3 ms·

Richard Feynman: This question of trying to figure out whether a book is good or bad by looking at it carefully or by taking the reports of a lot of people who

Richard Feynman: This question of trying to figure out whether a book is good or bad by looking at it carefully or by taking the reports of a lot of people who looked at it carelessly is like this famous old problem: Nobody was permitted to see the Emperor of China, and the question was, What is the length of the Emperor of China's nose? To find out, you go all over the country asking people what they think the length of the Emperor of China's nose is, and you average it. And that would be very "accurate" because you averaged so many people. But it's no way to find anything out; when you have a very wide range of people who contribute without looking carefully at it, you don't improve your knowledge of the situation by averaging. I think tarot cards are a lot simpler and more accurate. Mine suggest that Barack Obama should beware of the tall tower with the lightning bolts, and that a hooded figure carrying three sticks will bring him five golden discs. Meanwhile, Hillary Clinton will be menaced by a demon-like creature with the face of Rudy Giuliani, and will fight back by swinging swords at it while blindfolded. Or something.

byrneseyeview 19y ago
Your point? If they have no predictive power, you are guaranteed a profit by shorting front-runners and/or buying long-shots. If you believe that -- what are you waiting for?
- mechanical_fish 19y ago
  What does "no predictive power" mean? The Emperor's Nose technique has a certain amount of predictive power: it's very unlikely that it will be off by more than an inch, for example. Similarly, in the hands of a conscious reader such as myself, my tarot deck has lots and lots of predictive power. Right now it tells me that Mike Gravel does not have a 50% chance to become President this year, that Barack Obama will not live to be 200 years old, and that Bill Clinton, with 99.999% certainty, has not invented a perpetual motion machine. I'm not saying these markets are wrong. I'm suggesting that, at best, they're as correct as an off-the-cuff guess by the average reader of the New York Times. That's not all that impressive, but it's just good enough that I can't consistently outguess them without doing some research -- and, given that the total size of most of these penny-ante markets is less than a postdoc's annual salary, I'm never going to make any money doing that.
dejb 19y ago
People in this market aren't a random sample of people in the street. They are self-selected based on their willingness to bet money on the event. If they are poor predictors they will lose money on average and if they are good they will gain money. After many events the good predictors will tend to prevail over the poor ones because they have more money to bet and they'll be more eager to bet... Oh screw it I give up... some people just don't get odds or markets.
- foonamefoo 19y ago
  And you don't get Einstein's Random Walk.
  - dejb 19y ago
    I have a reasonable understanding of Brownian motion. Your comment indicates that you don't really understand markets. In an efficient market system (one which is perfectly predictive of chance of an event occurring based on observable information) it should not be possible to know if the price is going to go up or down. If could know for certain then obviously the price is wrong (and you could make a risk-free profit). So yes this would obviously look a lot like a random walk. However if you look at the graph of the odds through time (say for Obama winning the democratic nomination) you'll see it is responding to new information (such as poll results etc). It represents the aggregate opinion of the participants. If you were confident that your opinion is more accurate than this group then you could profit by betting. Oh crap I just wasted more time on this.
- mattmaroon 19y ago
  If there's one thing I know about gamblers, it's that their willingness to bet money on something generally has no correlation whatsoever to their knowledge about it. Their ability to continue betting over their lifetime certainly does, which of course means that any prediction market that doesn't run on real dollars is essentially useless. But that doesn't mean the for money ones are useful. This particular one is not. The infrequent nature of bets of this type, plus the small number of people able to get money onto sites like that in any reasonable timeframe keep it (at least where politics is concerned) from becoming anywhere near efficient. It's simply too small for much in the way of smart money to even bother with, so what you see is mostly just the result of sports bettors who watched an hour of CNN yesterday and have nothing else interesting to wager on. I wouldn't be surprised if they were extremely accurate for European sports, but for American politics they've been way off recently. The good news is, they're very easy to exploit in any close election year. Perhaps I'll blog about how after I do it in November.

byrneseyeview 19y ago

Your point? If they have no predictive power, you are guaranteed a profit by shorting front-runners and/or buying long-shots. If you believe that -- what are you waiting for?

mechanical_fish 19y ago
What does "no predictive power" mean? The Emperor's Nose technique has a certain amount of predictive power: it's very unlikely that it will be off by more than an inch, for example. Similarly, in the hands of a conscious reader such as myself, my tarot deck has lots and lots of predictive power. Right now it tells me that Mike Gravel does not have a 50% chance to become President this year, that Barack Obama will not live to be 200 years old, and that Bill Clinton, with 99.999% certainty, has not invented a perpetual motion machine. I'm not saying these markets are wrong. I'm suggesting that, at best, they're as correct as an off-the-cuff guess by the average reader of the New York Times. That's not all that impressive, but it's just good enough that I can't consistently outguess them without doing some research -- and, given that the total size of most of these penny-ante markets is less than a postdoc's annual salary, I'm never going to make any money doing that.

dejb 19y ago

People in this market aren't a random sample of people in the street. They are self-selected based on their willingness to bet money on the event. If they are poor predictors they will lose money on average and if they are good they will gain money. After many events the good predictors will tend to prevail over the poor ones because they have more money to bet and they'll be more eager to bet... Oh screw it I give up... some people just don't get odds or markets.

foonamefoo 19y ago
And you don't get Einstein's Random Walk.
- dejb 19y ago
  I have a reasonable understanding of Brownian motion. Your comment indicates that you don't really understand markets. In an efficient market system (one which is perfectly predictive of chance of an event occurring based on observable information) it should not be possible to know if the price is going to go up or down. If could know for certain then obviously the price is wrong (and you could make a risk-free profit). So yes this would obviously look a lot like a random walk. However if you look at the graph of the odds through time (say for Obama winning the democratic nomination) you'll see it is responding to new information (such as poll results etc). It represents the aggregate opinion of the participants. If you were confident that your opinion is more accurate than this group then you could profit by betting. Oh crap I just wasted more time on this.
mattmaroon 19y ago
If there's one thing I know about gamblers, it's that their willingness to bet money on something generally has no correlation whatsoever to their knowledge about it. Their ability to continue betting over their lifetime certainly does, which of course means that any prediction market that doesn't run on real dollars is essentially useless. But that doesn't mean the for money ones are useful. This particular one is not. The infrequent nature of bets of this type, plus the small number of people able to get money onto sites like that in any reasonable timeframe keep it (at least where politics is concerned) from becoming anywhere near efficient. It's simply too small for much in the way of smart money to even bother with, so what you see is mostly just the result of sports bettors who watched an hour of CNN yesterday and have nothing else interesting to wager on. I wouldn't be surprised if they were extremely accurate for European sports, but for American politics they've been way off recently. The good news is, they're very easy to exploit in any close election year. Perhaps I'll blog about how after I do it in November.