From theory to practice - page 324

 
Yuriy Asaulenko:

9%? Per month? On the depo? That's ridiculous). Absolutely not a serious system.

You have to do 20% a day, though, given the leverage. Well, at least 15%. Fuck your maths. Not the right kind of a stooge.

My leverage is only 10 and 5% a day is no problem for the system. The systems are similar, though.

Yuri, then what are you doing here?

If you've got a system like that, then... then this is what you do:

You take $1,000-- pennies.

You trust it to your system.

One day later, you got $1,000*1.05 = $1,050.

In two days, $1,000*1.05 = $1,102.50.

A year later 1000*(1,05)**365 = $54,211,841,577

54 billion!!!

Well we are honest arithmeticians after all, so let's accept that the result is overstated and correct for the fact that trading is only on weekdays - about 245 days a year.

Then in one year, 1000*(1.05)**245 = $155,373,973.

That's only $155 million per thousand.

But with THIS system you deserve more!

Get $5000 and wait not for a year but a year and a half...

Then 5000*(1.05)**(245*3/2) = $306,222,721,958

$306 billion. Three hundred and six billion dollars in a year and a half.

All those puppies on the Forbes list are shedding tears, because the best one of them has three times less!

Well, Yuri, with that kind of money you can solve the problem of low pensions in Russia, help state employees (except officials).

And in general to raise the economy of the country! After all, you will have to think where to invest that kind of money.

By the way from this amount the budget will receive about $40 billion from you, Yury, in the form of personal income tax.

Then Dmitry will say: "Hold on, there is money!!!"

Yuri, you're our darling! You are the savior of Russia! The diamond!

Do not waste your precious time on this forum, hurry up to get a measly $5k and in a year and a half....................


And seriously.

Yuri, I have read your posts and I believe that you are an intelligent, mature, thinking person.

And if even you, as a boy, measure pi...percentages, that scares me a lot more than the fact that we all don't know shit about non-entropy.

 
Alexander_K2:

Here's a look at what I'm talking about. EURGBP pair last week:

In case 1 the nonparametric skew coefficient = 0.44. In case 2 = 0.16.

Although even visually the two cases are identical. Well, you can't rely on nonparametric skew. We need something else.

Please understand my emotions - here I have a solution close to me, even a tiny percentage of profit, but I cannot complete the task.

And what do I see on the forum? Instead of discussing really serious problems - again some pair trading, arbitrage and other stuff.

Having looked at the figure my eyes tell me that there is no fish in this system and there should not be. Up to this point I had my doubts.

 
Alexander_K2:

I can officially confirm that the asymmetry coefficient by itself does not work.

That leaves Hearst and non-entropy. Okay, I can do my own research on non-entropy. But Hearst? Is it difficult to drop a link to studies where it is unequivocally confirmed that Hirst does not work either???

UNDERSTAND SanSanych Fomenko gave an extensive answer about Hearst.

Through him you can probably get access to libraries for Hearst calculations.

Of course, the libraries will not be under VisSim.

But apparently they are bolted to MT and can be raced in the tester.

It is also obvious that the most convincing answer about the applicability of Hurst researcher (you) can get only by putting yourself a numerical experiment - bibla-expert-tester.

And the last obvious thing is that it will take a lot of man-days.

 
Alexander_K2:

That's what I would like to know. I need a parameter that tells me exactly when the trend is over. That's it. This is the only difference between the market processes and the Wiener process. I know there is one. Only is Hearst?

I seriously want to ask. No quibbles, no banter.

I'm asking in all seriousness - how do you KNOW it exists???

If you believe it, I understand, but how do you know?

Can you mathematically prove it exists?

 
Alexander_K2:

Here's a look at what I'm talking about. EURGBP pair last week:

In case 1 the nonparametric skew coefficient = 0.44. In case 2 = 0.16.

Although even visually the two cases are identical. Well, you can't rely on nonparametric skew. We need something else.

So, take this mask and analyse it to prove the similarity that can be seen visually.
 
Uladzimir Izerski:

When the trend is over, you can calculate to the nearest couple of pips. But you will have to wait a long time for such an event. That's why I have a short-term trading tactic.

Fractal allows me to do it at any timeframe.

May I ask in more detail how to "calculate to the accuracy of a few pips"?

 
Alexander_K2:

Here's a look at what I'm talking about. EURGBP pair last week:

In case 1 the nonparametric skew coefficient = 0.44. In case 2 = 0.16.

Although even visually the two cases are identical. Well, you can't rely on nonparametric skew. We need something else.

Please understand my emotions - here I have a solution close to me, even a tiny percentage of profit, but I cannot complete the task.

And what do I see on the forum? Instead of discussing really serious problems - again some pair trading, arbitrage and other stuff.

All the problems are because of non-stationarity of the process, I built such a table here:https://www.mql5.com/ru/forum/221552/page300#comment_7026726

And here is a comment to the table:https://www.mql5.com/ru/forum/221552/page300#comment_7026811

Large single (rare) lags of increments (being in the tails of the distributions) give this distortion for statistics. So for historical values with a sufficiently large sample your coefficient will be much smaller than in the sliding window.

My opinion somewhere you should at least start from some stationarity of the process, or from equal arrival time intervals between ticks(operating time), from the quantization of the series as H-volatility by zz.

Time sliced TFs already contain all the "charms" of non-stationarity in the form of unstable volatility. That is why the most sophisticated methods are ineffective. If we find the right "reference point" for the process and the analysis methods are simple enough.

От теории к практике
От теории к практике
  • 2018.04.06
  • www.mql5.com
Добрый вечер, уважаемые трейдеры! Решил было на какое-то время покинуть форум, и сразу как-то скучно стало:)))) А просто читать, увы - неинтересно...
 
Serge:

...

And if even you, as a boy, are measuring pi...percentages, that scares me a lot more than the fact that we all don't know a damn thing about non-entropy.

What you're writing is utter nonsense - from beginning to end.

No one is measuring anything. What is being assessed is the A_K2 system, nothing more. The evaluation is a system that is not working.

A bit about your calculations.

If you trade currency without leverage and your income is only 1%/month - is it a lot? For trading it is not much. For any business it is not enough, and such a business is not worth doing at all.

For the same strategy with 100 leverage, you must inevitably have 100%/month.

I.e. A_K2 system, when trading without leverage, will only give you 0.09%/month. Is it worth doing such a business?

About millions-billions - let's leave the nonsense to the neighbours)).

 

Yuriy Asaulenko:

A_K2 system is assessed, no more. Assessment is a system that is not working.

It's not all that obvious.

9%/month with a 6% drawdown is quite acceptable, even though the profit may reach e.g. 12% at the end of the month, assuming that the drawdown does not increase.

 
Serge:

Can you elaborate on how to "calculate to the nearest couple of pips"?

I'm sorry. I wouldn't benefit from such a charitable act.