FR H-Volatility - page 40
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
Yes, that's exactly what I was counting on. I still need to figure out how to scale the x-axis.
What if we simply calculate Bars*TpB - Ticks, TpB is the average number of ticks per bar? More precisely, the derivative of this value in order to get rid of uncertainty with the datum. I mean to avoid drawing of equivolume bars.
Please be a little more specific. There are no equi-bars in my picture. The red line is MOW of ticks on Y and X axis. You can see a strong divergence although it shouldn't be like that (leading + smoother look). Or we may conclude that it is better not to use quotes of one brokerage company but analyze them by ourselves (collect them from different sources). Then our forecast will not depend on quotes (filters) of brokerage companies. Which may be better.
No, Prival, if ticks are used only for mercantile purposes, they should be taken from the brokerage company you work with. And the analysis of other brokerage companies ticks - only for searching regularities, not dependent on a brokerage company.
2 Candid: That is an interesting idea of course. However I think I have not exhausted all possibilities, associating exactly with equivolume bars. What is curious is the following: despite such a considerable local difference in numbers of bars, the difference is minimal on the full history. I have posted pictures for H4 a little earlier. If we take the entire history I have (since 1999), the amount of H4 bars differed from the amount of equivolume bars by 0.1%.
No, Prival, if we are going to use quotes for mercantile purposes, we should take them from the brokerage company we work with. And the analysis of quotes of other brokerage companies is just a search for stable patterns that do not depend on a brokerage company.
And I thought it was mercantile (from French and Italian: mercantile, self-serving), excessive prudence, haggling; self-interest. If using data of other quotes providers I will predict faster the start of price movement (for example if there is a delay in DC filter), then I will obtain more profit + my sampling frequency is higher, consequently I may trace more fine fluctuations. This statistic advantage may be small, but there and there a little may scrape up a loaf of bread :-). I simply trade according to quotes of brokerage companies, it is true, I cannot avoid it, but I am not obliged to use only quotes of brokerage companies in my analyses.
A little more specific please, I don't understand a little bit. I don't have any equi-bars in the picture. The red line is MOJ of ticks on Y-axis and X-axis. You can see a strong divergence, although it shouldn't be like that (leading + smoother look).
2 Mathemat: Well I wrote above that I was pursuing very limited goals.By the way, I remembered that I used to take the difference between long and short averages in volumes for the same purpose, but finally I gave up trying to use the tick volume, because I considered the long term stability of this source of information unsatisfactory. Regarding the diminishing difference in the number of bars with the growing period, I think it goes back to the daily (sessional) cyclic activity. In general, presenting history in the form of ordinary bars we destroy temporal information in a relatively simple way, while in the form of tick bars it is more complex. But it seems that the higher the timeframe, the closer the result is :). What I don't like about the "tick" transformation - on short times, if we can talk about its generality (i.e. uniformity for any point in time), it is very difficult. And on long times its results, judging by your work, turn out to be close to essentially simpler and more universal "bar" transformation of the time axis.
I can't explain what I'm building here ('I need an indicator which reflects the price in operational time'), but I'm not writing it right. I will try to explain in numbers.
Suppose we have 5 ticks (5 numbers Y=Bid+(Ask-Bid)/2), each coming at its own time t1, t2, t3, t4, t5. (these are also numbers). We calculate the LFL by prices (this is coordinate Y) and the LFL by time (coordinate X). We plot this point on the screen (X,Y). The next tick comes, the procedure is repeated, the next point is drawn. It is an ordinary Mashka, maybe it is clearer, only not only Price, but also arrival time is averaged (we can make it more precise, not just average, but according to ISC).
And the chart is superimposed, the first is what I calculated by ticks (red). The 2nd one is the minutes from Alpari for the same time period (blue). The beginning and the end of the charts should exactly coincide in time, if I am not mistaken in my calculations.
Had to take from different sources as I have no Alpari tick history, only minute history. If anyone has any, let me have them, I will repeat the calculations. Need 2 files 1 minute and the second ticks for the same period of time. (At least a week).
Z.I. What is "equi one tick :) bars", I still can't figure it out.
Z.I. What is "equi-capacity one-trick :) bars", I still can't figure out.
Yep, your notion I remembered. But in fact, we are talking about the average price of an equi-capacity bar, for the specific example of a 5 tick bar. It seems to me that this should not change anything in principle (compared to simple equi-capacity bars)
Z.I. What is "equi-capacity single-tick :) bars", I still can't figure out.
Yeah, your notion I remembered. But in fact we are talking about the average price of an equi-capacity bar, for the specific example of a 5 tick bar. It seems to me that in principle, it shouldn't change anything (compared to simple equi-bars).
I think it does. If you remember, I wrote about the sampling rate. If we take 1 value per minute, it takes a very long time to detect. This is important for the zigzag as well. (well its averaging time is 300 min). It is important to determine the break point as quickly as possible (in terms of time), so as to capture as much as possible from the zigzag.
With a build like mine, the incoming frequency is higher. Perhaps this will give some advantage in detection time + the curves are smoother + sort of ahead of the curve.
Edit. Although in my picture is something screwed up, but what and where I can not understand. Such a big difference like there should not be between the DC (unless of course it is a kitchen). One of them should have died a long time ago, as arbitrage in its pure form.
I think there is. If you remember, I wrote about the sampling rate. If we take 1 value per minute, it takes a very long time to detect. This is important for the zigzag as well. (well its averaging time is 300 min). It is important to determine the break point as quickly as possible (in terms of time), so as to capture as much as possible from the zigzag.
In general with regard to realtime I agree, I myself look somewhere similar. But the real gain is not as great as one would hope, because the conclusion (about the break of the zigzag, for example) should be more or less statistically secured. Working on minutes is good because having a good "reaction speed" we can test the algorithm on the history in conditions very close to the real ones.
P.S. By the way, I remember using the term "sliding bar" once :)
P.S. By the way, I remember using the term "sliding bar" once :)
I had to take it from different sources as I don't have an Alpari tick history, only a minute history. If anyone has any, please let me repeat the calculations. Need 2 files 1 minute and second ticks for the same period of time. (at least a week).
Catch. In the zip the minutes and ticks on the euro-dollar for August this year from Alpari.
By the way, has anyone tried to compare EUR/GBP ticks and real-time synthetic ticks by the ratio of EUR/JPY to GBP/JPY ticks. If the volume of EUR/GBP monthly bars is on average 5 samples/min, for the synthetic series (SR) the data receipt rate is on average 20 samples/min, moreover, the absolute values of these two VRs coincide to within a point! In other words, by analyzing the SR we obtain a tool that shows us the expected quote of EUR/GBP in advance.