Use random walk to make $$$

Quote from Corey:

In my opinion, Gil Blake proves you wrong. Fish swim together.
More...

I don't think anyone would disagree with you but that does not disprove a random walk. A random walk means that today your best guess for tomorrow is:

tomorrow's price = today's price + unknowable surprise

The fact that surprise for stock A is related to surprise for stock B does not disprove a random walk.

Sorry for the geek outburst. I do think lots of other things disprove it, at least for some stocks and indexes some of the time, but that doesn't.
 
I say the markets aren't random because there is always a reason for why they are doing something ... 99% of us just simply are not aware of it at the time that it happens, and normally will never become aware of it unless it is really news worthy or we are plugged-in to an information network.

So while the markets are not random in the true meaning of the word, they are probabilital.

This probabilital nature can be defined as: operating in a state of constant flux with the best awareness that we can have of them being what state they are in right now.

As an example, the different states can be labeled.

Up = Strong persistent price action in an upwards direction
Neutral = Non-persistent price action moving in an upwards and/or downwards direction
Down = Strong persistent price action in a downwards direction

I think this model of the financial market(s) (anyone of them) is a much more accurate description of what is taking place than saying that they are random, and how you choose to use the information in your trading is up to you.

GT
 
Moving on the next level of understanding of this phenomena is what successful traders do.

They use models (however you want to define the term "model") to determine the probability of the current state continuing, taken their position(s) accordingly, and use judicious risk management to insure that they can continue to engage in the process even if their judgement is flawed and they are proven to be wrong.
 
Quote from black diamond:

I don't think anyone would disagree with you but that does not disprove a random walk. A random walk means that today your best guess for tomorrow is:

tomorrow's price = today's price + unknowable surprise

The fact that surprise for stock A is related to surprise for stock B does not disprove a random walk.

Sorry for the geek outburst. I do think lots of other things disprove it, at least for some stocks and indexes some of the time, but that doesn't.
More...

Right, but Gil Blake is saying, if I am interpreting him correctly, that the 'unknownable surprise' is actually predictable, with high success, on certain days.
 
MandelbrotSet:

Thanks for that point. I agree. So far, I only feel close to perfecting the exit.

But I only scratched the surface at a models for the entry. Now that I understand exits, money management and position sizing, I have to completely revisit all my research into entries.

Currently, I do have one based on probability of continuing, like you said.

However, it is excruciatingly simplistic. So I personally have more work in that area.

Sincerely,
Wayne

Quote from MandelbrotSet:

Moving on the next level of understanding of this phenomena is what successful traders do.

They use models (however you want to define the term "model") to determine the probability of the current state continuing, taken their position(s) accordingly, and use judicious risk management to insure that they can continue to engage in the process even if their judgement is flawed and they are proven to be wrong.
More...
 
Quote from Corey:

Right, but Gil Blake is saying, if I am interpreting him correctly, that the 'unknownable surprise' is actually predictable, with high success, on certain days.
More...

Then I misunderstood you - sorry. I thought you were saying something about contemporaneous cross sectional correlation.
 
Quote from black diamond:

Then I misunderstood you - sorry. I thought you were saying something about contemporaneous cross sectional correlation.
More...

That's because I am. Gil Blake states that if you take a subset of stocks (typically related by industry) with a highly correlated, statistically significant price movement, that with high probability, the following day will have a movement in the same direction. In the short run, he claims, the subset will act "like a school of fish."
 
Quote from Corey:

That's because I am. Gil Blake states that if you take a subset of stocks (typically related by industry) with a highly correlated, statistically significant price movement, that with high probability, the following day will have a movement in the same direction. In the short run, he claims, the subset will act "like a school of fish."
More...

Then that is not contemporaneous cross sectional correlation, you have some lagged cross correlations in that story and that does violate the random walk assumptions. But it sounds like you know what you are talking about and we just disagreed about terminology.

That sounds very testable, I wonder if it holds up. If you ignore the industry relationship part, then that goes counter to what I would expect - small reversals of the biggest prior day moves from liquidity effects and/or overreaction correcting.
 
Quote from black diamond:

If I understand you this is possible. I'm probably too tired to explain this clearly but I'll try. The main feature of a random walk is that your expectation for a future price is the current price. You can have mean reversion or autocorrelation at a short frequency but a random walk at a longer frequency. If a price process is mean reverting at the daily frequency, then your expected value for tomorrow is somewhere between today's price and yesterdays. If it is a random walk at the 2 day frequency then your expected value for 2 days from now is the same as today.

I don't think anything prevents a price series from having both properties- tomorrow's price expected price is lower/higher than today's price, but the expected price 2 days from now is equal to todays price. I am not saying prices really work this way but I think it is possible to generate a time series that does. If this doesn't make sense I'll think harder about it later.
More...


What if the price moves randomly at all levels of frequency!
A modified version of the variance ratio test can be used to simultanously test the randomness at different frequencies. As a matter of fact, studies have shown that stocks' prices exhibit randomness at different levels.

As you just mentioned, price movements at short frequency levels show signs of mean reversion. However, this is just a delusion. This very short term mean reversion is simply a net result of the bid/ask spread. That's why intraday and one-day mean reversion is profound in illiquid and low price stocks.


Thumama..
 
Don't know about you but when I walk into a casino and see a craps table where the table is full to over flowing and whooping and hollering I want in. If the tables are desultory. I want no part of them. Hot tables are like an up stock. I don't give a damn whether it's random or not. It's going up.
As a swing trader my rules are simple.
" If it ain't going don't buy it.
If it ain't going down dont sell it.
Don't buck a strong market trend. "
If you want to add I don't care whether it's random or not. O.K.
I firmly believe that "How you exit a trade determines your profitability more than how you get in.".
Don't try to catch falling knives.
It's very old advice but I believe it still works, KEEP IT SIMPLE
 
You must log in or register to reply here.
Back
Top