Enough already! It's not random

I'd be hard pressed to find anyone (academic or otherwise) who actually believes in strong, classic random walk. Sure, we use iid random walk for some derivatives pricing for matter of analytical simplicity, but I don't think anyone thinks its right. Sometimes it just doesn't matter that much.

 

Ooooh boy.....Here we go again.

Let's all get on a level we can understand. Cut out the brainy boy talk and ask yourself a few questions.

For all you who think the market is not random, why do you have trades that lose?

Are you losing on purpose? Noooo.

Are you just not smart enough to figure it out? Nooooo.

Then what? The market is r_n_o_; right? (Go ahead; be brave; fill in the blanks.)

Some seem to think that if the market is random, that one can't make money. No, no, no.....This isn't true either.

Watch this:

A Ford worker goes to work everyday for 30 years installing the steering wheel on F150s. He gets it right every time because what he is doing is not random. He knows the exact outcome of his efforts.

A trader trades everyday for "x" number of years. However, to save his ass, he just can't figure out what works on every trade, even though he sees, feels, and does the exact thing he has done many times before. Yet, he still experiences losing trades. That is random.

Now if you can tell me truthfully that you can trade while knowing with certainty that you will have a successful effort every time, like the worker at Ford, then I will have to concede that the market is not random. I'm not holding my breath.

It's amazing how many of you want to believe that the market is not random, yet you lose, lose, lose.

The market is about probabilities and money management; not CERTAINTY.

Remember---

1. Trading is simple.

2. The market is random.

3. Everything works some of the time.

4. Nothing works all of the time. (And there is a reason why.)

 

OK, I understand. It was not really clear from your first post that you were doing a Wald-Wolfowitz test. Then I do agree with your results. It is the fact to find 1332 "runs" that is 3 sigmas and not the difference between up and down events. My fault, I misread you.

However, the problem with your approach is that it's not practically usable.

One example. Today is up. Then, from your study you would say that tomorrow would tend to be up too, but actually not, because your study says that there is a bias that may last more than just 2 days. I don't know if you follow me on that one.

Conditional probabilities are far more efficient and practical to work with. Then you know your odds (based on the past of course) of seeing precise events, not only saying that "as a whole" the market satisfies a statistical test or not.

 

You are correct. Detrending as a form of data normalization does make the price distribution “more” Gaussian. However, as you pointed out quite well, the problem with “fat tails” still remains even after the normalization is done. Of course if you still have “fat tails” the construction of a useful trading algorithm remains a challenge. However if it would be possible to do the normalization in real time (without using historical ATR or other things of this nature) as such that the result of it is a pure Gaussian distribution that would open an opportunity to expect a stable probabilities distribution with “no surprises” (or fat tails). That was the challenge that I addressed 5 years ago. In other words, my task was to find a real-time function that makes price variations normalization process as effective as possible (the result of the normalization in this case would be as close as possible to Gaussian with no Fat tails). The problem was solved by implementing the function that we called VQU that is based not just on price fluctuations (like ATR or similar functions) but also on the number of unique price changes in the unit of time. This number was also weighted by the buying power of the dollar and the value of the minimum price fluctuation unit. Applying this function has resulted in the distribution curve that did not have any “fat tails” and it was 99% Gaussian. As you can imagine constructing the trading algorithm based on that was quit easy.
Cheers.

 

confusion reigns ...

weather is often cited as the most random of events. the weather man has predicted sunny and higher temperatures every day this week. Is he likely to be right or wrong? it really depends on the confidence levels of his input. Often times he is throwing darts at a board, sometimes, he is 80-90% confident.

Three days running I've been basking in 30-40 degree sunny days.

So, does that render weather non-chaotic and predictable?

It seems to me that it doesn't really matter, the focus should be on those high-confidence days.

 

It's amazing that although you believe the market to be random, the chart contained in your paper written 5 years ago seems to predict the price of the S&P 500 with remarkable accuracy up until the middle of 2007.

(sorry, I couldn't resist)

 

Good catch, BigFunky!

Numbers aren't Maestro's strong suit. I haven't read this paper but his paper on zero-crossings was rife with math errors.

Have any of his papers been published? Peer reviewed? Where did he take his degree, if any? And in what field?

 

While we are on this subject, we might as well get into the same old fundamental analysis versus technical analysis thing too, or any of the other typical go-nowhere discussions... everything you're going to say about this topic has already been said by some other trader, in some other place, at some other time, and its all going to be said again.

wait wait... oh my god its a TREND! somebody shoot me.

 

I was wondering about the same.

 

Why is there so much opposition against Technical Analysis? Is it not a science? After all, such popular theories like the DOW (Day of Week), Candle Sticks, Elliott wave theory, Trend line, Relative Strength Index (RSI) , Stochastic etc. are all associated with Technical Analysis.

Is it so, because it is a science just like many do not like maths, but yet it is important.