Optimal values of SL and TP orders for an arbitrary TS. - page 12
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
For me the left picture here and the reasoning related to it looks superfluous, imho, one can directly start the reasoning from the right picture and the "optimization" of one deal. Especially because I, for example, would expect to see something like a distribution of zigzag segments by amplitude in the left picture, and this is a completely different diagram
P.S. This is of course my personal, highly subjective perception
It should be understood that f is the fraction of capital involved in the trade (which is how the author of the topic defined this variable). So the range of its definition is from 0 to 1 for arbitrary conditions (depo and point value).
My comments are primarily aimed at helping the author. I hope they will still be useful to him.
I don't doubt your willingness to help. However, if your comments are not well founded, then the author will have to look for your mistakes and explain them to you. So if you really want to help, you need to justify your comments. So the author won't have to do extra work.
For example, the one I have highlighted is an obvious mistake. You could have understood it from my post, but you didn't pay attention to it. As a result the author had to explain it again. Got it now ?
In my version, f this is the fraction of capital per point for the instrument on which we are working. So, one pip may account for a fraction of 1% of the deposit or even less (to be realistic). For example, when trading EURUSD at 100 leverage and 0.1 lot ($100 deposit), for one pip movement of a quote you would have $1, i.e. 1% of your funds.
Sergey, it seems to me that your definition of parameter f (underlined) needs to be clarified. I understood you, but it seems that others didn't. Maybe it would be better to say "the share of capital that is one point of profit"? Then it would be clear that the capital itself is in the denominator and the parameter f must satisfy the inequality I gave.
One point of profit is the product of the value of one point by the number of lots. Obviously, the capital itself cannot be less than the sum needed in the account to buy this number of lots. Hence the inequality and capital in the denominator.
Of course I do!
You're welcome.
So, we assume that our pricing process is random, i.e. similar to a one-dimensional Brownian motion. According to Einstein's law (it can be derived strictly if you wish), the square of the projection of the point displacement on the ordinate axis is proportional to the time t. Then, the amplitude (without taking into account the direction) of the price V depends on the time of holding an open position t, the volatility of the instrument V0 at timeframe t0, as:
Profitability per transaction is determined by the difference between this amplitude and spread (for the ideal TS all directions are guessed). Let's define profitability as the number of pips per unit time:
Assume that the type of price change function has the form of a cosine. At the beginning of the position we are at 0. Then for the full period of the cosine the amplitude will be 1 and the profit (profit) will be 0.
I think the rigid connection of the terms "amplitude" and "profit" for trading operations can be very nuanced.
...However, if your comments are not sufficiently substantiated, the author will have to find your mistakes and explain them to you. So if you really want to help, you need to justify your comment. So the author doesn't have to do extra work...
Any discussion can only take place if there are at least two opinions. The purpose of a discussion can be to explore the differences between opinions and bring them into some kind of common ground. So in any discussion, all parties are equal: I look for the author's mistakes and explain them to him, while he (or any other party involved in the discussion) looks for my mistakes and explains them to me.
Сергей, мне кажется, что твое определение параметра f (подчеркнуто) нужно уточнить. Я-то тебя понял, а вот другие судя по всему нет. Может быть лучше "доля капитала, которую составляет один пункт прибыли" ? Тогда будет понятно, что сам капитал стоит в знаменателе и параметр f должен удовлетворять неравенству, которое я привел.
Один пункт прибыли это произведение стоимости одного пункта на количество лотов. Понятно, что при этом сам капитал не может быть меньше того, что необходимо иметь на счету, чтобы купить это количество лотов. Отсюда и неравенство, и капитал в знаменателе.
Thanks, Yura. This is really more correct. I have been dealing with this for a long time and sometimes the obvious seems complicated to me, and vice versa. That's why such blunders arise.
Candid wrote(a) >>
For me the left picture here and the reasoning related to it looks superfluous, imho, one can directly start the reasoning from the right picture and the "optimization" of one deal. Moreover, I, for example, would expect to see something like a distribution of zigzag segments by amplitude in the left picture, and this is a completely different diagram
I thought the reasoning looked more understandable this way... Anyway, I agree.
Avals wrote(a) >>
Neutron, it fairly follows from your reasoning that TP and SL levels depend on position holding time - topt. From which it follows that optimal TP and SL should be dynamic depending on how much time left to topt. I.e. toptoppos. But then again not for all systems.
No, it shouldn't. Well, actually, you should, but indirectly, and we shouldn't necessarily look back on it, because I only use protective orders as protection for market force majeure events. Therefore, events are extremely rare and perhaps of little significance. I should not make any deep inferences from the above chart with illustrative TP. The truth is in integrals and when the general view of FR for optimal TS appears, these parameters will become dynamic and we will search for the optimum of profitability in the space of these parameters.
ystr wrote(a) >> Let's assume that the form of the price change function is cosine. Then for the full period of the cosine the amplitude will be 1 and profit will be 0. I think the rigid connection of the terms "amplitude" and "profit" for trading operations may contain a lot of nuances.
This example is not very correct, because we must use the obvious properties of a price series in our analysis. Random Brownian motion and martingian property are the closest to describing them. What cannot be said about a harmonic series like cos(x).
We, ystr, speak in terms of averages. Taking this approach, my statement that the price travels a path V in time t is quite correct and provable strictly. Let me remind you that I'm not talking about the average of the points the price has traveled (that's zero), but about the average of the distance modulus in points that the price has gained (traveled). Different things.
... уровень TP и SL зависят от времени удержания позиции - tопт.
If this is a cause-and-effect construction, then cause and effect should be reversed, the time to hold a position depends on TP and SL. If SL is low, the position will not reach the topt on average, and if SL is high, it will overstay. So, the aiming at the topt requires just a fixed SL, corresponding to the value of the topt.
If this is a cause-and-effect construction, cause and effect should be reversed. If the SL is low, the position will not reach topt, and if the SL is high, it will over sit. That is, the sight on the topt requires just a fixed SL, corresponding in magnitude to the topt.
I think that the reason for most systems is optimal position holding time, and SL and TP are consequences of it. In my opinion, the reason for most systems is the optimal time of position holding, and TP is a consequence of it. Practically every system has such a time after which holding the position makes no sense.
If we know that the pose is held for 24 hours and the initial TP was a figure. If we know that the initial TP was a figure - it is stupid to leave it the same when the position has been kept for one hour and it is a mile away from China.) It is almost always possible to find new TP levels, when due to volatility and positive system MO the result will be much better than without changes. Of course, the choice must be systematic and confirm the advantage and robustness on tests.
Если это причинно-следственная конструкция, то причину и следствие следует поменять местами, время удержания позиции зависит от TP и SL. При малом SL позиция в среднем не дотянет до tопт, а при большом пойдёт пересиживание. То есть прицел на tопт требует как раз фиксированного SL, соответствующего по величине tопт.
You should not be discussing the result for t opt so intensely. This is a model for an "ideal" TS, which cannot be implemented in practice. These conclusions are true only for highly predictable instruments. And I needed it as an auxiliary model for further constructions.You should not be discussing the result for t opt so intently. This is a model for an "ideal" TS which cannot be implemented in practice. These conclusions are true only for highly predictable instruments. And I needed it as an auxiliary model for further constructions.
can be realised in practice. Moreover, for real systems where the time in position is unlimited in advance, the use of TP is not reasonable at all. Examples of this are trend following systems and "let the profit grow" rules. imho.