From theory to practice - page 48
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
I googled "Integral quantile by Euler" and did not find it. May I ask what you called those words?
A share of humour in this topic.
I already wrote in another thread on this topic: If I count some statistics/filters on real ticks in realtime -> xilinix (from available Spartan - copy and modify project and get result, also there were industrial solutions for a lot of money). Increase in ticks makes no sense - different result on different brokers, but if you reduce all ticks to a minute you get ~similar result for several broker/dealer during active trading time. If you reduce it to a minute anyway to get a near true result, then where is the point of looking at ticks?
Previously (about 20 years ago), you had to be able to read the schematics of devices, and it's a lot of academic and practical knowledge at a good level, and today it's ordering in China ready-made device on alibaba with possible soldering extra (for Chinese) parts according to datasheet. I.e. today you need to take a ready-made filter circuit in some package and feed it with input data - academic arguments are unnecessary.
About the assumed baseline model for the study: if we talk about the frequencies of something, from available it is M1, the maximum frequency is M2, and the studied price series M4 (almost M5) - is if as a physicist assume the existence of a market analogy of the physical phenomena with a frequency of 2 times higher(M2) than the initial process(M4) . Or we construct 15-second bars by ticks, by which we study the price series M1. Or for example another model with 3rd and 5th harmonics, or their compilation, etc. DJ FXCM Dollar index launched with update every 15 seconds and only 4 currency pairs - using an interval of less than 15 seconds will give an advantage over the creators of this index (and other participants)? There is no author's model to explore in this thread. Then what's the fuss about?
Forum on trading, automated trading systems and testing of trading strategies
From Theory to Practice
I built it,Renat Akhtyamov, 2017.12.10 08:29
Has it ever occurred to you that ticks can come in packs to one, at regular intervals to another, and at some other intervals to a third?
Then when comparing the quotes all the above theory is nonsense?
The point is that you have to time-synchronise the graphs first, and only then compare them.
But I'm 100% sure that nobody will be able to do it with ticks, unless we have the same communication channels, the same receiving and transmitting equipment, etc.
And it is not for nothing that the basic reference point for comparison is a chart on TF M1.
But even on М1 the charts may not coincide because the first and the last ticks of the minute may fall in different candlesticks of different brokerage companies and will be different again.
And I didn't like the fact that some apparently geo-physicist tampered with the formula from the start and is still pounding his chest - I'm mol....Why do you need names? Brownian - not Brownian, chi-square or Statistic - what difference does it make... Why not use the identified regularity, even if it does not fit any distribution (of probability), even if the frequency does not have statistical stability (remember the reference to Gorban) and classical probability theory is not applicable. And what, to be afraid of this?
And why are you in such a hurry? Check my conclusions, double-check, just in case.
Yeah, yeah...
Double-checked again. Indeed, by taking tick data in this way (at exponential time intervals and with averaging as well) we almost completely "destroy" the non-Markovian process - it becomes a regular Markov chain with independent CBs. The probability distribution is regular bilateral geometric one and mathematics for Markov chains can and should be applied to such a sequence.
For example, for EURJPY I've posted earlier, p=0.14 (probability of success), q=0.86 (probability of failure). You can check it yourself.
But the shame is the shame - I was mistaken, I didn't see. I lost the non-marking of the process and I MUST NOT use historical data in this case!
I repent, weep and weep...
Sincerely,
Alexander.
Alexander_K2, do you have time to answer the questions I asked?
Holy shit. And my daughter already ordered louboutins on ali.
And paid with Dad's credit card ;) ?
Alexander_K2, do you have time to answer the questions I asked?
Yes, yes, of course...
Only in the evening - I need to get away from such a knockout... After all, I've already started the program on a demo account today, and it turns out all wrong!
The reason - I still cannot understand - how exactly one should accept tick data, not to destroy non-marking, to be able to use historical archives...
I don't know what you mean by non-marking in this case, but just changing the reading pitch will hardly break the dependency in price. Try just taking a 10 second pitch, maybe that will reduce the dependency on brokers. I don't think a bollinger-like system with long trade durations will suffer from this.
Thousands of people, including those with PHD, are looking for patterns and building systems on minutes, and they feel fine, while you, instead of studying their experience, are trying to reinvent the wheel. You can bang your head against the wall for years.
Financial markets are a huge industry with a lot of smart people, everything has been discovered before you. And for some reason you think it's for suckers, schoolchildren).
I don't know what you mean by non-marking in this case, but simply changing the reading pitch will hardly break the dependency in price. Try just taking a 10 second pitch, maybe that will reduce the dependency on brokers. I don't think a bollinger-like system with long trade durations will suffer from this.
Thousands of people, including those with PHD, are looking for patterns and building systems on minutes, and they feel fine, while you, instead of studying their experience, are trying to reinvent the wheel. You can bang your head against the wall for years.
Financial markets are a huge industry with a lot of smart people, everything has been discovered before you. And for some reason you think it's for suckers, schoolchildren.)
Yes, I accept all the rebukes.
But, it turns out that if there is no t2 distribution in the market, then it makes no sense to use historical data.... I find this extremely frustrating...
You are welcome to use it. The regularities aren't going anywhere. I'll tell you more, the kind of distribution doesn't matter at all. But it will take you a few more years of trial and error to realise that.
Your proverbial non-markness is the INDEPENDENCE of future changes from past ones. And the distribution (any distribution) contains no time dependence data at all.