Impulse - page 46

 

Gentlemen! Your main mistake lies in the original model. In the word "momentum". What makes you think there should be an analogy in price with the movement of a solid physical body?

If you recall the original source where the word came from, it is wave theory. And there the word is used in tandem with the word "correction". The mistake most beginners make is to repeat the phenomenon discussed throughout the thread. The answer lies on the surface. Find the meaning of the word correction in physics. You won't find it. Because there is no such thing there. The couple of words "correction" and "impulse" are not taken from physics, but from the same field, which is psychology. "Correction" behaviour and "impulse" behaviour.

A correction is a type of price movement where permanent deep reversals are used. It is not necessarily a "sideways" move. It can often be a trend. In wave theory, from which these concepts are derived, the type of corrective trend is non-standard corrective patterns with small x-waves. For this reason, many people combine the concepts of "correction" and "pullback", which is generally speaking incorrect. Has it ever occurred to you that if you already have a couple of definitions of trend - pullback, you don't need a second pair of impulse-correction at all? The pair "impulse-correction" is not intended to describe a chart shift in some direction or just stand in one range, but to describe how this shift takes place.

An impulse, on the other hand, is a unique type of trend. This trend has very shallow and gentle pullbacks. The appearance of this chart easily convinces those who see it to move quickly and correctly.

In a corrective trend, pullbacks are frequent and deep. Each reversal of the price movement scares the trader and makes him doubt the correct prediction, but it can be a big price move.

If we speak about ticks. An impulse is most often ticks in one direction without alternation. Moreover, even if they alternate, ticks in one direction are much larger than ticks in the opposite one.

In a correction-trend, very often ticks in both directions alternate, and the difference between the size of ticks in opposite directions is small.

There is one more purely practical difficulty. Impulses very often run away very quickly after the start. If you build an indicator that even registers the beginning of the impulse, you can easily miss the signal - it becomes irrelevant too quickly. You need alerts and push messages to your smartphone.

 
Dmitriy Bidulya:

Gentlemen! Your main mistake lies in the original model. In the word "momentum". What makes you think there should be an analogy in price with the movement of a solid physical body?

If you recall the original source where the word came from, it is wave theory. And there the word is used in tandem with the word "correction". The mistake most beginners make is to repeat the phenomenon discussed throughout the thread. The answer lies on the surface. Find the meaning of the word correction in physics. You won't find it. Because there is no such thing there. The couple of words "correction" and "impulse" are not taken from physics, but from the same field, which is psychology. "Correction" behaviour and "impulse" behaviour.

A correction is a type of price movement where permanent deep reversals are used. It is not necessarily a "sideways" move. It can often be a trend. In wave theory, from which these concepts are derived, the type of corrective trend is non-standard corrective patterns with small x-waves. For this reason, many people combine the concepts of "correction" and "pullback", which is generally speaking incorrect. Has it ever occurred to you that if you already have a couple of definitions of trend - pullback, you don't need a second pair of impulse-correction at all? The pair "impulse-correction" is not intended to describe a chart shift in some direction or just stand in one range, but to describe how this shift takes place.

An impulse, on the other hand, is a unique type of trend. This trend has very shallow and gentle pullbacks. The appearance of this chart easily convinces those who see it to move quickly and correctly.

In a corrective trend, pullbacks are frequent and deep. Each reversal of the price movement scares the trader and makes him doubt the correct prediction, but it can be a big price move.

If we speak about ticks. An impulse is most often ticks in one direction without alternation. Moreover, even if they alternate, ticks in one direction are much larger than ticks in the opposite one.

In a correction-trend, very often ticks in both directions alternate, and the difference between the size of ticks in opposite directions is small.

There is one more purely practical difficulty. Impulses very often run away very quickly after the start. If you build an indicator that even registers the beginning of the impulse, you can easily miss the signal - it becomes irrelevant too quickly. You need alerts and push messages to your smartphone.

An indicator? First we have to determine the algorithm of this phenomenon.
 

The arrows indicate the bars with the "characteristics" of the impulse = volume + spread, and they usually overlap, probably we should include some other variables in the analysis, i.e., the "limits" (levels, MA???) in relation to which "borders" (levels, MA??) the impulse can be used.

 
Veniamin Skrepkov:

The arrows indicate the bars with the "characteristics" of the impulse = volume + spread, and they usually overlap, probably we should include some other variables in the analysis, i.e., the "boundaries" (levels, MA???) around which an impulse can be used.

What is the algorithm for calculation of the arrows? Is it calculated on every tick or is it calculated on the previous bar (i.e., when a new bar appears)? Is the lower histogram repeatedly redrawn on the zero bar?
 
Karputov Vladimir:
What is the algorithm for calculating where to draw the arrows? Is it calculated on every tick or is it calculated on the previous bar (i.e. when a new bar appears)? Is the lower histogram repeatedly redrawn on the zero bar?
On the indicator linear_Price_bar, I put the spread level = 7 pips (the pair is one of the most volatile), on the Volume 120-180, the calculations are actually not there yet (I observe and conclusions are still ambiguous))))
 
Ruslan Kuchma:
  1. What is momentum in relation to Forex?

An impulse in my understanding is a unidirectional movement of the exchange rate of a currency pair per unit time. Naturally, the smallest unit of momentum is one tick. Therefore, most of the traders' developments lie in the plane of studying the structure of ticks in a minute, five minutes, or non-standard timeframes: two minutes, etc.


2. Is it possible to calculate momentum for a certain period of time without averaging the results over past periods of time?


Unfortunately, to automate the process of working with impulse, in any case, it is necessary to take the previous data as the basis of counting. But it all depends on the approach.


3. A strategy for working with momentum.


Personally, I use the probabilistic approach to determine the beginning of an impulse. Unfortunately, so far I haven't managed to find mathematical consistency in the structure of ticks on M1, so I have to use rough methods of observation. For example, impulse is highly probable after the price comes out of flat, etc.


I would be happy to read other insights from practicing traders on this topic.

I agree with the fact that "impulse is highly probable after the price comes out of the flat, etc.", i.e. one of the variants of "work on the border (corridor) of the flat". I think we should expand the boundaries of the "impulse", it's difficult to analyze it in M1. Probably it is difficult to make a classification on М1. Yes the classification as M1 forms M5 , and M5 forms M15 ? "In one case it may be a "trap" with preliminary buyers (preliminary bayers) or a target level when the fixing or buying is done, in one case the escape from the flat may be short (to release from the trap not to prevent the "averaging"), in the other case if the market performs its operations, the escape from the flat may be much stronger. Ideally, the indicator (Expert Advisor) should level out these "understandings" of the market that come out of the assumption, looking for a pattern)))))

 
Veniamin Skrepkov:
In the indicator linear_Price_bar I put the spread level =7 pips (the pair is one of the most volatile), for Volume 120-180, I do not have any calculations yet (I observe and conclusions are still ambiguous))))

No. This indicator is not suitable for analysing ticks - on the last candle, the indicator is redrawn many times. You will not be able to see the change of tick flow characteristics.


Veniamin Skrepkov:

I agree with the fact, that " the impulse is highly probable after the price leaves the flat, etc.". I agree, i.e. it is one of the variants of "working" on the border (corridor) of the flat, and we should widen the borders of the "impulse", because it is difficult to see it on M1. Probably it is very difficult to make a classification on М1.

I don't really care which timeframe the tick stream is on.
 
Veniamin Skrepkov:

I agree with the fact, that "the impulse is highly probable after the price comes out of the flat", i.e. one of the variants of "work on the border (corridor) of the flat". I think that it is one of the variants of "work" on the border (corridor) of the flat, and we should expand the boundaries of the "impulse", because in M1 it is probably difficult to make a classification. Probably it is difficult to make a classification on M1. Yes the classification as M1 forms M5 , and M5 forms M15 ? "In one case it may be a "trap" with preliminary buyers (preliminary bayers) or a target level when the fixing or buying is done, in one case the escape from the flat may be short (to release from the trap not to prevent the "averaging"), in the other case if the market performs its operations, the escape from the flat may be much stronger. Ideally, the indicator (advisor) should level out these "understandings" of the market that come out of the assumption, looking for a pattern)))))

You don't need M1, M5 and other timeframes. To search for impulse on tick data the chart can (and should) be hidden out of sight.
 

I think the author of these lines will not be offended if I quote him here:

Forum on trading, automated trading systems and testing trading strategies

Please advise on a good EA.

Yury Reshetov, 2015.09.18 15:14

...

If you want to approximate "clean" data, let's say tabulated data with small error, e.g. a sine wave, the curve will be smooth. In that case, the better the algorithm adjusts to the curve, the better.

If the data in the sample is "dirty", then the results will also be "dirty". Trash on the inputs is trash on the outputs.

Another thing is that there is no need to approximate pure table functions in application areas. Most often it is necessary to approximate the results of experiments. But there is no curve there, but a set of points scattered chaotically and crowded together in places where the curve should be. Well, nobody forbids to dissect, i.e. to pre-smooth these very scattered points into a curve using some algorithm before adding them to the inputs. I.e. pre-clean debris, but not to feed it to inputs. And then large degrees of freedom of the algorithm will not only prevent more accurate approximation, but will only contribute to it.

The last highlighted is a good idea applied to ticks. What if we compress ticks in packs (packs can be either of strictly defined size or floating) on the time axis. That is, it would be like this: take the last three ticks or five ticks. We get three or five points. You place these dots on top of each other - you get a certain cluster (such a small cloud). Take the next set of ticks and place them on top of each other again - you get a second cloud.
1

 
Karputov Vladimir:

I think the author of these lines would not be offended if I quoted him here:

The last highlighted is a good idea applied to ticks. What if we compress ticks in packs (packs can be either strictly defined or floating) on the time axis. That is, it would be like this: take the last three ticks or five ticks. We get three or five points. You place these dots on top of each other - you get a certain cluster (such a small cloud). We take the next pack of ticks and again place them over each other - we obtain the second cloud.

Thevolume profileindicator is a histogram of totalvolume at eachprice for the selected period of time (Market Profile). The price can reveal the level of accumulation (pro-trading) .