Matstat Econometrics Matan - page 14
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
The Hearst exponent is used in economics - intechnical analysisto justify trend prediction (in the above function the initial series will be price increment), in natural sciences - in analysis of various experimental data - to identify new process characteristics [5].I suggest comparing Hearst with PNB capabilities. I am ready to analyse any series. The condition: the data of series should be fixed in equal intervals of time and the same figures should not be next to each other. Any method will not describe the series better than PNB and will not give its characteristics. Any process is described by one parameter and two coefficients, analysing the difference of the numbers. In total at least 4 digits of the series are required, according to the principle the more the better. Finally, put a stop to the search for the best pattern in any series. Stop looking for a pattern at last, or give a decent fight to the PNB while I am alive!
I disagree, by thinning the tick flow you can achieve more accurate signals even in a trend.
If the HFT strategy is on ticks, it is possible. First of all using the correlation between adjacent ticks. And this correlation exists.
It is already meaningless on minutes. The ACF there does not differ significantly from the ACF of SB.
No magic! The usual counter-trend strategy. The counter-trend is unpopular for obvious reasons.
Trend traders lose a little bit on the flat, and earn a lot on the trend. On the flat, you drip yourself some valerian. On the trend, you drink champagne.
Counter-trenders make a little money on the flat, but lose a lot on the trend. On the flat, you pray quietly and keep your fingers crossed. On a trend, the ambulance takes you away with a heart attack.
Sasha bet $160. Well, that's like an admission ticket to the ride. You don't mind losing. But it's a lot of fun.
A clean counter-trend is not good for serious business.
A flat is a trend that hasn't developed. You need to switch from a trend to a counter-trend in a timely manner.
It won't. For example, calculate Hearst day and night. Or at low daily volatility and at high volatility (it doesn't change much in the market). During periods of news and without news. Disturbances are detected by the multicurrency analysis. This is what distinguishes the market from the SB - the "physics" of real life, not magic with formulas)
Surely it is possible to "tweak" the SB to achieve similar effects (while maintaining the inability to earn). For example, by introducing a variance dependence on time. Selective Hurst on realization of such "corrected" SB will probably fluctuate according to the party line, but the possibility of earning will not appear.
I'm not saying that Hearst means nothing, but only that one should not be lazy to study theory, which will help assess its importance in a quantitative sense, and not only with a harsh practical eye)
Here, for example, is an article estimating the confidence interval for Hearst on SB.
Surely it is possible to "tweak" the SB to achieve similar effects (while maintaining the inability to earn). For example, by introducing a variance dependence on time. Selective Hurst on realization of such "corrected" SB will probably fluctuate according to the party line, but the possibility of earning will not appear.
I'm not saying that Hearst means nothing, but only that one should not be lazy to study theory, which will help assess its importance in a quantitative sense, and not only with a harsh practical eye)
Here, for example, is an article where the confidence interval for Hearst is estimated.
I think it's obvious that a sample Hearst will have a non-zero variance, and will fluctuate around a certain value without any "edits". For SB about 0.5.
Surely it is possible to "tweak" the SB to achieve similar effects (while maintaining the inability to earn). For example, by introducing a variance dependence on time. Selective Hurst on realization of such "corrected" SB will, for sure, fluctuate according to the party line, but the possibility of earning will not appear.
Already on the minutes it makes no sense. There the ACF is no longer meaningfully different from the SB's ACF.
Minutes of currency pairs are reversible, but the average trade is below the spread.
I also became convinced of this when I drastically ran up a deposit. Even the multi-vault mode didn't help.
You beat me to it ))). I was about to write the same thing. Now I'll add to it.
Firstly, it can actually be measured. In my time I have done a lot of Hearst measurement on different markets and instruments.
Second, one has to measure selective Hirst with its non-zero variance. And in almost all cases the value 0.5 fell within the confidence interval. On large samples one can obtain a more exact value, for example 0.51, and then make sure that trend TS works on this instrument. Or vice versa, we can obtain 0.49 and check thattrend-trending TPs fail.
Third. Hurst allows us to perform an in-depth analysis. For example, the question: will thinning of ticks improve parameters of the trading system on a minute chart? The answer is no, because thinning does not change the Hurst, does not change the persistence of the series on the minutes and therefore does not increase the profitability of the trading system.
The Hearst exponent is used in economics - intechnical analysisto justify trend prediction (in the above function the initial series will be price increments), in natural sciences - in analysis of various experimental data - to identify new process characteristics [5].I suggest comparing Hearst with the PNB capabilities. I am ready to analyse any series. The condition: the data of series should be fixed in equal intervals of time and the same figures should not be next to each other. Any method will not describe the series better than PNB and will not give its characteristics. Any process is described by one parameter and two coefficients, analysing the difference of the numbers. In total at least 4 digits of the series are required, according to the principle the more the better. Finally, put a stop to the search for the best pattern in any series. Stop looking for a pattern at last, or give a decent response to PNB while I am alive!
There are (and quite a few) Chart Trading Systems/Methods that don't careat all which eye is on which tube, i.e. Hearst or H-Aolatility. The chart division into trend / flat is not a given and a feature of any series, but only a very narrow view of the chart and trade, of course one can do this (but then it is a dead end), or one can do it in another way. Hence, we can derive entirely different methods, which cannot be attributed either to the trend methods, or to the flat methods. The world and the market are multifaceted. But the way you look at it is the final result. They divide it into trend/flop and with 0.5 we reach a deadlock, we reach a deadlock with Hirst: "What is earned by the flat system is lost with the trend one and vice versa". There is a good cartoon on the subject - but I am too lazy and have no time to look for the link. (Expand your minds fellow traders:)))