Gamestudio Links
Zorro Links
Newest Posts
rMomentum always returns 0 when checking for peak/valley
by jcl
Today at 17:38
a little help here?
by 3run
Today at 16:38
How to set a pair trading order right
by jcl
Today at 11:46
Dual Momentum Algorithm - The way Zorro would have done it
by Hredot
Today at 05:21
HIgh FRequency Trading Script -- Highly Profitable
by Yosh
Yesterday at 19:27
Linear Regression in objective function
by jcl
11/22/17 15:40
MT4 vs Zorro
by pascalx
11/22/17 14:00
AUM Magazine
Latest Screens
RPG PARTY
Triton Wing now available on Steam
AckCon'17: Lotter vs The World
Triton Wing
Who's Online
14 registered (Lobo, Hredot, painkiller, 3run, alibaba, Kartoffel, FLD, RealSerious3D, Jerome8911, 1 invisible), 169 Guests and 4 Spiders online.
Key: Admin, Global Mod, Mod
Newest Members
ech87, Lobo, Andy60, AzzoCorp, franzz
17813 Registered Users
Page 1 of 3 1 2 3 >
Topic Options
Rate This Topic
#469326 - 11/11/17 00:44 Training Z strategy with different AssetsFix.csv
DdlV Offline
Serious User

Registered: 06/10/13
Posts: 1293
Hi all,

I'm probably not remembering something right, so would appreciate help/clarification!

Using Z3 since it Tests faster. laugh I install 1.66.5, and Test Z3. Results are per the manual.

I replace AssetsFix.csv with my own Broker's parameters and reTest. Results are different, as expected - unacceptably worse, in fact.

I Train Z3 and it downloads data and creates a new Z3.par. I reTest expecting better results, as presumably Z3 has now been Trained to my AssetsFix.csv. However, the Test results are exactly the same as before Training.

What am I missing? Why did the Train not improve the Test results? Is there no way to Train to my AssetsFix.csv?

Thanks.

Top
#469328 - 11/11/17 01:45 Re: Training Z strategy with different AssetsFix.csv [Re: DdlV]
JoFo Offline
Newbie

Registered: 07/03/17
Posts: 24
Loc: Bangkok
For me the Z Systems generate random trades and can only be profitable with a tuned Assetsfix what have nothing to do with reality.

Money for nothing.

Top
#469330 - 11/11/17 12:44 Re: Training Z strategy with different AssetsFix.csv [Re: JoFo]
jcl Online

Chief Engineer

Registered: 07/22/00
Posts: 25939
Loc: Frankfurt
Test results are not different when you modify Z3.par. Z3.par is for live trading, the backtest uses Z3_1.par .. Z3_12.par. - The only asset parameter with a possible effect on Z3 training is Rollover. But then the Rollover must be rather extreme - better check the entered parameters.

Top
#469338 - 11/11/17 16:03 Re: Training Z strategy with different AssetsFix.csv [Re: DdlV]
MatPed Offline
Senior Member

Registered: 02/28/15
Posts: 379
Loc: Milano, Italy
Hi Ddlv,
I trade with Oanda and have the same divergent results on Z systems. I guess that the main issue comes from the FAC parameters calculated from other brokers .

I do not find z systems very useful, without having the opportunity to fully train and re-train from beginning and the freedom to modify the ENDDATE.

I hope this will change soon or later.

Ciao

Top
#469343 - 11/11/17 20:22 Re: Training Z strategy with different AssetsFix.csv [Re: MatPed]
DdlV Offline
Serious User

Registered: 06/10/13
Posts: 1293
Thanks all!

@jcl:
Thanks for the clarification. Is this in the manual somewhere? Sorry if I missed it.

Followups:
Do I understand correctly that there is therefore no way to verify the results produced by the Train? I.e., what Trade results would be expected using the new Z3.par and my Broker's values?

My AssetsFix.csv came from the Download script without any changes - are Roll values still suspect in that case?

@MatPed:
Agreed. The Z Systems have yet to completely satisfy my Proof-Of_Concept perspective. But (aside from Z5 frown ) they haven't completely crashed and burned yet either...

How long have you been on Oanda? Has it been an OK/good experience?

Thanks again all.

Top
#469361 - 11/12/17 15:24 Re: Training Z strategy with different AssetsFix.csv [Re: DdlV]
MatPed Offline
Senior Member

Registered: 02/28/15
Posts: 379
Loc: Milano, Italy
Originally Posted By: DdlV
Thanks all!

How long have you been on Oanda? Has it been an OK/good experience?

Thanks again all.


One year, mor or less. So far so good, but tts a MM broker not the best... No ETFs, Stocks, Futures and Options are the real downside...

Top
#469385 - 11/14/17 07:03 Re: Training Z strategy with different AssetsFix.csv [Re: MatPed]
DdlV Offline
Serious User

Registered: 06/10/13
Posts: 1293
Hi jcl. Would appreciate your comments on the followups:

Do I understand correctly that there is therefore no way to verify the results produced by the Train? I.e., what Trade results would be expected using the new Z3.par and my Broker's values?

My AssetsFix.csv came from the Download script without any changes - are Roll values still suspect in that case?

Thanks.

Top
#469387 - 11/14/17 11:52 Re: Training Z strategy with different AssetsFix.csv [Re: DdlV]
jcl Online

Chief Engineer

Registered: 07/22/00
Posts: 25939
Loc: Frankfurt
Depends on whether you train or retrain. Retraining affects only Z3.par. Full Training affects all .par files. Of course you will normally still get the same backtest results, since you're using the same historical data.

I do not think that a different AssetsFix can have a noticeable effect on training. Except maybe in extreme cases, such as a very different Rollover, by factor 10 or so. In such a case the training would possibly generate parameters that close trades a bit earlier. If one of your AssetsFix values is completely wrong, you'll get bad results in the backtest. But the problem should be normally easy to spot in the asset list and the performance report.

Top
#469389 - 11/14/17 18:05 Re: Training Z strategy with different AssetsFix.csv [Re: jcl]
DdlV Offline
Serious User

Registered: 06/10/13
Posts: 1293
Thanks jcl.

It appears that the Z systems (at least Z3) can only be reTrained - whether Trading or not, pressing Train only updates Z3.par. Is this as expected? Or is there some way to full Train Z3?

Thanks.

Top
#469431 - 11/15/17 17:40 Re: Training Z strategy with different AssetsFix.csv [Re: DdlV]
jcl Online

Chief Engineer

Registered: 07/22/00
Posts: 25939
Loc: Frankfurt
You're right. They have no difference in the code between training and retraining.

Top
Page 1 of 3 1 2 3 >



Gamestudio download | chip programmers | Zorro platform | shop | resources | magazine | manual | support faq | bugs

oP group Germany GmbH | Birkenstr. 25-27 | 63549 Ronneburg / Germany | info (at) 3dgamestudio.net