What makes an unsteady graph unsteady or why oil is oil? - 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
I'm definitely writing on the wall.
Go on, you're doing a good job. Maybe when you reread our discussion you will understand what I was talking about and find your mistake.
Случайное блуждание это мартингейл, соответственно, мат.ожидание для него равно не нулю, а текущей цене. Естественно, если тупо сдвинуть ЦБ на размер текущей цены, т.к. она уже свершившийся факт и ни на что не влияет, то МО для любого момента в будущем будет ноль.
Well, to be quite precise, yes - I totally agree.
Продолжайте, у вас хорошо получается. Может быть, когда вы перечитаете наше обсуждение, вы поймете о чем я говорил и найдете свою ошибку.
Reread what, your questions to me?
У меня такой вопрос ко всем участникам этой ветки: Какими статистическими свойствами цены должны обладать, чтобы их можно было предсказать или прибыльно торговать?
Even though the ACF of the first difference is deltaic, this does not prove the independence of the increments. But the opposite - their dependence - can be proved in several ways. To be sure of this and not to return to this question, I did the following:
1. I calculated a distribution of frequencies on a certain timeframe TF1 - it has been verified more than once on this forum that it is exponential (or at least very close to it).
2. the sum of two adjacent increments is in itself an increment of time series twice as big as TF2=2*TF1, frequency distribution of which we can evaluate by the same method as in paragraph 1.
3. if increments of TF1 series are independent, theory requires (remember 1 semester theorist), that TF2 distribution is exactly a convolution of TF1 distribution with itself.
4. Everyone can check that item 3 is not true, at least by empirical calculation of the convolution from available data, or by checking on paper, by presenting an exponential distribution in the integral.
5. It follows from items 1-4 that the increments of series TF1 are dependent.
У меня такой вопрос ко всем участникам этой ветки: Какими статистическими свойствами цены должны обладать, чтобы их можно было предсказать или прибыльно торговать?
Naturally, SB is predictable, since p = 1 - q = const
If p = q, then there is a proven theorem about returning a point to the initial (any historical) value with probability 1. Therefore, any channel can be drawn, through the levels already present on the history and bluntly following the counter-trend inside the channel. The arc sine theorem should also be taken into account, i.e. prices may take quite a long time to return to the initial levels. And time intervals between returns will systematically increase.
If p <> q, then bluntly bet towards the higher probability. see the problem of player ruin.
MO in case of SB is a constant and it equals MO = p - q
The prices of financial instruments, unfortunately, are not SB since p = 1 - q <> const, MO <> const, and there are gaps and other uncertainties.
2. сумма двух соседних приращений является сама по себе приращением временного ряда вдвое большего таймфрейма ТФ2=2*ТФ1, распределение частот которого мы можем оценить тем же методом, что и в пункте 1.
Естественно, что СБ предсказуемо, т.к. p = 1 - q = const
Если p = q, то есть доказанная теорема о возвращении точки в начальное (любое историческое) значение с вероятностью 1. Следовательно можно отрисовать любой канал, через уже имеющиеся на истории уровни и тупо по контртренду внутри канала...
Yeah, predictable. It will be out there somewhere...
and with probability 1 it will come back, but only a small addition: n -> infinity. That is, if you have eternity to spare, you can wait out any drawdown. And if you draw levels on history, do not forget that the bell constantly expands and goes to infinity.
"Trends, levels, channels" in a random rambling will be stronger than Goethe's Faust, it's even cooler than oil at 12 bucks a barrel.
Ага, предсказуемо. Оно будет где-то там...
и с вероятностью 1 вернётся обратно, но только маленькое дополнение: n -> бесконечность. Т.е. если у тебя в запасе вечность, то можно переждать любую просадку. И если будешь рисовать уровни на истории, то не забудь что колокол постоянно расширяется так же уходя в бесконечность.
"Тренды, уровни, каналы" в случайном блуждании это по-сильнее "Фауста" Гёте будет, это даже круче, чем нефть по 12 баков за бочку.
Have a rest. We already know your opinion. But we are not interested in it. We are only interested in proven theorems, not in the selves of an illiterate dissident.
Actually the average number of returns - ret to the starting point of 2*n ticks for SB with p = q is calculated by the formula:
ret = sqrt(2 * n) / sqrt(2 * Pi)
I.e. there will be approximately 4 ret for 100 ticks, 40 ret for 10 000 ticks, and 400 ret for 1 000 000 ticks.
In 100 ticks the point will reach the border that is 10 pips away from the initial level in any direction, in 10 000 ticks it will be approximately 100 pips, in 1 000 000 ticks it will be approximately 1000 pips. I.e. the time between returns grows squared, while the distance doesn't exceed the square root. It doesn't exceed because the point can not only move away from the initial level, but also approach it with the same speed, i.e. it can easily return to the initial level and reach the border on the opposite side. Therefore, MO = p - q = 0
Отдыхайте. Мы Ваше мнение уже знаем. Но оно нас не интересует. Интересуют только доказанные теоремы, а не отсебятина безграмотного диссидента.
And the last thing I will say.
On forex 0 and infinity are rare...
And the indices have to be watched simultaneously through the inverse pair (USDEUR for example).
Otherwise - coordinate dependence...
;)