What is everyone looking for? - page 17
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
If we have property Ci of some object O, then this property can have only one numerical value, but otherwise it is not one property, but several - here is an example of what has not one property, but several. I think everybody mixes up or does not express his thoughts exactly, - but the point is that the ability to generalize - to combine objects having the same properties is what we need as a result of evolution. In the process of cognition of the world a man discovers new properties and gives them a characteristic. So a property is what the man sees. Let's take weather for instance: are there any indicators which can help to predict rainfall with high level of probability? A thermometer, a barometer, a humidity meter, and a clock (calendar). Well, essentially none of these devices is an indicator of the likelihood of rain. That is, making a projection of the properties of the world around the object of rain, none of them is an indicator. They are only characteristics. Once again, an object is a set of properties-we can pick any property and combine it into an object. For example, the object is air, which has humidity, temperature, pressure. Is it enough to analyze it as a probability of rain? No! We also have wind speed, solar radiation, air composition etc. Where is the boundary where you have to stop in detail? Probably in the transition from simple to complex until the desired accuracy is achieved. We need to distinguish indicators from indicators - there is an indicator that reflects numerical value of some properties of one of the objects that make up the object we investigate. There is also an indicator that reflects the properties of the other object.
In other words, there is pixel colour and there is photo beauty. But at transition from the low level to a higher level of detail no new properties are added in the (alphabet of the lower) level - this is man's ability to generalize.
But when the basic (atomic) properties are few - only they have to be analysed - but you cannot call them indicators of properties of the higher level alphabet.
Первую производную, и спектр, ну сам спектр включает в себя амплитуду на некой частоте - итого мы имеем по сути машки ( свойства спектра ) и первую производную . И все. Первая производная считается так - по средней цене ( (High(i) + Low(i))/2 ) - ( (High(i-1) + Low(i-1))/2 ) . Ну сами (Open-+-Close) это по сути наиболее вероятное значение ВР. Да и то Open и Close я бы не учитывал - это есть в сперктре.
1. A time series has NO derivative - not by any HIS definition in mathematics. If you have another definition of the derivative, please give it to me, citizen "mathematician".
2. Spectrum is such a controversial concept in modern science that it must be explicitly defined in any case of its use. Can you tell me what a "spectrum" is?
3. The relationship of "Mach to spectrum" is not as REAL as you probably think it is.
"The concept of AFC and FFC exists only for steady-state signals." (C) Mandelstamm, "Lectures on Oscillation Theory".
from some trend indicators you can get a vola, as both trend and volatility indicators use time and price (actually the change in price over time). The rest is answered in the previous post
Learn to read: you can't get vola from trend, but you can get both vola and trend from BP.
Из одного индикатора нельзя аналитическим путем получить другой индикатор. Из индикатора тренда нельзя получить индикаторы объема, а из них обоих - волантильности. Хотф все моожно получить из исходного ВР.
Not so fast. Personally, I'm ready to agree with you that it would be AWESOME to have market slices (pictures) given by different, slightly independent (orthogonal) indicators.... But there are a few reservations:
1). for example tick volume (and we speak mainly about trading on FC) is actually smoothed and then quantized Noise of a time series, and therefore it hardly can be strongly independent from the time series. And then there is another problem - the circumstance you specified - about the deductibility of one indicator from another because of the number of methods used to obtain them.
2). So when speaking about "orthogonality" of indicators, if you do it directly and roughly, without detailed analysis of how an indicator is defined analytically and how it is calculated, you can easily miss the point.....
I personally agree with Svinozavr's comment - it should be a VERY FINE, ADVANCED WORK, not a rush.
I believe that most price range models fail all the time precisely because of this - trying to cavalierly attack the market.
Not so fast. Personally, I'm ready to agree with you that it would be AWESOME to have market slices (pictures) given by different, slightly independent (orthogonal) indicators.... But there are a few reservations:
1). for example tick volume (but here we are mainly talking about forex-editing on DT) - it's actually smoothed and then quantized noise of a time series, and therefore it hardly can be independent from the time series. And then comes the circumstance you indicate - about the deductibility of one indicator from another indicator - due to the number of methods of their obtaining.
2). So when speaking about "orthogonality" of indicators, if you do it directly and roughly, without understanding precisely how an indicator is defined analytically and how it is calculated, you can easily miss the point.....
Therefore, I personally agree with Svinozavr's comment - it must be a VERY FINE, ADVANCED WORK and not an easy one.
Proving orthogonality is a tricky thing. But do we need it at our level. Suppose a TS is made on different mash-ups, which do highlight some new BP properties. But it doesn't take into account overbought/oversold market or volatility. I argue that a TS including trend, overbought and volatility indicators is potentially more promising than one based on the wizards.If we have property Ci of some object O, then this property can have only one numerical value, but otherwise it is not one property, but several - here is an example of what has not one property, but several. I think everybody mixes up or does not express his thoughts exactly, - but the point is that the ability to generalize - to combine objects having the same properties is what we need as a result of evolution. In the process of cognition of the world a man discovers new properties and gives them a characteristic. So a property is what the man sees. Let's take weather for instance: are there any indicators with which one can predict rainfall to a high extent of probability? A thermometer, a barometer, a humidity meter, and a clock (calendar). Well, essentially none of these devices is an indicator of the likelihood of rain. That is, making a projection of the properties of the world around the object of rain, none of them is an indicator. They are only characteristics. Once again, an object is a set of properties-we can pick any property and combine it with an object. For example, the object is air, which has humidity, temperature, pressure. Is it enough to analyze it as a probability of rain? No! We also have wind speed, solar radiation, air composition etc. Where is the boundary where you have to stop in detail? Probably in the transition from simple to complex until the desired accuracy is achieved. We need to distinguish indicators from indicators - there is an indicator that reflects numerical value of some properties of one of the objects that make up the object we investigate. There is also an indicator that reflects the properties of the other object.
In other words, there is pixel colour and there is photo beauty. But at transition from the low level to a higher level of detail no new properties are added in the (alphabet of the lower) level - this is man's ability to generalize.
But when the basic properties (atomic) are not numerous - only those must be analysed - but they cannot be called indicators of properties of the higher level alphabet.
This is already a philosophy, although proponents of behavioural economics would take a different view. They take the properties of the human psyche (the list is limited and universally recognised) and derive market movements from it.
Я утверждаю, что ТС, включающая индикаторы трендовый, перекупленность и волантильность потенциально перспективнее, чем построенная на машках.
Now, that's a question. It will be necessary to see HOW YOU MANAGER the readings of these indicators, how they interact with each other in your control system. Perhaps, you will be able to build such a harmonious system with the wrenches, which will be stronger than a combination of three quite different indicators.
They are linear (more or less). That is, in each case, you have to look at it. What you have said is not a principle, not a law, but simply a statement of your personal experience.
Learn to read: you can't get an ox from a trend, but you can get both an ox and a trend from a BP.
Learn to write: many trend indicators consider vola and/or contain price change over time and you can get vola from them ;)
Take a wahm which is considered a trend indicator. Count its change over a fixed period and you will get volatility :)
I argue that a TS that includes trend, overbought and volatility indicators is potentially more promising than one built on wols.
P.S. overbought/oversold, volatility and trendiness are not orthogonal or independent. There are only 3 dimensions: price, time and volume. Fixing one "coordinate" counts changes in the other or 2 others. All combinations are not difficult to enumerate. Although besides change, it is possible to count other - integral characteristics (max, min, sum, average, etc.)
learn to write: many trend indicators take vola into account and/or contain price change over time and you can get vola out of them ;)
Take a wave that is considered a trend indicator. You can get volatility with it per fixing period :)
No way. Intuitively, they say a change in volatility precedes a change in trend, and being stuck in overbought areas leads to something. I attribute it to the general culture of writing TS. When writing a program the nested blocks may or may not be shifted - one day non-shifting will play a mean trick.My attitude to indicators is the same as to the Chukchi song, which goes and sings what he sees. You have to start with BP characteristics: stationary, non-stationary or both at the same time at different sites or even at the same site. Then using mathematics that has worked fine for many years in other areas.
learn to write: many trend indicators take vola into account and/or contain a change in price over time and you can get vola out of them ;)
Take a wave, which is considered a trend indicator. Count its change over a fixed period and you will get volatility :)
Well, brother bent. Volatility is the difference between price and average. It is the higher frequency component of BP. A wavelet decomposition gives a decomposition into these frequencies, the first of which is the trend. The sum of all this is the initial BP. There is an orthogonality proof for this decomposition.