"Whenever I go into a restaurant, I order both a chicken and an egg to see which comes first"

Friday, March 9, 2012

BETTING vs. POLLS–BETTING WINS!

There has been a lot of discussion recently about how markets are better predictors of outcomes than polling; and how an aggregation of random individuals is better than selected experts; and finally how an aggregation of cognitively diverse individuals is perhaps the best of all.

This is a fascinating field of inquiry, for it challenges the received wisdom that specialists, scholars, pundits, and other experts are the final authority on predicting events.  However, research carried out over the last few years shows just the opposite – that groups of random individuals either betting on future events or simply predicting them, are better than those ‘in the know’.  This collective ‘wisdom’ is not restricted to electoral politics, but foreign affairs as well.  We should better trust the collective wisdom of 100 million American bettors on the outcomes of a war in Iraq or Vietnam or the chances that Israel would bomb Iran than on experts, politicians, or Presidential advisors.

These theories are fascinating because they are counter-intuitive, valuing numbers of random individuals over individual expertise.  In short, get enough people together and ask them if gas prices will hit $5.00 next year, they will probably generate the right answer.

Betting is the best example of how random individuals can predict better than experts.  The first case is that of Iowa’s Electronic Market (IEM) where random individuals are invited to bet on a particular result, for example a political campaign.  They must actually bet money or they can take the more conservative play money option.  These betting results are compared to expert pundit predictions and/or the results from polls.  The markets consistently outperform the pundits and pollsters.  Money Talks Online News had the following comments and explanations:

Joyce Berg mentioned in the video, the markets beat the polls 75 percent of the time. On another popular site, Intrade, the majority of gamblers correctly guessed the winner of every state in the 2004 presidential election, and only missed two in 2008. How come?

There are whole books about this subject, like The Wisdom of Crowds. But here are three simple reasons it make sense…

1. Polls ask for opinions and personal preference – who do you want to win, do you agree with the candidates’ stance on issue X, and so forth. Bets are made to win money, not to express personal views, which may be subject to personal preferences and prejudices.

2. Randomness doesn’t ensure informed responses. Any pollster worthy of the name understands the importance of random sampling and sample size – more people from more backgrounds usually gets a result with higher accuracy and better representation of the group. With polls, there’s no particular reason for the people who get randomly picked to be well-informed about the candidates or issues. Even though pollsters recognize that and try various techniques to weight responses and avoid bias, they can’t match the obvious incentive gamblers have to pay attention: money. In short, those who are putting their money where their mouths are should be better informed.

3. Markets have more data. Individual polls might survey a thousand random people about a particular question, but the people putting their money on the line look at all of the polls and plenty of other factors – debates, the economy, political history, the latest news scandals, and so on. MONEY TALKS ONLINE NEWS

Betting on presidential elections was legal for many years, and the outcomes were far more accurate than any ‘informed’ prediction: Rohde and Strumpf, academic researchers have written: 

Wagering on political outcomes has a long history in the United States. As Henry David Thoreau noted in 1848, “All voting is a sort of gaming,… and betting naturally accompanies it” (Thoreau, 1967, p. 36). This paper analyzes the large and often well organized markets for betting on presidential elections that operated between 1868 and1940. Over $165 million (in 2002 dollars) was wagered in one election, and betting
activity at times dominated transactions in the stock exchanges on Wall Street.Drawing on an investigation of several thousand newspaper articles, we developand analyze data on betting volumes and prices to address four main points.

First, we show that the market did a remarkable job forecasting elections in an era before scientific polling. In only one case did the candidate clearly favored in the betting a month before Election Day lose, and even state-specific forecasts were quite accurate. This performance compares favorably with that of the Iowa Electronic Market (currently the
only legal venue for election betting in the U.S.). Second, the market was fairly efficient,despite the limited information of participants and attempts to manipulate the odds by political parties and newspapers

Regarding the second alternative – simple random aggregations – Michael Rozeff cites various sources:

James Surowiecki in his book The Wisdom of Crowds tells the story of the game show "Who Wants to be a Millionaire" in which a contestant could ask an expert for help with a question or ask the audience. The experts were right 65 percent of the time, and the audience was right 91 percent of the time.

Jude Wanniski related a story told to him by Jack Treynor, a finance guru. Treynor had his class guess the number of jelly beans in a jar holding 850 beans. The average guess was within 3 percent of the total. Wanniski, by the way, correctly realized that this supported the efficiency of financial markets. He also, in my opinion incorrectly, construed this as proof of the efficiency of political markets, an opinion he expanded upon in The Way the World Works.

At 3:15 p.m. on May 27, 1968, the submarine USS Scorpion was officially declared missing with all 99 men aboard. She was somewhere within a 20-mile-wide circle in the Atlantic, far below implosion depth. Five months later, after extensive search efforts, her location within that circle was still undetermined. John Craven, the Navy's top deep-water scientist, had all but given up. As a last gasp, he asked a group of submarine and salvage experts to bet on the probabilities of different scenarios that could have occurred. Averaging their responses, he pinpointed the exact location (within 220 yards) where the missing sub was found."

Finally, Surowiecki, the proponent of ‘cognitive diversity” wrote the following in The Wisdom of Crowds:

What was missing most from the [analysis of the Apollo 7 near disaster], of course, was diversity, by which I mean not sociological diversity but rather cognitive diversity. James Oberg, a former Mission Control operator and now NBC News correspondent, has made the counterintuitive point that the NASA teams that presided over the Apollo missions were actually more diverse than the MMT. This seems hard to believe, since every engineer at Mission Control in the late 60's had the same crew cut and wore the same short-sleeved white shirt. But as Oberg points out, most of those men had worked outside of NASA in many different industries before coming to the agency. NASA employees today are far more likely to have come to the agency directly out of graduate school, which means they are also far less likely to have divergent opinions. That matters because, in small groups, diversity of opinion is the single best guarantee that the group will reap benefits from face-to-face discussion."

Michael Rozeff concludes by saying that

“Prediction markets in general perform exceedingly well compared to individual forecasts. In his article on prediction markets, Philip O'Connor writes: "In fact, studies of prediction markets have found that the market price does a better job of predicting future events than all but a tiny percentage of individual guesses. The analysis below of the Virtual Super 12 shows the average selection, an average or constructed market price, to be better than 99% of participants' selections."

If betting is so much better than polling or punditry, why do we continue to rely on the latter?  Are we really so mistrustful of the unwashed masses?  Is there something in us which refuses, or prefers to ignore the oxymoronic ‘wisdom of the uninformed’? Are we so in awe of educated experts and do we feel so inferior to them that we feel obliged to listen to them? 

There may also be a political/ideological bias to the theory that markets – once again – are right.  It may be hard for many academics to accept that money talks; that when people have to put money up front, they make better decisions than when they do not; and worse, when a political stock market, like IEM, is created, it somehow sullies the more pure form of informed dialogue, intelligent discussion, the community of ideas. 

In any case, times are changing; and we should accept what works not what should work.

No comments:

Post a Comment