[Bodhi Dharma Zen]

Originally Posted by uexkuell »

"Trend" is an artificial concept.
That is the reason why it cannot be quantified.

I believe that trends are real, just impossible to spot BEFORE they happen. I mean, in an historical chart it is obvious when some instrument goes up or down in a mostly regular fashion. But spotting it before it happens is the tricky part.

My stance? You can't. No one can.

If I happen to believe that some price will go up and then it goes up I call my belief a prediction, and thus I confirm that "I know". But with the same eagerness that we remember the times we score we forget the times our "predictions" didn't work.

Originally Posted by uexkuell »

With a pretty simple experiment this can be proven:
Enter e.g. EUR/USD with 1,000,000 EUR or USD. You will get immediately the feeling that there is a "trend" because the account goes up or down numbers that will probably seem significant.

With other words: Trend is a function of account size, profit expectancy and risk tolerance.

If huge leverage is used there is no trend-less phase, "chop" or whatever people call it.

Ok, I do not understand where are you coming from in here. Care to explain it a bit more?

Originally Posted by uexkuell »

When strategies seem not to work in "choppy" times this is due to some other problem - probably bad entry point.

If the entry point is chosen correctly it's always possible to choose one of:
- increase leverage
- increase time frame = wait until the symbol moves far enough

But can we correctly "chose" it? Or is it more a matter of "luck"?

[Bodhi Dharma Zen]

Originally Posted by sevensa »

I don't see how the market can be a fuction of account size, profit expectancy and risk tolerance as the market has no "knowledge" about any of those. An indivudual leverage has zero impact about the trendiness of the market.

Exactly what I think. Maybe he is talking about something else??

[Bodhi Dharma Zen]

Originally Posted by sevensa »

How do you know that statistically it doesn't mean anything? I would be interested to know what kind of testing you did to reach your conclusion?

Simple. I followed several graphs with different intervals. I couldn't spot a pattern at sight, so I didn't have to write a system to see it (I use tradestation BTW).

But let's not forget, maybe I did something wrong.

[Bodhi Dharma Zen]

Originally Posted by ckait »

The definition of trend is making higher highs and higher lows. The problem most people have is they dont understand what time frame they are trading. And it varys from time frame to time frame. A 1 min trend may only last 5 minutes or it could last 20. has nothing to do with account values.

You are right, still, what we are trying to figure out is not that if trends exists or not.. but if we can SPOT THEM before they happen.

[ckait]

Markets can act erratically, but they also do the same thing alot of the time. I just know there are areas whrer i expect to get a reaction. If I see the reaction im looking for i'll take the trade. The profile provides the structure and the devopement druing the day in its relationship with the previous day or days gives me my likely trade location. I know some people break down the day into its 30 min periods I definately look at them but like the daily profiles better.

[Bodhi Dharma Zen]

ckait,

Yes, I know what you mean. Sometimes it appears that some behaviors tend to repeat themselves. What I doubt is that this could be coded in a mechanical system. Now, what I want to say with this is that if you develop a system that detects those areas where you expect a reaction (say a pattern of momentum + volume + previous day activity) I'm almost certain that you will get a system that will not do better than chance.

This is because we all have a hardwired "confirmation bias" system going on at all times, we believe we recognize a pattern and then act by it, if it doesn't work we make excuses, and if it works we confirm our belief.

This is (part of) the reason I have developed a mechanical system, I don't want my emotions to dictate my actions on the market. Still, it is tempting to ride it manually...

[ckait]

Thats cool, Ive tried several times to come up with a machanical system and have a few set ups to test but i cant find any one to program them. I have no skills in that area. But i do have defined trade criteria If someone is up to the task to take on programming and backtesting the set ups. If nothing else I know my set ups would give a fresh out look on a few things and possible something that could be automated. Thinking of getting my wife to do it, she used to be an electrical engineer and could probably muddle her way through it but Id rather have a traders prospective.