Predicting randomness

Dr. Alexander Elder's books are good arguments that technical analysis works and markets are not random. Let's not confuse the random noise with the market randomness.

Comparison with gambling should be limited only. I'm not aware of any technical analysis that works with gambling.

On the other hand money management (position sizing) works well both in the markets and in gambling.
 
Quote from kjkent1:

Although I do not find that the markets are random, I cannot agree with your proffered proof. The autocorrelation formula that you provide is just another name for what is commonly known in mathematics as the linear coefficient correlation function. The mathematical "proof" of this function, depends on the axiom that "All input data points are located between the data endpoints." To state it simply, the mathematical proof specifically denies the possibility that any regression function can be used, by itself, to predict future behavior.

The function merely shows how non random the observed and input data is, as compared to completely random input data.

My contention is that the markets are non random because they are manipulated by people with money, and that if you "follow the money," you can do better than the average person.

However, no one who routinely depends upon a regression function can make more money in the long run, than the average buy and hold investor. If someone "is" making money using such regression functions, then they are adding something else into the mix that is permitting them to defeat what would otherwise be a breakeven proposition.

KJKent,

Yes, it is just a linear correlation function, but the main point was that this function shows that price at t depends on price at t-1. Price t could be the price 5 min ago and t-1 is price 10 min ago. The fact that there is positive autocorrelation by definition shows that the prices are not a stochastic process. I guess i used the wrong terminology - I am not saying you can predict future prices by regression., ie price t+1 by using price t. I am saying that past prices are dependant on each other.

The million dollar (maybe billion) question is how to project the price :)
 
Quote from nicholaf:

KJKent,

Yes, it is just a linear correlation function, but the main point was that this function shows that price at t depends on price at t-1. Price t could be the price 5 min ago and t-1 is price 10 min ago. The fact that there is positive autocorrelation by definition shows that the prices are not a stochastic process. I guess i used the wrong terminology - I am not saying you can predict future prices by regression., ie price t+1 by using price t. I am saying that past prices are dependant on each other.

The million dollar (maybe billion) question is how to project the price :)

That's easy. Just ask the marketmaker/specialist. LOL!
 
Quote from kjkent1:

I posted my suggestion of the "game" as a means of demonstrating how unprepared most people are to trade the markets. People paper trade, thinking that this gives them experience. Paper trading doesn't work, because the market doesn't respond to their moves, so there's no feedback to tell the trader how he's actually doing.

My game, however, gives immediate feedback -- there's $1,000 at stake and someone can win it all, if they figure out the right strategy.

Of course, in the real market, you can't win it all, but the strategies are the same. And if a technical indicator won't work within a closed universe game of 10 people, it damn certain ain't gonna work in an open universe with millions of participants.

Call me lazy, i'll call myself too dumb. Would you mind to give the solution to this puzzle you brought up an odd 20 pages back? And the implication of the outcome as well please.
 
Quote from trade4succes:

Call me lazy, i'll call myself too dumb. Would you mind to give the solution to this puzzle you brought up an odd 20 pages back? And the implication of the outcome as well please.

Please PM me.
 
Quote from mogul:

on that note, does randomness even exist? Or are events merely classified as random based on our level of perception.

A computer generating a random number is not random, but for all intents and purposes it can be used as such

Now you are talking!
 
Quote from mogul:

on that note, does randomness even exist? Or are events merely classified as random based on our level of perception.

A computer generating a random number is not random, but for all intents and purposes it can be used as such

There's a difference between "random" and "unpredictable." A truly random event cannot be predicted, even if the observer is able to accurately measure every causal action operating upon the potential future effect.

A merely chaotic event is potentially predictable, because it is the product of completely measurable causal factors.

At the most fundamental level, uncertaintly in the universe is still an open question. The majority of theoretical physicists subscribe to the "Copenhagen Interpretation" of uncertainty, which holds that it is indeed impossible to simultaneously meausre both the speed and location of an electron, and therefore true randomness exists in the universe.

The minority, however, believes that there are other mathematically plausable explanations (e.g., "many-worlds uncertainty interpretation") that would permit simultaneous measurement, and that we simply arent yet able to implement the math with any physical technology. If true, then all things are measurable, and randomness is merely the product of insufficient data.

The above represents the "naturalistic" explanations for randomness and uncertainty. There is, of course, the "supernatural" explanation, which would be basically, that God created the universe, and therefore randomness exists or doesn't exist, depending upon God's requirements at any given moment.

If you accept this last theory, then you don't need any technical or fundamental tools to make money in the market. You need only prayer.
 
Back
Top