Why do I see "Trends" in Randomly Generated Data?

[jack hershey]

Quote from MAESTRO:

Holly, Jack, you wrote a book! However, can you please tell me? Do you believe that trends exist in purely random generated time price series?

Beliefs about working with math (generators) could only work for market analysis if human psychology were part of the deal.

For me, the roughout would be to use price and volume and about seven fractals to have a composite.

For me the pace vs volatility of past periods serves to instruct me on the general picture.

I do not use induction nor do I get into the future any further than the hard data available or storable on feeds relative to the posted games being played by people who have the capital to "show".

My take on trading is that certainty is a requirement for all monitoring, analysis, decision making and action. Binary vectors achieve eliminating uncertainty in the present where all trading takes place.

For me, it turns out my mind has been "printed" for over 50 years of trading. Before that I was obliged to do systems analysis to get the better answers. I only view the market as a system which has a structure, process going on and results of the process.

I feel that everyone who has an open mind and is able to think critically can build their minds in total harmony with the markets.

I know this is a long answer but it is important for anyone reading what I say to have a context.

the modus of the market boils down to two alternatives that are not opposites. Therefore, a couple of independent variables are required to use mathematics to generate informative models.

qualified people have done terrific runs on my stuff and it has given them a great deal of confidence in the factors and values I use.

 

[MAESTRO]

Quote from jack hershey:

Beliefs about working with math (generators) could only work for market analysis if human psychology were part of the deal.

For me, the roughout would be to use price and volume and about seven fractals to have a composite.

For me the pace vs volatility of past periods serves to instruct me on the general picture.

I do not use induction nor do I get into the future any further than the hard data available or storable on feeds relative to the posted games being played by people who have the capital to "show".

My take on trading is that certainty is a requirement for all monitoring, analysis, decision making and action. Binary vectors achieve eliminating uncertainty in the present where all trading takes place.

For me, it turns out my mind has been "printed" for over 50 years of trading. Before that I was obliged to do systems analysis to get the better answers. I only view the market as a system which has a structure, process going on and results of the process.

I feel that everyone who has an open mind and is able to think critically can build their minds in total harmony with the markets.

I know this is a long answer but it is important for anyone reading what I say to have a context.

the modus of the market boils down to two alternatives that are not opposites. Therefore, a couple of independent variables are required to use mathematics to generate informative models.

qualified people have done terrific runs on my stuff and it has given them a great deal of confidence in the factors and values I use.

It was a simple question. Do you believe that trends exist in randomly generated price series? Yes or NO, that is all I need to know.

 

[Thunderdog]

This is indeed a very interesting thread.

 

[jack hershey]

Quote from MAESTRO:

It was a simple question. Do you believe that trends exist in randomly generated price series? Yes or NO, that is all I need to know.

On a scale of 1 to 10, they will yield (provide a quality of information) of about 4 to 6. This score is a yes for you and a no for me.

If a person puts up a good display, seeing price bars (generated or otherwise) is not required. to do much better than 6.

 

[5yrtrader]

I think this discussion is very interesting, but I didn't read every post so excuse me if I am restating something. Charts created from random number generators do seem to display trends, when in reality the previous up or down move gives no information on what number will be generated next. But there are two problems I see with comparing these charts to real markets.

1) You are stating that markets are random, when in fact they are not. Buying begets buying, as we have seen in the Nasdaq in the 90s and the housing market earlier this decade. Prices moving up draw more buyers to the market, which drives the prices higher.

2) As far as I know all trendfollowers say they make their profits in the "fat tails." Your random number generator assumes a normal distribution. It can't, by definition, have fat tails. The random number generator will never create charts like the ones below, follow the links. These would be trades where trendfollowers make their profits.

http://exchanges.barchart.com/intra/mgex/mgedmwh8.htm


http://finance.yahoo.com/q/bc?s=ABK&t=1y&l=on&z=m&q=l&c=

5yr

 

[dtrader98]

Quote from 5yrtrader:

1) You are stating that markets are random, when in fact they are not. Buying begets buying, as we have seen in the Nasdaq in the 90s and the housing market earlier this decade. Prices moving up draw more buyers to the market, which drives the prices higher.

Good Points. Ultimately, my argument here comes down to the notion that reading pure price charts does not offer much more predictability than say reading a randomly generated price chart and determining the future direction. One way to visualize this is to think of the following scenario; suppose you analyzed a random chart (not knowing whether it was random or real)-- if you had predicted the future direction properly based on your perception of a trend, you would also have proved to yourself, in retrospect, that you were fooled (hence the often quoted mandelbrot line, fooled by randomness).

There is no question that market behavior much like weather patterns, is NOT the resultant output of a completely random system (like say a coin toss). However, it is the difficulty in determination of the individual forces (each market participant) and the complexity of their interactions (huge untraceable feedback loops) that make the result of all their aggregate actions displayed on a price chart, for all intents and purposes appear much the same as a randomly generated chart would appear.

The best we can do (outside of having inside information showing where the aggregate supply/demand is originating) is to try to model the aggregate behavior at a higher level to have some level of confidence about it's future direction. And most mathematicians, have modelled it after the gaussian distribution, which in essence is random, but has some nice predictable properties (i.e. it is nicely bound by stdev limits) that allow one to develop some type of system to profit off this probabilistic model. Unfortunately, even this model can be thwarted by large unexpected events (outside of 5 sigma bands), similar to charts that suddenly create large gaps against chart readers expectations of a trend.

If you don't believe a model is necessary then it is not worth arguing the point, since no objective grounds can be made for either party to argue on. And the subjective realm is rife with those who believe they have the power to see things that others do not.

There are other ways to model markets outside of the present thread (i.e. chaotic vs deterministic or random), but I think they would add more confusion to the thread, and for all intents and purposes would not contradict the threads arguments much.

All that being said, as depressing as it may sound, remember that this entire thread is arguing about the comparison of pure market and price action (and its multitude of derivatives) to randomly generated price action.
There are other tools available to a small trader as the outsider however, that greatly enhance his opinion to make wise bets based on extremes (as some have mentioned), but again that is for another thread.

 

[Indrionas]

Quote from Rahula:


Now if the markets are random and the market closed up,up,up,up,up,up,up,up,up,up,up,up 12 days in a row - most people would likely be talking about it as a n-sigma event or a "black swan" even though it is just as unique and random an outcome as a up, up, down, up, down, down, down, up, down, up, down, down 12 day series of closes.

Direction is of normal distribution. The n-sigma events in the markets happen for range size, that's where the fat-tails are.

 

[MAESTRO]

Quote from 5yrtrader:

I think this discussion is very interesting, but I didn't read every post so excuse me if I am restating something. Charts created from random number generators do seem to display trends, when in reality the previous up or down move gives no information on what number will be generated next. But there are two problems I see with comparing these charts to real markets.

1) You are stating that markets are random, when in fact they are not. Buying begets buying, as we have seen in the Nasdaq in the 90s and the housing market earlier this decade. Prices moving up draw more buyers to the market, which drives the prices higher.

2) As far as I know all trendfollowers say they make their profits in the "fat tails." Your random number generator assumes a normal distribution. It can't, by definition, have fat tails. The random number generator will never create charts like the ones below, follow the links. These would be trades where trendfollowers make their profits.

You, guys really need to read “Fooled by Randomness: The Hidden Role of Chance in the Markets and in Life” - by Nassim Nicholas Taleb. It will put a lot of thing in perspective. As far as your two points, you have no idea how wrong you are. But before you argue more please read this book, it will help.

 

[ProfLogic]

Quote from MAESTRO:

You, guys really need to read “Fooled by Randomness: The Hidden Role of Chance in the Markets and in Life” - by Nassim Nicholas Taleb. It will put a lot of thing in perspective. As far as your two points, you have no idea how wrong you are. But before you argue more please read this book, it will help.

I don't know about Spyder but I just finished Taleb's last book and will be in the process of selling it on ebay. Not worth keeping it for sure.

I specifically researched all major random events to effect the markets in the last 120 years, with all of the available data I could muster up and found that each one was pre-empted by market action to telegraph a reversal in price direction or negated the price reversal of the market action immediately following the event sending the market back into the trend it came from . . . 100% of the time.

I repeat . . . if YOU can not define the environment you are trading in then every action YOU take is completely random but not ALL of us trade with YOUR blinders on.

 

[Thunderdog]

I read Taleb's FBR when it first came out and thought it was a great book. And yet, I continue to believe in expoitable price trends. More specifically, I believe in exploitable price action that develops into a trend of indeterminate duration with reasonable regularity. Don't you just wish that you were on the opposite side of all my trades?