Intrade is an interesting suggestion, but I don’t think he could make enough on it. He wants 8000 USD a year, and even if we assume he can get 10%, he’ll still need 80k invested.
I don’t think he has 80k to spare, and I have to wonder—is 10% feasible in the long run? I could see getting it in an election year easily, because the markets are so volatile and heavily traded, but what about off-years?
Well, I should clarify that I think a smart bias-educated person can beat the prediction markets fairly easily—I doubled my (small) investment in the IEM just by exploiting some obvious biases in the last presidential election, and I know I’m not the smartest bear around. My doubt is whether he can beat the market enough: any sum of money CronoDAS has is likely small enough he would need really absurd returns.
Are there differences between prediction and markets that make it easier for a “smart bias-educated person” to win fairly easily?
Besides what Nick said, people seem to treat prediction markets more as entertainment than seriously. For example, Ron Paul or Al Gore should never have broken 1%, and Hillary shares were high long after it became obvious she wasn’t going to make the nomination. These were all pretty clear to anyone suspicious of fanciful wouldn’t-it-be-fun? scenarios and being biased towards what one would like to happen.
If you think its fairly easy, then I’d be curious to know whether you’re putting your money where your mouth is… how much have you invested?
I started in the IEM with ~$20, and even after taking some heavy losses in 2004 and whatever fee the IEM charged ($5?), I still cashed out $38 in 2008. If you’re interested in more details, see my http://www.gwern.net/Prediction%20markets
I appreciate your careful documentation. And I thought these words of yours were wise:
“I often use them [prediction markets] to sanity-check myself by asking ‘If I disagree, what special knowledge do I have?’ Often I have none.”
Words are vague, lets use numbers. Say you were forced to invest $1000 in the prediction markets over the next year. What probability would you assign various outcomes: e.g. [-100%,-50%], [-50%,-25%] [-25%,-10%], [-10%,0] [0,10%], [10%,25%] [25%, 50%], [50%,100%], [100%, 200%], and [200%, 1000000%]
One must be wary of faux precision. But I think I would put the odds of >100% or <-40% at under 30%; I’d assign another 10 or 20% to a gain between 30% and 100%, and leave the rest to the range of small losses/gains.
The ten categories I suggested may be a bit excessive, but it would be much easier to judge if you were a little more precise. You acknowledge a non-trivial chance of losing a non-trivial amount of money. The confusion is that I thought your previous statement that a “smart bias-educated person can beat the prediction markets fairly easily” would preclude this.
You acknowledge a non-trivial chance of losing a non-trivial amount of money. The confusion is that I thought your previous statement that a “smart bias-educated person can beat the prediction markets fairly easily” would preclude this.
There are arbitrage opportunities, but they’re not what I’m thinking of.
An analogy: knowing about biases and how to play optimally is important to play poker at any high level; but that still doesn’t mean you’re going to win every hand. I might correctly call an election for Obama, but that’s not going to help me as a trader if he abruptly dies of a heart-attack or Sarah Palin stages a coup with a crack unit of Alaskan hunters—I’ll still lose my money. I don’t see any contradiction here.
My answer to whether there are differences between prediction [markets] and markets was no, except in as much as prediction markets that are currently active are far larger (noise cancellation), more heavily traded (more information from more experts is represented already), and have had longer for biasses to be exploited and so corrected for.
Intrade is an interesting suggestion, but I don’t think he could make enough on it. He wants 8000 USD a year, and even if we assume he can get 10%, he’ll still need 80k invested.
I don’t think he has 80k to spare, and I have to wonder—is 10% feasible in the long run? I could see getting it in an election year easily, because the markets are so volatile and heavily traded, but what about off-years?
Agreed. We should always be skeptical of an individual’s ability to beat the market.
Well, I should clarify that I think a smart bias-educated person can beat the prediction markets fairly easily—I doubled my (small) investment in the IEM just by exploiting some obvious biases in the last presidential election, and I know I’m not the smartest bear around. My doubt is whether he can beat the market enough: any sum of money CronoDAS has is likely small enough he would need really absurd returns.
Are there differences between prediction and markets that make it easier for a “smart bias-educated person” to win fairly easily?
If you think its fairly easy, then I’d be curious to know whether you’re putting your money where your mouth is… how much have you invested?
Besides what Nick said, people seem to treat prediction markets more as entertainment than seriously. For example, Ron Paul or Al Gore should never have broken 1%, and Hillary shares were high long after it became obvious she wasn’t going to make the nomination. These were all pretty clear to anyone suspicious of fanciful wouldn’t-it-be-fun? scenarios and being biased towards what one would like to happen.
I started in the IEM with ~$20, and even after taking some heavy losses in 2004 and whatever fee the IEM charged ($5?), I still cashed out $38 in 2008. If you’re interested in more details, see my http://www.gwern.net/Prediction%20markets
I appreciate your careful documentation. And I thought these words of yours were wise: “I often use them [prediction markets] to sanity-check myself by asking ‘If I disagree, what special knowledge do I have?’ Often I have none.”
Words are vague, lets use numbers. Say you were forced to invest $1000 in the prediction markets over the next year. What probability would you assign various outcomes: e.g. [-100%,-50%], [-50%,-25%] [-25%,-10%], [-10%,0] [0,10%], [10%,25%] [25%, 50%], [50%,100%], [100%, 200%], and [200%, 1000000%]
One must be wary of faux precision. But I think I would put the odds of >100% or <-40% at under 30%; I’d assign another 10 or 20% to a gain between 30% and 100%, and leave the rest to the range of small losses/gains.
The ten categories I suggested may be a bit excessive, but it would be much easier to judge if you were a little more precise. You acknowledge a non-trivial chance of losing a non-trivial amount of money. The confusion is that I thought your previous statement that a “smart bias-educated person can beat the prediction markets fairly easily” would preclude this.
There are arbitrage opportunities, but they’re not what I’m thinking of.
An analogy: knowing about biases and how to play optimally is important to play poker at any high level; but that still doesn’t mean you’re going to win every hand. I might correctly call an election for Obama, but that’s not going to help me as a trader if he abruptly dies of a heart-attack or Sarah Palin stages a coup with a crack unit of Alaskan hunters—I’ll still lose my money. I don’t see any contradiction here.
Yes. Prediction markets are far smaller, and have far less intelligence devoted to exploiting away their irrationalities.
Efficiency.
my question was about how much more efficient the stock market is, and why.
My answer to whether there are differences between prediction [markets] and markets was no, except in as much as prediction markets that are currently active are far larger (noise cancellation), more heavily traded (more information from more experts is represented already), and have had longer for biasses to be exploited and so corrected for.
Efficiency.