do markets trend more then range?

Hi all,
I wonder if markets tend more often to trend or more often to range? I wonder since I am developing a trend based system and while testing last 30 days I am receiving strange results I did not expect - they do not relate much to profit or loss but rather to the shapes of profit curves - (one month automatic trading system forward test on 30 currencies based on Gain Capital data, the same settings for all pairs) that make me very very confused, I really am not sure what to think about them.

It is my intent not to disclose whether I am in plus or minus per month because I need an unbiased answer from some rookie, some experienced trader.

I have been told that markets tend to trend, so I developed a trend trading system. I would really need an advice on how to measure if markets trended or not - how should I measure it? Is it even measurable? Are there any statistical/mathematical models that can measure it? Should I consider parameters of my system when considering general trend/range measurement of a specific pair? How should I consider timeframe when thinking about that? Can one say that it is trending on 15M timeframe while ranging on say 60M?

Can one think that if the system works fine on some timeframe and not on another that that is ok? Or that it works on TF between 8-20 minutes and not higher on some pairs while other need much higher timeframe in tens of minutes? Some even hours to make profit with this system? Does that mean that the market is not trending on higher, respectively smaller timeframes? Is it a smart idea to reverse rules for pairs that lose a lot on smaller timeframes? while using reverse version on higher timeframe?

It has been programmed that it can be run/tested on any timeframe and any asset so there is no fitting to curve on any pair. Everything is calculated purely from the closing and median price, no news, it trades over weekends, all the time, it changes long/short positions all the time. Maybe some optimization might improve it. But I am very reluctant to do any optimization since I fear of overfitting to current market conditions.

Do you have answer to any of my questions?
Any advice or hint really appreciated in advance
N.

As far as i know, the market range 75% of the time. I think in one of your question you already see the problem of any system:

In the eyes of a swing trader, maybe a market is in a range (Several hundred pips range) but in the eyes of a short term trader, the market is trending big time and during this trending period you will have periods of range. (fractal structure of the market)

I think a better system will be to identified market turning points. For any market, whether it is in a range or trending, the way it turns is the same.

I have studied this subject but from a different stand point. If you look at this thread http://www.forexfactory.com/showthread.php?t=171552 you will see the beginning of my studies. Since then I have refined the findings and I think, judging by your post, you will be able to see what I was looking for and how it might help you.

I recently also having similar question my self, and later on I found out that calculating directional movements maybe is the solution. So I spelled out several codes and testing and refining it. So far it look promising..

I tried to answer these exact same questions for years, and all it did was become a waste of time. Programming EAs over short time frames is basically trading the noise fluctuations, very unpredictable.

Instead you should be looking into WHAT and WHY a market trends or ranges. Then you can start to "see" trends before they happen and get in at market swings where you have the least risk and maximum rewards, this is where the real profits are made.

-Razorjack

To be honest, that answer is almost impossible to answer. The markets are fractal, so what you see as a range on one time frame is probably just a retracement of a trend in another time frame.

But if you wanted to quote a general rule of thumb - markets trend 30% of the time and they are range bound 70% of the time. Meaning if you are trading a trend following strategy, expect to only win about 1/3 of your trades and if you are trading a range-bound strategy, expect a slightly higher number of winning trades!

I would almost bet that an 80/20 rule is imbedded in this thing somehow. 80% of a person's profits can be made 20% of the time when the market is trending. The other 80% of the time when market is rangebound is when most traders fail.

Wouldn't it be profitable to trade a range bound market? The edges of the range are known so picking up the retracements should not be too hard, right?

http://i49.tinypic.com/28threq.gif

I have heard people saying that markets trend only about 15% of the time, others saying they trend around 25% of the time etc... So it depends on how you define a 'trend'. If you define it as a strong ADX reading of X or above then perhaps they do only trend 15% of the time, if you have slightly less strict criteria for how you define 'trend' perhaps it's more like 25% of the time.

Personally, I'd say there is an up trend if the close is higher than it closed 8 days ago AND higher than it closed 20 days ago AND higher than it closed 120 days ago, and vice versa for shorts. Using this as a definition of a trend, pairs like the EUR/USD and GBP/USD trend about 47% of the time time.

Trend following is the easiest way I know of to make money.

A range on the weekly chart can be many trends on the hourly. Price is a lagging indicator anyway, a range can turn to a trend and a trend can reverse, price is less lagging than the other calculated indicators but lagging non the less.

It depends. Late 2008 was very trending, today's markets are really ranging. I'm a 1m trader and I can only find 2-3 "real" trends per month. That's 10% of the time.
And given that the markets are fractal.........

Yeah, problem is by the time you see the established range, and then set you limit orders a breakout happens. Basing you trading on patterns in historical price for me is gambiling. There are patterns that are tradeable but you need to understand what is going on behind the pattern.

This is the way I think of it. The markets are like many dynamic systems which seek a steady state. In the case of the markets that would be when buyers and sellers agree on the price and it never changes. Of course that state is rarely achieved. But when the system is knocked off kilter, it will tend to try to finding the steady state. Trending happens when something new is introduced which causes prices to change suddenly in one direction. When prices approach the next level, they will tend to oscilate or range for a while until they reach the steady state or more likely they get knocked off kilter again. I think based on this model, it is obvious that trending will last for a shorter period of time than ranging.

Hello guys,
thank you very much for your posts. I am getting general idea. Based on my results I was afraid that trends are not as common as ranges. Folowing is my discussion. For you not having much time, my questions are simply put:

-Sticking to timeframe where it makes money is a good idea? Wouldn't that mean that system is not so robust when it does not work on other timeframes as well or even goes negative?
-Sticking to part of day when the pairs are heavily traded (and trends are more probable due to news announcements) it makes more money, isn't doing so basicly curve fitting?

My discussion...
Do you think that it would be ok, that I would only stick to that timeframe on which it made maximum of money?
Like EURGBP makes me plenty of money on 10M, GBPAUD on 15M, some other even on 12 or 24. Vast majority of results per pair are like from Gauss distribution. I mean when I say GBPAUD is best for me for 15M, then 10M and 20M (and usually even 12M/18M) return second and third best results, that is based on one month of testing. on GBPAUD it gave me like 70-80 trades per month, EURGBP on 10M 105 trades. On 30M it makes like 30-40 trades per month per pair.

If I limit trades only to time when pair is heavily traded I improve average profit per trade in multiples, like per EURGBP from few pips into two digit levels, GBPAUD average profit increased by 1/2. You advised me to stick to times when there are news announcements that can move the market - to create a trend. I guess that is the point.

So I guess I should just not care if markets trend or range and only test the system, if it makes money, I do not need that information I was looking for, right? or am I totally missing something?

Thank you guys, I really appreciate your advices, when I get experienced a little I should come here more often and give advice to some less skilled traders to repay the comunity...
N.

The market is always trending....

And always ranging

When I started many years ago and paid for an excessively expensive stock trading course containing less info than you can find on the free babypips site, I was taught the markets range 70% of the time, technically this maybe correct but pan out and it is a pause in a up or down trend.

As you point out it looks different on different time frames, trending on larger time frames ie. daily, hourly much easier than on 5m for example. Personally I trend trade my signal is a 300 sma price cross on hr chart. Do I occasionally get burned by it ranging? Sure but not so often that it's a problem long term.