Avalanche - page 303
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
Profit and drawdown in the netting variant are superior to previous variants.
Are the lots the same in "netting" and "locking" variants? In order to try to compare the two variants, they must be different (in the "netting" variant - the difference of the two previous lots of the "locking" variant).
And with a slightly different closing of the chain, than in the lot variant, the picture changes - it is possible to get/without getting into the swing.
There is no sense to fit one variant to another, but for the sake of interest I got very close results.... Although I am not so sure, that losses amount to the last order is smaller than margin of the whole chain :)
А лоты у Вас в "неттинговом" варианте и в "локирующем" одинаковые? Для того, чтобы попытаться сопоставить два варианта, они должны быть разными (в "неттинговом" - разница двух предыдущих лотов "локового" варианта).
Да и при чуть другом закрытии цепочки, чем в локовом варианте, картина меняется - можно попасть/непопасть в раскачку.
Смысла подгонять один вариант к другому не видел, но ради интереса добился очень близких результатов.... Хотя в то, что сумма убытков до последнего ордера меньше маржи всей цепочки, пока не убедился :)
I do not quite understand what you mean. By the example of lot allocation.
About "just the difference" I was wrong (my example from another thread below is also "behind the ears")
My humble request :)
Please write, which "joint order" I should open instead of . any, starting from second
2010.03.22 00:38 buy 0.01 1.35339
2010.03.22 08:01 sell 0.02 1.35026
2010.03.23 00:06 buy 0.03 1.35643
2010.03.23 08:37 sell 0.09 1.35026
After your reply, I will make sure to follow up on the 'live history' of all positions.
'
The aim of the Holivar has never been the truth! I just want to understand ... 'pro netting' for these positions. ;)
What is there to understand?
2010.03.22 00:38 buy 0.01
2010.03.22 08:01 sell 0.02 = Close 0.01 buy, Open 0.01 sell
2010.03.23 00:06 buy 0.03 = Close 0.01 sell, Open 0.02 buy
2010.03.23 08:37 sell 0.09 = Close 0.02 buy, Open 0.07 sell
---
total after 2010.03.23 08:37 0.07 sell
I too have not tried to reproduce the experiment - the last time I did it I limited myself with (conditional) 0.06 lots, but I have noticed a tendency - the sets have become closer.
If I find old versions of EAs and, more importantly, "adjusted sets" (I need to hit the same starting points of the chain), then I'll give an example of "identical lots" of the "netting" and "locking" versions.
I repeat - the interest was sporting, i.e. aimless ;)
The point of my post was that the simple replacement of the opening position by a stop with a flip will not produce identical systems. And the fact that the multiplication factor should be handled ... It is a fact.
ZS. I think I found a fragment of how EAs work
Loki
128 2008.03.07 06:58 buy stop 50 0.01 1.54414 0.00000 0.00000 0.00000 0.00 10082.83
129 2008.03.07 06:58 sell stop 51 0.01 1.53591 0.00000 0.00000 0.00000 10082.83
130 2008.03.07 14:30 buy 50 0.01 1.54414 0.00000 0.00000 0.00000 0.00 10082.83
131 2008.03.07 14:30 delete 51 0.01 1.53591 0.00000 0.00000 0.0000 10082.83
132 2008.03.07 14:30 sell stop 52 0.02 1.53591 0.00000 0.00000 0.00000 10082.83
133 2008.03.07 15:37 sell 52 0.02 1.53591 0.00000 0.00000 0.00 10082.83
134 2008.03.07 15:37 buy stop 53 0.03 1.54414 0.00000 0.00000 0.00 10082.83
135 2008.03.11 11:07 buy 53 0.03 1.54414 0.00000 0.00000 0.00 10082.83
136 2008.03.11 11:07 sell stop 54 0.05 1.53591 0.00000 0.00000 0.00 10082.83
137 2008.03.11 13:51 sell 54 0.05 1.53591 0.00000 0.00000 0.00 10082.83
138 2008.03.11 13:51 buy stop 55 0.08 1.54414 0.00000 0.00000 0.00 10082.83
139 2008.03.12 11:05 buy 55 0.08 1.54414 0.00000 0.00000 0.00 10082.83
140 2008.03.12 11:05 sell stop 56 0.12 1.53591 0.00000 0.00000 0.00 10082.83
141 2008.03.12 21:05 close 55 0.08 1.55689 0.00000 0.00000 102.00 10184.83
142 2008.03.12 21:05 close 54 0.05 1.55702 0.00000 0.00000 -105.60 10079.24
143 2008.03.12 21:05 close 53 0.03 1.55689 0.00000 0.00000 38.21 10117.44
144 2008.03.12 21:05 close 52 0.02 1.55702 0.00000 0.00000 -42.27 10075.17
145 2008.03.12 21:05 close 50 0.01 1.55689 0.00000 0.00000 12.71 10087.88
Netting
76 2008.03.07 07:51 buy stop 33 0.01 1.54522 1.53699 0.00000 0.00 1131.38
77 2008.03.07 07:51 sell stop 34 0.01 1.53699 1.54522 0.00000 0.00 1131.38
78 2008.03.07 09:21 sell 34 0.01 1.53699 1.54522 0.00000 0.00 1131.38
79 2008.03.07 09:21 delete 33 0.01 1.54522 1.53699 0.00000 0.00 1131.38
80 2008.03.07 14:30 s/l 34 0.01 1.54522 1.54522 0.00000 -8.23 1123.15
81 2008.03.07 14:30 buy 35 0.01 1.54561 1.53699 0.00000 0.00 1123.15
82 2008.03.07 15:19 s/l 35 0.01 1.53699 1.53699 0.00000 -8.62 1114.53
83 2008.03.07 15:19 sell 36 0.01 1.53697 1.54522 0.00000 0.00 1114.53
84 2008.03.11 11:11 s/l 36 0.01 1.54522 1.54522 0.00000 -8.27 1106.27
85 2008.03.11 11:11 buy 37 0.02 1.54523 1.54105 0.00000 0.00 1106.27
86 2008.03.11 13:39 s/l 37 0.02 1.54105 1.54105 0.00000 -8.36 1097.91
87 2008.03.11 13:39 sell 38 0.03 1.54095 1.54513 0.00000 0.00 1097.91
88 2008.03.12 11:14 s/l 38 0.03 1.54513 1.54513 0.00000 -12.57 1085.34
89 2008.03.12 11:14 buy 39 0.04 1.54516 1.54098 0.00000 0.00 1085.34
90 2008.03.13 04:26 close 39 0.04 1.55800 1.54098 0.00000 51.19 1136.53
I will not publish the "hats" as the purpose of this sample was to illustrate the "red"
SZY. My charts are less nice than yours, so the above is just an illustration for the "systems identity" thesis, nothing more.
I have a build-up pattern of 0.02 0.06 0.18 in both variants ...
In this example the coefficient is 3, but it could be different. I don't adhere to any strict recommendations of the build-up scheme given in the 1st avalanche post.
It doesn't play a special role.
I have a build-up pattern of 0.02 0.06 0.18 in both variants ...
The same lots for the netting version will give a completely (or slightly :) ) but different result.
In this example the coefficient is 3
I have one of the "demo variants" spinning with the following parameters
starting coefficient 2 at the fifth iteration is coefficient 4, and after the seventh iteration it is 1
I was searching for long periods of history when the price was "jumping" for such a long time.
The same is with the channel width, but not so obvious.
I don't adhere to any strict recommendations of the build-up scheme given in the 1st avalanche post.
And I didn't read any "lot requirements" in the first post at all, just "examples". :)
Also, in the report cap that you posted a few pages ago, there is an option that is not too much of a drawdown:
A variant with not too much drawdown.
I apologize for not answering some of the questions of the forum visitors, I have no time to theorize, I'm busy with practical work.
Do you think that the question of mate expectation in points is a purely theoretical question???
Let's agree then, that all the disadvantageous for you questions, you will refer to the theoretical and will not answer them.
And you will be engaged in pure practice, from time to time posting pictures from the tester.
I will stop asking inconvenient questions, and there will only be silence and graphs steadily creeping upwards.
But it turns out - "Theatre of one actor".
Те же лоты для неттинговой версии дадут совершенно (или немного :) ), но другой результат.
Один из "демо-вариантов" у меня крутится с такими параметрами
старртовый коэффициент 2 на пятой итерации - коэффициент 4, а после 7ой - 1
Я долго искал участки на истории, когда цена "скакала" так долго.
Аналогично и с шириной канала, но не так наглядно.
А я вообще не прочитал в первом посте никаких "требований на лоты", только "примеры". :)
Can you explain why, it's not clear to me.
And this is in the 1 post
First turn - 0.01 / 0.02Second turn - (0.01 +0.03) / 0.02
Third turn - (0.01 +0.03) / (0.02 +0.06)
Fourth U-turn - (0.01+0.03+0.12) / (0.02+0.06)
Turn five - (0.01+0.03+0.12) / (0.02+0.06+0.24)
Isn't it a lot build up scheme?
Sorry, but I'm not posting the results for theoretical research, but for those programmers who are developing similar EAs to have a reference point.
For example, if I see that someone else's drawdown is smaller, it means that I am sure I can decrease it too. In the same way, we can compare and draw conclusions by other parameters. You must admit that when you don't have the information about their results from your competitors, you may think that your variant is the best and there is no sense in following it up. And what can the information about the expected payoff, or in fact, the average profit of a deal give me? For example, we have 2 extreme variants, 1. IR is small, but there are many trades 2. IR is large, but there are few trades. Conclusion with very different IR, the net profit can be the same. So, it is not necessary to strive for it to be large.
Sorry, but I'm not posting the results for theoretical research, but for those programmers who are developing a similar EA, in order to have a reference point.
For example, I see someone's drawdown is smaller, so I'm sure I can get it reduced, too. In the same way we can compare and draw conclusions on other parameters. If you have no information from your competitors about their results, you may think that your variant is the best and there is no point in continuing your work. And what can the information about the expected payoff, or in fact, the average profit of a deal give me? For example, we have 2 extreme variants, 1. IR is small, but there are many trades 2. IR is large, but there are few trades. Conclusion with very different IR, the net profit can be the same. So, it is not necessary to strive for it to be big.
1) You don't need anything, and that's good. Give it to us, we need it more. ))
2) I would like to ask you to post two tests with different models.
It's a matter of two minutes.....
1) Вам ничего, и хорошо. Вы нам дайте, нам оно нужнее. ))
2) И очень просил Вас выложить два теста с разными моделями
Это дело двух минут.....
Explain what information you can get from the expectation value.
Excuse me, but I will spend my free time doing more useful things.