First sacred cow: "If the trend started, it will continue" - page 8
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
Here's my point. Look at chart 1:
Do you see trends here? Do you see technical analysis, resistance lines, support, Elliott waves? Personally, I don't see anything here . These are random walk charts on an accumulative basis. The trends that appear here cannot be traded, neither their beginning nor their end. All this is a complete randomness.
Now look at chart number 2:
The chart shows the same trend and the same trend with a 5% positive bias. The difference between one and the other is obvious. So, the way I see it. a trend is the difference between a random walk and a market shift. Simply put, it's a spread between the blue line and the red line.
There are such shifts in real markets. They are very weak, usually 2-3%. It is practically impossible to make money on these shifts because you have to feed your broker's commissions. But it is easy to make a break-even system, and it is even possible to use standard indicators.
There are trend shifts and counter-trend shifts. Trend shifts usually occur on very long timeframes and they increase proportionally to the directional price movement. The counter-trend shifts appear on much smaller timeframes. The uniqueness of the market is that it includes two opposites and allows you to earn on both of these properties at once!
Here is an example, a simple system which uses counter-trends:
And here is the system using the trend:
Although both systems are preeminent to the extreme, they technically beat the market. The question is, how can both systems end a long testing period in profit if one of them uses pullbacks and the other uses trends? This question requires additional research, but most likely it is connected with different phases of the market (trend-flat-backward), but all these phases appear in different time scales. I came to the conclusion that in the short term the market is a pullback, and in the long term the market is trending.
Now a word or two about the predictability of a trend start. Here's an axiom in logic that traders should pay attention to: you can't predict the existence of object A. You can't predict long-term price movement based on price. You need other data. We would be much more effective at predicting future price movement based on not price indicators.
It is better to use non-price indicators, they are much more helpful.
Vita, thank you.
"In terms of profit, trends don't exist": well, what about the very experienced traders who coined the saying? For them a trend exists and they even make profit from it...
2 S: You've gone too far with the aces. They have different statistics, not like the market rows.
Here's where I wouldn't trust the legends, and would have checked the realities. What experienced trader's definition of a trend comes to your mind? Is there any profit in it? Experienced trader's tale + pictures of success != trend definition + profit on it. Generally speaking, experienced traders with profits are not proof of the existence of trends in terms of profits.
Then who are these most experienced traders? Aren't they the ones who got lucky at the lottery in a certain period? Anyway, that's the tip of the distribution iceberg. You know it yourself, after all.
How can one come up with such a saying? Easy. Those who are lucky enough to "hit" what can then be called a trend many times in a row will boldly say trend is your friend and if it started, it is likely to continue. Those who were not lucky enough to fall into a trend will boldly say not to trade against it, because their stops were blown off by a reversed, unguessed, "trend". The rule about chopping losses, letting profits grow is also invented. If witches existed, they would fly on brooms. If trends existed, they would continue, once they started.
This is where I wouldn't trust legends, but rather scour the realities. What experienced trader's definition of a trend comes to your mind? Is there profit in it? Experienced trader's tale + pictures of success != trend definition + profit on it. Generally speaking, experienced traders with profits are not proof of the existence of trends in terms of profits.
Then who are these most experienced traders? Are not those who were lucky at the lottery in a certain period of time? Anyway, that's the tip of the distribution iceberg. You know it yourself.
How can one come up with such a saying? Easy. Those who are lucky enough to "hit" what can then be called a trend many times in a row will boldly say trend is your friend and if it has started, it is likely to continue. Those who were not lucky enough to fall into a trend will boldly say don't trade against it, because their stops were blown off by a reversed, unguessed, "trend". The rule about chopping losses, letting profits grow is also invented. If witches existed, they would fly on brooms. If trends existed, they would continue, once they started.
+100
I would like to add. I'm going to draw a zigzag curve by hand. I should draw it in such a way, that support-resistance levels would be visible and classical figures would be visible. I'll show it to the trend experts and invite the eliotchiki-fibonauts. Then I will say: "Gentlemen! This is an indicator of the zigzag of such a trading instrument. Please continue the line." And after analyzing it, they will continue! They don't even suspect that the zigzag line is a figment of my imagination. And I calmly, quietly, continue the line in the other direction! What then? Why can't they ever guess where my broken line will go next? Because they know nothing about my motives and motivations, just as they know nothing about the true (all in the aggregate) forces driving quotes. The market price of currencies, and indeed of all things tangible and intangible, is a figment of human imagination. A person cannot know what he will be thinking in the next moment, much less what a crowd of millions will be thinking. After all, that is what it takes - "to know" - to determine the beginning and the end, again, of a man-made "trend".
+100
I'd like to add. I will draw a zigzag curve by hand. I will draw it in such a way that the support-resistance levels would be visible, and the classic figures would be visible. I'll show it to the trend experts and invite the eliotchiki-fibonauts. Then I will say: "Gentlemen! This is an indicator of the zigzag of such a trading instrument. Please continue the line." And after analyzing it, they will continue! They don't even suspect that the zigzag line is a figment of my imagination. And I calmly, quietly, continue the line in the other direction! What then? Why can't they ever guess where my broken line will go next? Because they know nothing about my motives and motivations, just as they know nothing about the true (all in the aggregate) forces driving quotes. The market price of currencies, and indeed of all things tangible and intangible, is a figment of human imagination. A person cannot know what he will be thinking in the next moment, much less what a crowd of millions will be thinking. After all, that is what it takes - "to know" - to determine the beginning and end, again, of a humanly invented "trend".
I am FOR such a position.
variant1
An upward trend is a sustained change in price due to an excess of purchase bids over sales bids during a period of conditionally even bids
a downward trend is ...
.
option2
An up trend is a combination of elementary particles of a chart (2 candlesticks), satisfying the condition of keeping the maximums and minimums of each elementary particle above the previous one within the combination itself
a downward trend is ...
.
Variant3an up-trend is the price movement from the support level to the resistance level (the levels are last in time)
downward trend is ...
.
option4
a trend is a conditionally equal sequence of support/resistance levels formation
.
Options 2,3,4 ... are consequences of Option 1As I see it, a trend is a correspondence to a model being tested, no matter what model it is (it's a very crude definition, but I'm not going to spend a sunny Saturday afternoon perfecting stupid definitions until late evening :o). If one takes a "growth" model as a trend, no matter what it is based on - then objectively there are no trends, but if one really wants to, one can, as they say, consider a trend only in a scaling aspect, as an element of a multifractal (by the way, price series are not self-similar at all).
There is a book "Applied Mathematical Statistics" by Kobzar. The whole section is devoted to trend and randomness criteria, starting with Abbe-Linnick criterion and so on (though not all of them can be applied to such series). But nevertheless my searches led to the following - there are no trends for price series as understood by speculators, but there is an illusion. "Life" confirmation is very simple - there are no "sustainable" champions among us, there are local victories :o) But that's just so, a philosophical aside.
But, the most interesting one, given in the book, is a trend criterion based on autocorrelation, which I have modified a little, added time series memory, tweaked here and there and used. This is what an example of "trend" search looks like:
The analyzed time series in the picture is the price itself:
The winner is the part of the series with the maximum AR model probability estimate. Recall, but the trend in this case (in my case) is the fit of the AR model. The market model itself is in the picture (already cited):
It is essentially "oscillations" around local average levels with jumping (micro disasters) to new levels. By classification, my robot is a very, very, very simplified self-organising stochastic system with a fixed structure.
Here's a test run: https://forum.mql4.com/ru/20423/page20. Right now I'm doing some re-surveying, gathering statistics and thinking about one problem. The system is only theoretically profitable, the calculation on one bar takes about 15-20 minutes. I cannot test it on the history, so I am thinking how to simplify/optimize it.
PS: ADDENDUM
PS: by the way, the picture is 4 years old, names need to be corrected
Eh...
A trend is not something specific, like - this is a trend and this is not.
A trend is a TENDENCY. It's a peroblastic wind rose, it's a current in the sea, it's a growing population of guinea pigs on Chucca Caca.
The trend lasts for months.
What you (expletive) call trends is not that. :)
It's waves or something, whatever you want to call it.
The trend is wonderful because no one can change it.
And what you're discussing is dealing cards in a game of pref. :)
As I see it, a trend is a correspondence to a model being tested, no matter what model it is (it's a very crude definition, but I'm not going to spend a sunny Saturday afternoon perfecting stupid definitions until late evening :o). If a trend is a "growth" model, no matter what it is based on - then objectively there are no trends, but if one really wants to, one can, as they say, consider a trend only in the scaling aspect, as an element of a multifractal (by the way, price series are not self-similar at all).
.....
It is essentially a "fluctuation" around local average levels with a jump (micro disasters) to new levels. By classification, my robot is a very, very, very simplified self-organising stochastic system with a fixed structure.
Here's a test run: https://forum.mql4.com/ru/20423/page20. Right now I'm doing some re-surveying, gathering statistics and thinking about one problem. The system is only theoretically profitable, the calculation on one bar takes about 15-20 minutes. I have no opportunity to test it on the history and now I think, how to make it simpler/optimize.
trend is a consequence
let's start with the cause
trend is a consequence.
>> let's start with the cause.
From the cause? You mean from the emergence of the first mind on earth bent on enrichment and hoarding?
There is a book called Applied Mathematical Statistics, by Kobzar. There is a whole section devoted to trend and randomness criteria, starting from Abbe-Linnick criterion and further in the text (though not all of them can be applied to such series). But nevertheless my searches led to the following - there are no trends for price series as understood by speculators, but there is an illusion. "Life" confirmation is very simple - there are no "sustainable" champions among us, there are local victories :o) But that's just a philosophical aside.
....
Here's the test samples: https://forum.mql4.com/ru/20423/page20. Right now I'm doing a little revision, collecting statistics and thinking about one problem. The system is theoretically profitable, calculation on one bar takes 15-20 minutes. I cannot test it on history, so I am thinking how to simplify/optimize it.
How do you combine two thoughts in bold? On the one hand it sounds like there is no profit in trends. On the other, profits are there. In trends? In some other movements from level to level?