Renter - page 30
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
Wouldn't it be better to have this kind of performance...
.
.
I should probably dig into the special functions...
.
Prudnikov, Brychkov, Marichev. Integral and Series. Moscow, Nauka. 1981.
Prudnikov, Brychkov, Marichev. Integral and series. Supplementary chapters. M. Nauka. 1986.
.
is a collection, like Kamke's ODE Handbook.
Searching through this sea of information is a lot of work.
but it may be worth it!
After this replacement it all comes down to a derivative of a complex function (if I remember correctly): df(s(k))/dk=df(s)/ds*ds/dk
ds/dk is taken, but df(s)/ds is no easier than the original df(k)/dk and no worse than a dick.
This language is quite adequate to describe linear dynamical systems. Oleg, your reasoning about lattice functions, frankly, just killed me. There were no such difficulties in the original problem.
About flexibility, I agree.
1. this language is quite adequate to describe both linear and non-linear dynamical systems, both deterministic and stochastic. Of course, it also has its limitations and area of applicability.
.
2. I will not present here the theory of lattice functions. I will only draw your attention to the behavior of the function you have compiled: with every new count its degree increases by one. As long as we are talking about several counts, there is nothing wrong... even with thirty to fifty to a hundred counts... But if you have to work with signals whose frequency is measured in kilohertz, with your approach you have to increase them to degrees measured in thousands. For signals with frequencies in the MHz range, degrees measured in millions... and so on.
That's what I'm talking about.
After this replacement it all comes down to a derivative of a complex function (if I remember correctly): df(s(k))/dk=df(s)/ds*ds/dk
ds/dk is taken, but df(s)/ds is no easier than the original df(k)/dk and the dick is not sweet.
.....
Man, postnumerando annuity, for fuck's sake...
Decent man, moderator, expert in wave theory, and you swear like a cobbler. :)
Sorry for the off-topic.
When the effective interest rate was introduced in banks (Basel does not recognise any other), it was even worse...
Same rake - just a little on the side.
;)
Alexey!
I didn't look at the formula, but read it:
15% от всего накопленного депозита мы снимаем:
Therefore I mistakenly thought that you too were solving the right problem - without contrived constraints...
We have been solving this one for a week already, FreeLance... let alone the correct one... But the crux of the problem seems to have been pinpointed by Oleg:
avtomat: 2. Я не буду здесь представлять теорию решетчатых функций. [...] Но если надо работать с сигналами частота которых измеряется килогерцами, то при твоём подходе надо уже возводить в степени, измеряемые тысячами. Для сигналов с частотами в области МГц -- степени, измеряемые миллионами... и т.д.
I have a vague idea of it, it was a long time ago. I remember a splendid, grey book, devoted almost exclusively to the Laplace transform. There were also sections devoted to working with lattice functions - with quite unexpected formulas in which number theory functions miraculously appeared (say, the Riemann zeta function).
As for degrees measured in thousands and millions... what's the second great limit on what? Look a dozen pages ago, it was already in this thread: in the region of t and q, denoted by Sergei, the dumb binomial expansion invariably fails, because the exponent multiplied by the addition to unity (here a value of the order q*t) is not small.
We should probably dig around for special functions...
Prudnikov, Brychkov, Marichev. Integral and series. M. Nauka. 1981.
Prudnikov, Brychkov, Marichev. Integral and series. Supplementary chapters. M. Nauka. 1986.
We know these treatises. They are horrible, but in their time they were useful, especially the second one. Only here we have a purely elementary case, it cannot be simpler...
We've been solving this one for a week already, FreeLance... let alone the right one... But the crux of the problem, it seems, has been accurately defined by Oleg:
I have a vague idea of it, it was a long time ago. I remember a splendid, grey book, devoted almost exclusively to the Laplace transform. It included sections on working with lattice functions - with quite unexpected formulas in which number theory functions miraculously appeared (say, the Riemann zeta function).
Regarding degrees measured in thousands and millions... what's the second great limit on what? Look a dozen pages ago, it was already in this thread: in the region of t and q, designated by Sergei, dumb binomial decomposition invariably limps.
We know these books. They are horrible, but at one time they were useful, especially the second one. Only here we have a purely elementary case, it cannot be simpler...
no fun in that, but - 10-30% (stable!) per month for the fora and for all other parts of the world - a GREAT miracle...
And the author has indicated that he won't keep such a "squishy" deposit in one place for a long time, so he has limited the period and degree value.
For the task of these organizational questions - where and how to generate cache, as far as I see it, do not matter - a hell of a mess.
But never mind, he who paid for dinner, that's the girl who dances.
I'll keep an eye on the rambling theme.
----
I hope that everyone understands the difference between annual and monthly rates (annual is not equal to monthly*12) in these formulas - through exponent, or effective rate chewed postnumerando...
;)
I set large t and k in the problem condition in the hope of obtaining an analytical solution. In this case I thought to hold decomposition by parameter k up to degree 3 inclusive and solve cubic equation... But, life turned out to be more complicated than usual. Even within these limits it is necessary to hold the higher powers of the expansion for acceptable accuracy.
The problem is nevertheless very interesting. It seems to have a direct bearing on optimal deposit taking in the forex market. Indeed, an optimal MM implies a profitable TS, which provides reinvestment of funds and, consequently, a constant percentage growth of the deposit (exponential growth ideally). It cannot continue indefinitely - sooner or later the account will collapse and all of the deposit funds will be destroyed. Thus, we will be left with only the withdrawn funds. And here we have a situation, when knowing reinvestment percentage rate q (it depends on expected payoff of TS) and typical life time of the deposit t we need to maximize funds withdrawal f.
It seems the problem can be solved in full only with numerical methods. As I understand my colleagues have no ideas, I suggest we stop on this note and consider the subject closed. As the dry residue of the work done, we can state that there is an analytical expression which connects together all the parameters entering into the problem with the sum of the deduced means:
If necessary, a numerical value can be obtained for the optimum withdrawal percentage k. There is also a question, how often should you withdraw funds (once a year, once a month or once a week)? If you play with parameters (of course q will change), the optimum is the most frequent withdrawal, which is limited by the percentage of withdrawals. But, this is a complication of the model (as well as introduction of inflation percentage into the formula, etc.) and requires further study, which can be left for personal digging.
I would like to express special thanks for the help and useful discussions to Oleg and Alexey.