Market model: constant throughput - page 23
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
There are two objects that are connected to each other by a communication channel. The objects can be people, computers, etc. The communication channel has a limited bandwidth.
The objects on the communication channel can exchange data.
Are you suggesting that there is limited bandwidth in the communication channel between the MT terminal and the DC and something is lost? That's for sure! These very objects are not telling us anything, they are hiding something. They are very suspicious.
:о)
The information contained in the data is the smallest set of bits that can be transferred and restored to the original data. The amount of information is the number of bits in that very minimal set.
There is a "theory of information". Why not start familiarising yourself with it? "Bit" is just a chosen "scale" - a logarithm on base two, which is taken from the probability of one of the states the system can be in. You can choose whatever scale you like, but it doesn't add any magic.
Information is essentially the difference between the current uncertainty and the fact of receiving a message about the unambiguous state of the system (entropy is zero, we already know exactly what state the system is in). So far I don't really understand what you are writing about.
Obviously, if BP1 contains an amount of information N, then BP2 = BP1 * Const contains an amount of information equal to N as well. Reasoning a little further, but also simply, it can be argued that if there is a rigid functional relationship between BP1 and BP2, they contain the same amount of information. And if you need to transmit both BP1 and BP2 at once, you don't need to transmit N+N amount of information. You'll have to transmit a bit more than N
That's not true, to put it mildly. Although I admit that I understood it in my own way. I will say briefly - if you by happy coincidence have two rows with the same entropy (information) it will not give anything and about anything fundamental does not speak.
PS: There is an NS based on information theory. I read it briefly, it's quite interesting. But didn't get to the practical use.
The topic is totally lacking not only in terms but also in the general concept of the theory of information transfer.
If the topic is going to develop into something informative, then:
1.Объективность информации. Объективный – существующий вне и независимо от человеческого сознания. Информация – это отражение внешнего объективного мира. Информация объективна, если она не зависит от методов ее фиксации, чьего-либо мнения, суждения.
Пример. Сообщение «На улице тепло» несет субъективную информацию, а сообщение «На улице 22°С» – объективную, но с точностью, зависящей от погрешности средства измерения.
2.Объективную информацию можно получить с помощью исправных датчиков, измерительных приборов. Отражаясь в сознании конкретного человека, информация перестает быть объективной, так как, преобразовывается (в большей или меньшей степени) в зависимости от мнения, суждения, опыта, знаний конкретного субъекта.
3.Достоверность информации. Информация достоверна, если она отражает истинное положение дел. Объективная информация всегда достоверна, но достоверная информация может быть как объективной, так и субъективной. Достоверная информация помогает принять нам правильное решение. Недостоверной информация может быть по следующим причинам:
преднамеренное искажение (дезинформация) или непреднамеренное искажение субъективного свойства;
искажение в результате воздействия помех («испорченный телефон») и недостаточно точных средств ее фиксации.
Полнота информации. Информацию можно назвать полной, если ее достаточно для понимания и принятия решений. Неполная информация может привести к ошибочному выводу или решению.
4.Точность информации определяется степенью ее близости к реальному состоянию объекта, процесса, явления и т. п.
5. Актуальность информации – важность для настоящего времени, злободневность, насущность. Только вовремя полученная информация может быть полезна.
6.Полезность (ценность) информации. Полезность может быть оценена применительно к нуждам конкретных ее потребителей и оценивается по тем задачам, которые можно решить с ее помощью.
These are the postulates to start with and then what to do with the information - compress it, filter it or not perceive it at all is another matter
As for trading, I can agree that the information is in the crosses, and it is absent in the majors. Why? Because all the majors work for the market - there is a trend, there are players and the players either strengthen the trend or weaken it, and so on, until suddenly all the players are under the intervention
With the crosses it is a little bit more complicated, there are not many players, but in the crosses there is an economic and geographic dependence, at least I think so, because the currency indexes are calculated mainly through the crosses
to hrenfx
Are you suggesting that there is limited bandwidth in the link between the MT and the DC terminal and something is being lost?
We are not talking about an MT or a DC. The object from which the information comes is the market. The subject that receives this information - the trader. The information loss due to filters and other things is not considered.
There is an "information theory". Why not start to get acquainted with it? "Bit" is just a chosen "scale" - logarithm on base two, which is taken from the probability of one of the states the system can be in. You can choose whatever scale you like, but it doesn't add any magic.
There's a lot of stuff in information theory. And the definition of information that is given there does not seem to apply here. I gave the definition above. If you do not like the word information, you may find another one.
Information is essentially the difference between the current uncertainty and the fact of receiving a message about the unambiguous state of the system (entropy is zero, we already know exactly in what state the system is). So far I don't really understand what you are writing about.
If you understood my definition of "information", the hypothesis of constant market bandwidth says that the market generates the same amount of "information" per unit time.
Why do I say that all "information" is contained in majors? If you transmit EURUSD, GBPUSD and EURGBP, then you only need to transmit EURUSD and GBPUSD to receive data on EURGBP as well. I.e. there is enough "information" in the majors.
Why are we talking about the amount of information in the first place? The point is that the "information" of the market contains all the regularities of the market, because it is the knowledge of all the regularities that makes it possible to generate the very minimal set of bits for transmission.
Information is a set of bits that cannot be compressed in any way to transmit.
It is assumed that the market, as a relatively closed system, generates a constant (or slowly changing) amount of information per unit of time.
There's a twist here. I will say for sure - it works if you apply it correctly. And the model is much simpler than the one you describe.
Why do we even talk about the amount of information? The point is that market 'information' contains all the regularities of the market, because it is through knowledge of all the regularities that it is possible to generate that very minimum set of bits to pass on.
Quote from the first post:
Compressing as much as possible is a theory. There are a lot of compression algorithms. The harder an algorithm compresses, the closer it is able to estimate the amount of information contained in the available data. That is, we cannot determine the exact amount of information, but we can estimate it.
Quote from the first post:
no we can't evaluate with standard archiving algorithms. The redundancy during encoding is estimated and that's it. No regularities of the process itself are extracted by these algorithms
Universal archivers do not use the search for patterns of a particular data type, but eliminate the redundancy of coding for any file. In order for the archiving algorithm to find patterns in a particular row and compress it optimally - these patterns must be embedded in it. JPEG has features of two-dimensional images, mp3 has features of audio information etc. There are no miracles, that some algorithm can find unique features of any information. For video: they write new codecs all the time - there is no ideal. Therefore your task cannot be solved without such special archiver. :) And since no ideal compression method can be found here either, you can only compare the degree of correspondence of a particular BP to the regularities inherent in a particular codec
Let me add to what you have justly said. If we consider some alternation of ticks as a forte regularity - the topikstarter should have used for "using" archivers in the initial example (Prival is not present - I'm for him :) the information about price change as a result of a tick in one direction or another, using one byte for representation of this event ( and not double, as it most likely happened ;). 128 pips per tick is quite informative and rare...
As for HrenFix's plausible reasoning - I support it by definition. Otherwise it's just flooding and moaning about unsteadiness and "thick-tailing".
But proofs are even better!
Unfortunately there is no time to be engaged in it all - there is a backlog in: trading; programming; personal or/and family life ... (choose according to taste).
And Alexey is right, a post like this
I disagree. If you replace the word "currency" with "potatoes", it sounds even more dubious.
To measure an object, you have to have a unit of measurement. Length can be measured in metres or inches. Weight in kilograms or pounds. EUR in USD or JPY.
Often in physics problems, there are situations where, for example, length is not important. And to simplify logic, it is taken as a unit (kilometres or feet - it doesn't matter. The main thing is that the other quantities are measured in the same units). So you can take the same EUR always as a unit.
In general, you can't measure anything without units.
P.S. Of course, you can also measure USD in potatoes.
As far as I'm concerned "stream of the twilight of consciousness". We measure within a pair - then the currency of profit/loss is important. Measuring momentum - sign and some relative amount of deviation from previous rest is important.
But if you don't provoke - the swamp will become muddy. So kudos to DickPhoenix!
Write again ;)
We cannot estimate with standard archiving algorithms. The redundancy in coding is estimated and that is all. These algorithms do not retrieve any regularities of the process itself.
It is always possible to estimate. The question is the reliability of the evaluation.
The possibility of using media compression has been discussed. Finding and extracting a little, but still a lot, of redundancy from the data is better than nothing. Elementary compression algorithms have shown that they can handle it. Estimating the amount of information is, of course, a crude estimation.
You can theorise a great deal. And one can do minimal research, which has been done.
It is possible to theorise a great deal. And one can do minimal research, which has been done.
Only a correct one. rather than revel in the uniformity of the mantissa...
;)