Actually, I think a truly efficient market shouldn’t just skip around across orders of magnitudes, just because expectations of future prices do. I think truly efficient markets show some degree of “drag”, which should be invisible in typical cases like publicly-traded stocks, but become noticeable in cases of order-of-magnitude value-uncertainty like Bitcoin.
Imagine if everyone agreed that the best way to calculate Bitcoin’s expected future value was to look at a single source of news: the Bitcoin guru. Every day, if the Bitcoin guru goes on TV with his thumbs up, then everyone agrees that that is worth a Bayesian update of 1.25x to the expected future price, while thumbs down is worth a Bayesian update of 0.8x. And no one has a better model of how the Bitcoin guru’s thumb works than the fair-coin-flip model.
In other words, pretend you know that everyone’s expected future value of Bitcoin follows a log random walk. Now you can use inductive reasoning to conclude: If the expected future value of BTC is $1000/coin today, then it will be either $1250 or $800 tomorrow.
I used to think an “efficient market” was necessarily a market which current price captures people’s consensus expectation of future prices. But in my example, it seems possible to have a guaranteed-positive-return trading strategy: investing say 10% of your portfolio in BTC, and constantly trading as required to rebalance your 10% asset allocation.
There are a few names you can use for that...
“Drag Trading”: trading that causes market prices to be dragged toward their past value and away from their current expected future value.
“Shannon’s Demon”: a kind of Drag Trading—trading to maintain a constant % of BTC in your portfolio, i.e. near-continuous portfolio rebalancing.
“Kelly Criterion Trading”: An instance of the Shannon’s Demon strategy, where you choose your asset percentages in proportion to their long-term future relative value, which you derive by asking yourself what relative percentages of the world’s wealth are stored in your different asset classes.
My big point has two parts:
I think I’ve shown that an “efficient market” might not have prices that directly reflect a consensus of future prices, but instead might show some “price drag”.
Bitcoin’s wildly swinging prices is evidence that it is lacking this “price drag”, unless the model of expected future price drag is even more volatile than the exponentially-volatile one in my thought experiment.
But in my example, it seems possible to have a guaranteed-positive-return trading strategy: investing say 10% of your portfolio in BTC, and constantly trading as required to rebalance your 10% asset allocation.
You assume trading. Who will be your counterparty?
If the expected future value of BTC is $1000/coin today, then it will be either $1250 or $800 tomorrow.
Given that the expected value for the change between today and tomorrow ((+250-200)/2=+25) is publicly known, I wonder who will sell him bitcoins for $1000 today.
In other words, the situation as described is unstable and will not exist (or, if it will appear, it will be arbitraged away very very quickly).
What Vaniver said. Also, emperically, you can look at the current price/order book on an exchange and see that people are in fact willing to sell you these things. If my holdings represented a life altering sum of money it would be time to take less risk and I would be one of those people.
Sigh. Again, look at the context. There is a claim
it seems possible to have a guaranteed-positive-return trading strategy: investing say 10% of your portfolio in BTC, and constantly trading as required to rebalance your 10% asset allocation.
Ah, you’re disagreeing with the model and phrasing it as “if that model were true, no one would sell you btc, but people are willing to sell, therefore that model is false.” Do I understand?
If so, I do not agree that “if that model were true, no one would sell you btc” is a valid inference.
Given that the expected value for the change between today and tomorrow ((+250-200)/2=+25) is publicly known, I wonder who will sell him bitcoins for $1000 today.
If someone has a log utility function, a half chance of $800 and $1250 is as valuable as a certainty of $1000. Basically, people who have more risk than they want are selling to people that have less risk than they want.
(The stated example- of 2.5% growth in absolute terms per day- is very exaggerated compared to actual asset prices, I think. If this were about an asset that had an expectation of 2.5% growth in absolute terms per year, but the high variance, then it would be reasonable to imagine the market being much happier with the $1000 today than the gamble, because of how risky and low-growth it is compared to other options.)
I’m having a little trouble understanding the model.
I used to think an “efficient market” was necessarily a market which current price captures people’s consensus expectation of future prices. But in my example, it seems possible to have a guaranteed-positive-return trading strategy: investing say 10% of your portfolio in BTC, and constantly trading as required to rebalance your 10% asset allocation.
Is the Bitcoin Guru correct in his predictions? If so, is he taking into account that people might be using your drag trading strategy? If not, then it sounds like the strategy no longer offers guaranteed returns.
No, he doesn’t take into account people’s drag trading behavior, because his conclusion is dominated by evidence about the underlying value of Bitcoin as usable money in the long-term.
To make slightly more concrete what it means to say that the Guru’s predictions are accurate, could we say that all his predictions are for what the price will be on Jan 1st, 2019? So when we say he’s accurate, we mean that on that day the actual price will be some starting price, times the product of a bunch of 1.25s and .8s according to whatever his predictions were on all the days in between.
Can you elaborate on why you think this is true?
O hai.
Imagine if everyone agreed that the best way to calculate Bitcoin’s expected future value was to look at a single source of news: the Bitcoin guru. Every day, if the Bitcoin guru goes on TV with his thumbs up, then everyone agrees that that is worth a Bayesian update of 1.25x to the expected future price, while thumbs down is worth a Bayesian update of 0.8x. And no one has a better model of how the Bitcoin guru’s thumb works than the fair-coin-flip model.
In other words, pretend you know that everyone’s expected future value of Bitcoin follows a log random walk. Now you can use inductive reasoning to conclude: If the expected future value of BTC is $1000/coin today, then it will be either $1250 or $800 tomorrow.
I used to think an “efficient market” was necessarily a market which current price captures people’s consensus expectation of future prices. But in my example, it seems possible to have a guaranteed-positive-return trading strategy: investing say 10% of your portfolio in BTC, and constantly trading as required to rebalance your 10% asset allocation.
There are a few names you can use for that...
“Drag Trading”: trading that causes market prices to be dragged toward their past value and away from their current expected future value.
“Shannon’s Demon”: a kind of Drag Trading—trading to maintain a constant % of BTC in your portfolio, i.e. near-continuous portfolio rebalancing.
“Kelly Criterion Trading”: An instance of the Shannon’s Demon strategy, where you choose your asset percentages in proportion to their long-term future relative value, which you derive by asking yourself what relative percentages of the world’s wealth are stored in your different asset classes.
My big point has two parts:
I think I’ve shown that an “efficient market” might not have prices that directly reflect a consensus of future prices, but instead might show some “price drag”.
Bitcoin’s wildly swinging prices is evidence that it is lacking this “price drag”, unless the model of expected future price drag is even more volatile than the exponentially-volatile one in my thought experiment.
Therefore drag-trading Bitcoin seems promising.
You assume trading. Who will be your counterparty?
Wat? A liquid market is a standard assumption.
Are you talking about your assumptions or are you taking about reality?
Bitcoin is plenty liquid right now unless you’re throwing around amounts > $1 mil or so.
Look at the grandparent:
Given that the expected value for the change between today and tomorrow ((+250-200)/2=+25) is publicly known, I wonder who will sell him bitcoins for $1000 today.
In other words, the situation as described is unstable and will not exist (or, if it will appear, it will be arbitraged away very very quickly).
What Vaniver said. Also, emperically, you can look at the current price/order book on an exchange and see that people are in fact willing to sell you these things. If my holdings represented a life altering sum of money it would be time to take less risk and I would be one of those people.
Sigh. Again, look at the context. There is a claim
Which happens to be wrong.
Ah, you’re disagreeing with the model and phrasing it as “if that model were true, no one would sell you btc, but people are willing to sell, therefore that model is false.” Do I understand?
If so, I do not agree that “if that model were true, no one would sell you btc” is a valid inference.
Essentially, the model says “there is free money lying on the ground, just picking it up is a ‘guaranteed-positive-return trading strategy’.”
I am pointing out that free money lying on the ground is an illusion.
If someone has a log utility function, a half chance of $800 and $1250 is as valuable as a certainty of $1000. Basically, people who have more risk than they want are selling to people that have less risk than they want.
(The stated example- of 2.5% growth in absolute terms per day- is very exaggerated compared to actual asset prices, I think. If this were about an asset that had an expectation of 2.5% growth in absolute terms per year, but the high variance, then it would be reasonable to imagine the market being much happier with the $1000 today than the gamble, because of how risky and low-growth it is compared to other options.)
I’m having a little trouble understanding the model.
Is the Bitcoin Guru correct in his predictions? If so, is he taking into account that people might be using your drag trading strategy? If not, then it sounds like the strategy no longer offers guaranteed returns.
Yes, the Bitcoin Guru makes accurate predictions.
No, he doesn’t take into account people’s drag trading behavior, because his conclusion is dominated by evidence about the underlying value of Bitcoin as usable money in the long-term.
Hmm, I might just be misunderstanding.
To make slightly more concrete what it means to say that the Guru’s predictions are accurate, could we say that all his predictions are for what the price will be on Jan 1st, 2019? So when we say he’s accurate, we mean that on that day the actual price will be some starting price, times the product of a bunch of 1.25s and .8s according to whatever his predictions were on all the days in between.
Does this match what you had in mind?