the mast listening to the Sirens; the rest of the crew has their ears plugged to avoid the severely biased Siren Song. 

—

What Is The Meaning Of His Blog Header? No idea. I wonder whether his 2005 Marginal Revolution post (”Hanged For Accuracy”) gives us a clue:

I came across an even more dramatic example of such thinking in Dava Sobel’s Longitude (1995:11-12):

Returning home victorious from Gibraltar after skirmishes with the French … the English fleet … discovered to their horror that they had misgauged their longitude … the Scillies became the unmarked tombstones for two thousand of Sir Clowdisley’s troops. [Admiral Sir Clowdisley] had been approached by a sailor, … who claimed to have kept his own reckoning of the fleet’s location during the whole cloudy passage. Such subversive navigation by an inferior was forbidden in the Royal Navy, as the unnamed seaman well knew. However, the danger appeared so enormous, by his calculations, that he risked his neck to make his concerns known to the officers. Admiral Shovell had the man hanged for mutiny on the spot. … In literally hundreds of instances, a vessel’s ignorance of her longitude led swiftly to her destruction.

Even though shipmates had a strong common interest in knowing their longitude, other social incentives apparently prevented them from sharing their information. As a consultant on the use of prediction markets within organizations, I’ve also noticed that managers are often surprisingly uninterested in the prospect of more accurate forecasts and more informed decisions. Could these phenomena have similar explanations?

—

About Overcoming Bias:

How can we better believe what is true? While it is of course useful to seek and study relevant information, our minds are full of natural tendencies to bias our beliefs via overconfidence, wishful thinking, and so on. Worse, our minds seem to have a natural tendency to convince us we that are aware of and have adequately corrected for such biases, when we have done no such thing.

—

Overcoming Bias dot com will be authored by 14 (academic or not) scholars. Among them, 5 usual suspects from the field of prediction markets, including the owner of this microscopic little blog (who would do just anything to get linked to by Midas Oracle).

—

Speaking of bias, is our good doctor Robin Hanson as innocent as Snow White? Let’s take a look at “The Wisdom Of His Crowd“. Here’s an excerpt of the 2005 poll he asked his acquaintances (the author of these lines being one among many) to fill in, so as to discover what could be his next academic project.

Here are the ten main choices as I see them now:

1. Disagreement Book – Expand “Are Disagreements Honest” and related papers into a book, adding new material on data about who is right in real disagreements. I’ve been telling people this is my plan. This could establish my reputation as a deep thinker on a big issue. Fun, as there are still things for me to learn on this topic. No real competition on this topic (as least re the more technical angle), and it is nicely not aligned with an ideology. But not clear this will really change much in the world.

See the key sentence??? “I’VE BEEN TELLING PEOPLE THIS IS MY PLAN.” Ha. ha. ha. Totally biased poll. And, SURPRISE, SURPRISE, of course, that poll gave the option #1 (”the disagreement book”) as the most popular answer. NO WONDER. And so we are here, today, with our Robin Hanson opening a group blog on “overcoming bias”. (The “idea futures book” came as a close #2. Had he suppressed the bias in his poll, we would have had Robin Hanson opening a group blog on prediction markets, today, instead.)

Maybe the first topic of discussion among these 14 luminaries (or so they think they are) should be: How to overcome Robin Hanson’s biased polls?

—

How To Subscribe To Robin Hanson’s Group Blog:

His Royal Highness publicizes the “RSS 1.0″ site feed, on his right sidebar. It’s an old format; complete crap.

Summary of the Tradesports DEMS.HOUSE.OVER29.5 issue, and the TS credibility gap

Posted on by
1

No Gravatar

In the past weeks, there was something of a dust-up over TS&#8217- handling of the DEMS.HOUSE.OVER29.5 contract. According to all American media, the Democrats needed 15 new House seats to gain control of the House of Representatives. And, indeed, in the final days leading up to the election, there an excellent synchronization between the DEMS.HOUSE.OVER14.5 contract and the inverse of HOUSE.GOP.2006, indicating a perceived equivalence between 100 – p(HOUSE.GOP.2006) and DEMS.HOUSE.OVER14.5. Thus, the (presumably overwhelmingly American) market inferred that TS would follow the same convention as all American media, mainstream and otherwise, did. And according to American media, a Democrat replacing independent Socialist Bernie Sanders didn&#8217-t count as a Democratic pickup, because he already caucused with the Democrats anyway. However, TS initially disagreed with that, and counted Sanders as a non-Democrat. (More on that soon.)

Speaking personally, TS reinforced this perception&#8211-that independents who caucused with the Democrats counted as Democratic seats, and if they were replaced by an official Democrat, that wouldn&#8217-t increase the Democrats&#8217- vote total&#8211-by specifically stating, in regards to SENATE.GOP.2006, that a &#8220-loss&#8221- of the Connecticut seat to Joe Lieberman (who had switched from Democrat to Independent) would still be counted as a Democratic seat, because Lieberman already caucused with the Democrats anyway.

The fine print of the Tradesports DEMS.HOUSE.OVERXX contracts was, for most of the time, fairly clear in stating that the Democrats started from 201 seats, and anything over that number would constitute a gain. Judging from the activity of DEMS.HOUSE.OVER14.5, however, I believe that most of the market inferred (as did I) that 201 was the initial starting number of of Democrat seats according to American convention as well as TS, and didn&#8217-t realize the difference between the two systems. A very technical mistake, but not one for which TS deserved blame.

However, TS threw a monkey wrench in the system by telling forummer &#8220-gekko6&#8243- that Bernie Sanders&#8217- Independent seat going Democratic would not count as a Democratic pickup, because Sanders had already caucused with the Democrats in the first place. In this decision, as in its Connecticut Senate decision, TS showed an impressive grasp of the vagaries of the American political system&#8211-namely, that the size of each caucus was the real issue- and Congressional majorities being determined on the basis of caucus, not party affiliation, it only made sense to calculate the shift in power in the House on the same basis.

Unfortunately, that also meant that TS had, at the same time, repudiated its own convention for what constituted a &#8220-gain.&#8221- &#8220-gekko6&#8243- had already been very aware of this issue, because he had pointed out that the actual starting count according to the American system was 203, not 201&#8211-due to Sanders&#8217- being an independent endorsed and unopposed by the Democrats, and the vacancy of Bob Menendez&#8217-s seat after Menendez was appointed to the Senate by NJ Gov. Corzine. (From a foreign, technical perspective, this was defensible&#8211-Menendez&#8217-s seat was vacant, so the winner on Nov. 7 would count it as a &#8220-gain.&#8221- However, the Republicans did not contest Menendez&#8217-s very Democratic seat, so in the American convention, it was never counted as a Democratic pickup.) Tradesports&#8217- convention effectively said that the Democrats gained two more seats than the American convention did, and several days after the November elections, the American convention said that the Democrats had gained 28 House seats with 1 certain to go Democratic in a runoff (in Louisiana), so 29 seats total, while the Democrats had effectively gained 31 seats according to TS. Hence the problem with DEMS.HOUSE.OVER29.5. And while a legalistic interpretation would favor TS, I believe that the synchronization of DEMS.HOUSE.OVER14.5 and 100 – p(HOUSE.GOP.2006) indicated that most trading during the final few days showed that the market was unaware of TS&#8217- own convention for the election outcome.

I sent Tradesports an e-mail about it (apparently they don&#8217-t accept new entrants to their forum anymore, because my application has been pending for about a month), and they replied that the new number of Democrats minus 201 would constitute the number of Democrat gains, thus contradicting what they had told &#8220-gekko6.&#8221- I then publicly denounced TS for waffling the issue. TS did nothing, and apparently hoped the controversy would blow over. A bunch of recounts in close races later, TS appears to have lucked out, because according to the American convention, the Democrats now have at least 233 House members, up from 202-plus-one-Socialist, so the American system now says the Democrats have gained 30 seats and DEMS.HOUSE.OVER29.5 has been fulfilled either way.

Honestly, I was impressed that TS understood the American system as well as they did. Unfortunately, TS&#8217- subsequent &#8220-flip-flopping&#8221- showed that it did not, in fact, understand its own contract specifics. It also fit a larger pattern of cavalier disdain for its clients, often interpreting an ambiguous outcome significantly contrary to that of the market-majority (Harriet Miers confirmation, NK missile test contract) without appropriate compensation, setting up a joke &#8220-Arbitration Committee&#8221- that, if it even exists, has done nothing except infuriate customers, and most recently expiring sports contracts before the games were even concluded &#8211-and coincidentally raking in tons of expiry fees from people who weren&#8217-t given a chance to liquidate.

The latter, in fact, has happened with enough cavalier consistency that one can only wonder whether TS simply plans on milking its American consumers for as much as possible before closing the site to new American participants as a result of the recent US legislation. SportsBook has downgraded Tradesports to a C+ rating, which is at the very low end of what SportsBook can vouch for.

If TS plans on more effectively structuring its futures contracts, it should structure them along the lines of, &#8220-At 11:59:59 PM GMT on dd/mm/yyyy, Democrats will control ON or OVER XXX seats in the US House of Representatives.&#8221- If it wants to undertake a broader effort to restore its own credibility, it needs to stop the caprice that is fast becoming the norm for how it adjudicates contract outcomes&#8211-whether that adjudication occurs before or after the outcome has actually occurred. [added:] TS&#8217- infrastructure and trader base are both excellent, and there&#8217-s no point in wasting those assets on sloppy legalese and interpretation.

&#8211-Alex Forshaw

http://the-ts-maven.blogspot.com

Tradesports forum homepage: http://forum.tradesports.com/

Addendum: As for 100 – p (HOUSE.GOP.2006) vs DEMS.HOUSE.OVER14.5, I remember them consistently mirroring one another, and thinking to myself, &#8220-Aha! Efficient markets at work.&#8221- However, what little historical data TS makes publicly available makes it hard to judge that, and there were snapshots when the two were de-coupled, so perhaps a minority of perceptive traders _did_ trade on the differences in the rules. But in the sample of snapshots that I looked at, the two were coupled (within 2 points) much more often than not during the final days, when liquidity was high enough to make arbitraging the two contracts a worthwhile use of capital.