What is not understood and taken into account when discussing the random walk model (RBS) - page 7
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
This is the essence of it. Fair enough for an individual implementation. It is to exclude the influence of individual realisations that the post I quoted https://www.mql5.com/ru/forum/221552/page19#comment_6168925 analyses the minute-by-minute incremental statistics for 625 days. So that the conclusions are general and reproducible. As Alexander_K2 found out, the obtained information works not only for the selected usdchf pair at the selected range (intraday per minute activity over 2 years, 2015-2017), but for all currency pairs and any other time. The conclusion of lack of stationarity is based on 625*1440=900000 (almost a million) measurements, not a single realization.
I have to say that in the same way not only does it reveal a statistically significant dependence of increments on time of day, but also their dependence on day number within trading week 1-5. However, it is weaker than the intraday dependence.
Move on from models where time introduces a distortion of what is being observed. The bar cutoff and the number of wanderings within a bar at different times of day can indeed distort the observed greatly.
But I suspect that extreme value theory (and the probabilistic models used in it) can really change many perceptions and dispel some misconceptions.
Also, many "refutations" of an efficient market are built on the assumption that we have a Gaussian distribution of elementary increments.
And this is likely to be far from the case. )
Nothing against using SB for any purpose when describing price dynamics. I agree and have even used in trading the fact that over an infinitely long period of time zero expectation of rate increases will also work. I can't understand, on the basis of what, without specifying the purpose of modelling, the model "with STATIONARY random increments" suddenly turned out to be "most adequate" to all unknown purposes at once? It seems to me that this forum is discussing possibilities to detect local regularities in price dynamics that would lead to deviation of the expectation of the sum of rate's increments from zero during the time from opening to closing of a trade. See, for example, right on this forum https://www.mql5.com/ru/forum/221552/page430#comment_7945889 and about ZKK describing these changes https://www.mql5.com/ru/forum/221552/page19#comment_6168925. That is, there is no stationarity of the increments during the day and the SB model with stationary increments is not suitable for describing the price dynamics during the day.
Indeed!
Once again I will respond to the substance of the proposed strategy, most likely the last, if there is no implementation on your part
I don't speak to empty-nesters in that tone of voice.
You are on ignore until you apologise.
All the best, communicate with others elsewhere.
Move away from models where time introduces a distortion of the observed. Bar cut-offs and the number of wanderings within a bar at different times of day can indeed greatly distort the observed.
But I suspect that extreme value theory (and the probabilistic models used in it) can really change many perceptions and dispel some misconceptions.
Also, many "refutations" of an efficient market are built on the assumption that we have a Gaussian distribution of elementary increments.
And this is likely to be far from the case. )
It's a shame that you, seemingly a new character in this thread, are just as vacuous, without a single formula or graph, making a completely unscientific, unsubstantiated assertion about something.
One more meaningless post from you in this thread and you will be in my ignore.
Write statements based on scientific knowledge, not on your feelings or your hunches.