Another coin-flipping game
The classical investment philosophy book “Random Walk down Wall Street”, by Burton Malkiel likes to tell repeatedly the coin-flipping story. Keep flip a coin and log down the sequence of heads and tails. The appearance of heads and tails is analogous to stock prices’ up and down (up or down direction, not absolute magnitude of up and down). It is not surprised to see at the end of the game the heads/tails chance even out. Given the plotting of a coin-flipping results and a real stock’s market performance to experienced trader, he can’t tell which one is which. This is a strong argument that the market moves randomly.
This experiment has one key assumption: it is no difference doing the flipping in sequence or throwing a large amount of ordered coins at once and then record the results. This is because each flipping has no effect to the previous or next flipping.
Many other coin-flipping games have been developed to prove or disapprove the random nature of the market. Most of them assume two consecutive flipping events are independent. There are some others, at minority, that assume flipping events have some connections. The fundamental is that if we acknowledge that coin flipping event’s independence and the market’s up-and-down are alike. In more practical thought, we can’t predict head-and-tail of a coin flipping but can we predict what the market will move?
We can make a new coin flipping gain. In the game, if a series of heads appears, say we’ve seen 3 heads in a row, the probability of a 4th head is diminishing exponentially. And if there are 4 heads in a row, the probability of a 5th head is also further less likely. The probability can be modeled in many ways. And let the game go on for large amount of data. The eventual outcome is not as easily as the game described in Malkiel’s book because that involves how we describe the probability is modeled. No one can do that correctly because that means he can predict the market accurately.
But this new game is more realistic to the market.
This experiment has one key assumption: it is no difference doing the flipping in sequence or throwing a large amount of ordered coins at once and then record the results. This is because each flipping has no effect to the previous or next flipping.
Many other coin-flipping games have been developed to prove or disapprove the random nature of the market. Most of them assume two consecutive flipping events are independent. There are some others, at minority, that assume flipping events have some connections. The fundamental is that if we acknowledge that coin flipping event’s independence and the market’s up-and-down are alike. In more practical thought, we can’t predict head-and-tail of a coin flipping but can we predict what the market will move?
We can make a new coin flipping gain. In the game, if a series of heads appears, say we’ve seen 3 heads in a row, the probability of a 4th head is diminishing exponentially. And if there are 4 heads in a row, the probability of a 5th head is also further less likely. The probability can be modeled in many ways. And let the game go on for large amount of data. The eventual outcome is not as easily as the game described in Malkiel’s book because that involves how we describe the probability is modeled. No one can do that correctly because that means he can predict the market accurately.
But this new game is more realistic to the market.
posted by Intl.investment.services at
12:51 PM
0 Comments:
Post a Comment
Subscribe to Post Comments [Atom]
<< Home