intraday DAX strategy 5min mini1€ spread 1

Viewing 15 posts - 331 through 345 (of 503 total)

← 1 2 3 … 22 23 24 … 32 33 34 →

Author

Posts
01/31/2017 at 7:35 PM #23321 quote

Nicolas
Keymaster

Master

Could someone post exact time of different real trades taken by the ProOrder strategy? I mean the difference between backtests and live trading lies here…mainly.

01/31/2017 at 8:07 PM #23331 quote
Raul Vg
Participant

Senior
Hi Nicolas, this has been today, backtest to the left and demo to the right.

I understand that in the backtest it puts the operations just to the point, without using the close + 2 and the close-1, that is to say, that I will try putting:
```
buy n*nn contract AT market




sellshort n*nn contract AT market
```
today-3.png
01/31/2017 at 8:07 PM #23334 quote

Paris
Participant

Veteran

@ nicolas tu veux quoi exactement ?

@ nicolas can you explain what you mean ?

edit : ok you ve got it .

01/31/2017 at 8:18 PM #23338 quote

Paris
Participant

Veteran

@ raul .

in all yours system i see thoset trades most importantly the long position , it modulates only the long .

what about a system that modulates the sort positions . a kind of mirror of that system .

Could you code it ? Maybe we could run both to duplicates gains . what do you think ?

01/31/2017 at 8:21 PM #23339 quote

Raul Vg
Participant

Senior

@paris 75000

I’m sorry, but I do not understand what you want to tell me, you could give me an example.

01/31/2017 at 8:43 PM #23341 quote

Paris
Participant

Veteran

@ raul

in your 3 systems , when you buy , it is 1 , 2, 3 , 4 , 5 , 6 …..contracts and when you sell it is always 1 contract .

can you code an opposite ” twin ” …

i mean

when you sell , it is 1 , 2, 3 , 4 , 5 , 6 …..contracts and when you buy it is always 1 contract .

regards

01/31/2017 at 8:54 PM #23342 quote

Raul Vg
Participant

Senior

When he sells he also adds contracts like when he buys. Unless there has been an error of some kind, it is configured as such. In the code that I have shared with Nicolas veras that puts the same contracts in purchase and sale (n * nn)

01/31/2017 at 8:58 PM #23346 quote

Paris
Participant

Veteran

@ raul

ok , sorry for the question .

01/31/2017 at 9:23 PM #23349 quote
manel
Participant

Veteran
Hi Nicolas – Attached is an example of the difference between backtest and live (excuse the crude drawing of the boxes :-)). On the left is LIVE 10.2 Spreadbet and on the right is DEMO 10.3 Backtest (with 1pt spread). The strategy is Dax-5-min-v2 listed earlier in the topic. The main issues are :
1. Demo has 1 extra trade on the 8am bar but Live does not – This is also the case with other strategies with the first trade being missed so it is a common problem.
2. Boxed in blue are matching trades between Live and Demo, so some definitely match – Note, I have allowed for upto 3 points of market slippage in Live
3. Boxed in Red are where levels of entry do not match and in some cases by a lot
4. The trade on 08.05 on 31/01 is a perfect example where the entry trades are almost perfectly matched but the exit trade ends up being 5 points different.
5. There is a also a discrepancy on the no. of contracts traded but that could be due to the code which accumulates and adds contracts since the first trade was missed in Live but I have yet to check and verify this. It’s probably a secondary issue to the execution prices problem at the moment though.
There are other strategies which I can list which exhibit similar behaviour if you need, just let me know. Hope this helps.

Many thanks.

5m-test-1-310117.jpg
02/01/2017 at 7:40 AM #23369 quote

Nicolas
Keymaster

Master

@manel

Thanks for the well explained example, it’s valuable to find how and why dissimilarities exist. I keep this post as a clue for further investigation.

If I’m correct, the strategy still use pending stop orders to enter the market? If the answer is yes, I’d like to have the IG tickets list with each pending order number with their exact time, please.

In the picture you post, I see that the orders are mainly put at their correct prices (spread and slippage should explained a lot the differences between backtest and demo for this), the main problem is with trailing not operating the same way in live trading, are you ok with this conclusion? (we’ll find later why the first trade is not taken).

02/01/2017 at 8:26 AM #23371 quote

dreif123
Participant

Average

hi everybody,
my FB on yesterdays execution is very strange on this strategy … I tradet onliveaccount on 10,2 version with a loss ….the demoaccount on 10,2 generatet a profit.
meanwhile I am not wondering that this is happening , ….also on very similar strategys (for example simple crosses on EMA) the results are often different between live and demo ….

02/01/2017 at 9:30 AM #23378 quote

CN
Participant

Senior

Day 2 of big losses…..

It’s so stupid right now, it gets long when the dax already went 100p, stops it out, then it goes down 50pts at that time it thinks that its a great idea too short…
Maybe a RSI indicator to help the scalp decide if its a good trade or not?

02/01/2017 at 10:16 AM #23380 quote

Paris
Participant

Veteran

hi guys ….

ok another day of losses , not big but losses ……

now i ve a great news for do you , want to know what it is ?

02/01/2017 at 10:17 AM #23381 quote

CN
Participant

Senior

Tomorrow is a new day?

02/01/2017 at 10:26 AM #23382 quote

Paris
Participant

Veteran

no cn , i am not kidding , ( killing ? ) ….

look what is attached ….

Capture-du-2017-02-01-10-24-20.png
Author

Posts