| Pages:
1
..
27
28
29
30
31
..
98 |
saycem
Junior Member
Posts: 48
Registered: 13-7-2018
Member Is Offline
Mood: No Mood
|
|
I have 4 strategies that are using setexitonclose trading live from GSB for the last couple of months and I have not seen an issue, so I'm confused as
I was unaware of other options. What are you using if not setexitonclose?
|
|
|
admin
Super Administrator
Posts: 5060
Registered: 7-4-2017
Member Is Offline
Mood: No Mood
|
|
Quote: Originally posted by saycem  | | I have 4 strategies that are using setexitonclose trading live from GSB for the last couple of months and I have not seen an issue, so I'm confused as
I was unaware of other options. What are you using if not setexitonclose? |
If thats the case then fine.
It may depend on your execution method.
|
|
|
appengineer
Junior Member
Posts: 61
Registered: 8-4-2019
Member Is Offline
Mood: No Mood
|
|
| Quote: Originally posted by saycem | | I have 4 strategies that are using setexitonclose trading live from GSB for the last couple of months and I have not seen an issue, so I'm confused as
I was unaware of other options. What are you using if not setexitonclose? |
I am using this to manually close all trades at 1500 for ES
sessionEnding = CalcTime(SessionEndTime(0,1), 0);
If Time >= sessionEnding and MarketPosition(0) <> 0 Then
|
|
|
admin
Super Administrator
Posts: 5060
Registered: 7-4-2017
Member Is Offline
Mood: No Mood
|
|
| Quote: Originally posted by appengineer | | Quote: Originally posted by saycem | | I have 4 strategies that are using setexitonclose trading live from GSB for the last couple of months and I have not seen an issue, so I'm confused as
I was unaware of other options. What are you using if not setexitonclose? |
I am using this to manually close all trades at 1500 for ES
sessionEnding = CalcTime(SessionEndTime(0,1), 0);
If Time >= sessionEnding and MarketPosition(0) <> 0 Then
|
that should work fine.
|
|
|
LucaRicatti
Junior Member
Posts: 16
Registered: 12-10-2019
Member Is Offline
|
|
Hi,
I'm evaluating GSB and it's impressive. I would like to test forex systems but got aggregate exceptions errors few seconds after hitting Play (ES and
stocks are ok).
Is there a working forex setup to share?
Thanks
|
|
|
admin
Super Administrator
Posts: 5060
Registered: 7-4-2017
Member Is Offline
Mood: No Mood
|
|
| Quote: Originally posted by LucaRicatti | Hi,
I'm evaluating GSB and it's impressive. I would like to test forex systems but got aggregate exceptions errors few seconds after hitting Play (ES and
stocks are ok).
Is there a working forex setup to share?
Thanks |
Thanks for your comments.
this is fixed in 55.61. There was no volume in your price data
https://trademaid.info/forum/viewthread.php?tid=227#pid3769
I will get the build to you in the next 24 hours after its had some testing
GSB 2.0 is going to be working better in for forex. We will add pattern filters etc
|
|
|
LucaRicatti
Junior Member
Posts: 16
Registered: 12-10-2019
Member Is Offline
|
|
Many thanks, added the volume to my data and now it's running.
If you need help in testing beta or forex I got Multicharts/Interactive Brokers data subscription (and time to devote)
|
|
|
admin
Super Administrator
Posts: 5060
Registered: 7-4-2017
Member Is Offline
Mood: No Mood
|
|
| Quote: Originally posted by LucaRicatti |
Many thanks, added the volume to my data and now it's running.
If you need help in testing beta or forex I got Multicharts/Interactive Brokers data subscription (and time to devote) |
Your welcome to give me a list of additional indicators you want added. Pattern filters are more important for forex, but we have a list of them to be
done already. Ive emailed you 55.63 version
|
|
|
Sten
Junior Member
Posts: 35
Registered: 25-10-2019
Member Is Offline
|
|
I tried to run an optimization process in GSB 1.0.54.97 / 2019-08-31 (latest available trial version). And what I notice is that when optimization
process goes on and I change a system by selecting, let's say, OOS column in the Unique Systems window, I get an exception (probably any column except
"ID" produces an exception, and exceptions also occurs after optimization process ends).
See the screenshot below.
Is it a known issue?
Another observation is that after some number of exceptions GSB starts to function incorrectly - does not show workers running in Resource Monitor
window, optimization process never ends e.t.c. I was not even able to complete the full methodology as described on the help page (stuck somewhere on
Walk-Forward Optimization) because of these exceptions and incorrect product functioning after them. I am not sure, maybe this was caused by a lack of
RAM or pretty old Windows Server 2008 R2 x64 OS. I now try to run the process again on Windows 10.
But exceptions, when I switch a system in a Unique-Systems window, do occur on Windows 10 as well. And they are quite annoying. Is it possible to fix
them?
I attached a .txt file with one of these exceptions from C:\GSB\Exceptions folder.
Attachment: Login to view the details
|
|
|
admin
Super Administrator
Posts: 5060
Registered: 7-4-2017
Member Is Offline
Mood: No Mood
|
|
| Quote: Originally posted by Sten | I tried to run an optimization process in GSB 1.0.54.97 / 2019-08-31 (latest available trial version). And what I notice is that when optimization
process goes on and I change a system by selecting, let's say, OOS column in the Unique Systems window, I get an exception (probably any column except
"ID" produces an exception, and exceptions also occurs after optimization process ends).
See the screenshot below.
Is it a known issue?
Another observation is that after some number of exceptions GSB starts to function incorrectly - does not show workers running in Resource Monitor
window, optimization process never ends e.t.c. I was not even able to complete the full methodology as described on the help page (stuck somewhere on
Walk-Forward Optimization) because of these exceptions and incorrect product functioning after them. I am not sure, maybe this was caused by a lack of
RAM or pretty old Windows Server 2008 R2 x64 OS. I now try to run the process again on Windows 10.
But exceptions, when I switch a system in a Unique-Systems window, do occur on Windows 10 as well. And they are quite annoying. Is it possible to fix
them?
I attached a .txt file with one of these exceptions from C:\GSB\Exceptions folder.
|
this isnt a known issue and the exception doesnt help. Often exceptions can be ignored.
Im going to email you a possible fix
Peter
|
|
|
Sten
Junior Member
Posts: 35
Registered: 25-10-2019
Member Is Offline
|
|
Peter, thanks! The build you e-mailed me fixes the exception.
I was not able to repeat the exception while running optimization process.
Now I am trying to walk through all the steps from Methodology and see if it works. I'll let you know if I find other issues.
By the way, I have a question. The methodology document says:
https://trademaid.info/gsbhelp/Provingthemethodology.html
------
2) Select the best systems
Change nth to all. Dates are still pre 20150630. <---
Sort on NP/DD and get the top roughly 2000 systems, put into favorites A.
------
When I change nth to all, GSB asks me if I want to re-backtest all selected systems or not. Do I understand correctly that I should re-backtest
systems so that GSB recalculates systems stats after I change nth to all?
|
|
|
admin
Super Administrator
Posts: 5060
Registered: 7-4-2017
Member Is Offline
Mood: No Mood
|
|
| Quote: Originally posted by Sten | Peter, thanks! The build you e-mailed me fixes the exception.
I was not able to repeat the exception while running optimization process.
Now I am trying to walk through all the steps from Methodology and see if it works. I'll let you know if I find other issues.
By the way, I have a question. The methodology document says:
https://trademaid.info/gsbhelp/Provingthemethodology.html
------
2) Select the best systems
Change nth to all. Dates are still pre 20150630. <---
Sort on NP/DD and get the top roughly 2000 systems, put into favorites A.
------
When I change nth to all, GSB asks me if I want to re-backtest all selected systems or not. Do I understand correctly that I should re-backtest
systems so that GSB recalculates systems stats after I change nth to all?
|
You can change dates, nth use wf params, overrideOrignal settings....etc
the important thing is at the last change you hit yes. You can hit yes after every change, but that takes longer. The re-calculations are only hit
when you do yes. So can see this if you edit the stats macro too.
|
|
|
Sten
Junior Member
Posts: 35
Registered: 25-10-2019
Member Is Offline
|
|
Here is another issue. I'm using GSB 1.0.55.90 / 2019-11-07 and I started 1000 walk-forward tests from Favorite-A, sent them to the cloud (as
described in the Methodology). And left Manager and workers working for the night.
Next morning I see the following picture in the GSB Manager in Unique Systems window:
Some systems have completed WFO tests, some are Waiting, and few are running. That is ok.
However, when I look into Manager's Resources Monitor window I see this:
It shows Stubs are stopped, "Optimization/Status" tree is empty (I use this to monitor progress and see how worker are loaded with the tasks). User
"Tasks / Active" tree is also empty, if I try to expand it, it shows "Active" contains empty list, and then withing a second becomes "> Active" so I
can expand it again - but in reality "> Active" tree does not contain any user tasks.
The only thing that shows that workers are loaded with the tasks is "Optimization / Speed" section. I can see here that there are some workers which
are doing something and GSB calculates their speed.
So the problem is that after some time when WFO tests are running, GSB Manager shows in the status that all stubs are stopped. So it becomes
uncomfortable to monitor the task progress.
Here is how it looks on one of the worker machines, worker's Resource Monitor window:
We can see that worker load's "Sys. CPU" and that suggests it is actually doing something. At the same time in the "Optimization / Status" on the
worker shows "Stopped" status.
And Unique System's window on the worker shows:
that worker is actually processing some WFO tests. So, once again, the "Optimization / Status" is displayed incorrectly.
The numbers in the WFS column are increasing (both on the Manager and in the Worker machines). So it looks like the system is doing something useful -
WFO calculation goes on, so I'm leaving it working for some more time.
But it would be nice to fix this bug with incorrect Status for workers. As now I do not feel comfortable when I try to monitor the task progress - I
always ask myself a question, if GSB still works, or has stuck somewhere in the middle of the process, and doing nothing.
|
|
|
admin
Super Administrator
Posts: 5060
Registered: 7-4-2017
Member Is Offline
Mood: No Mood
|
|
Hi Sten
The methodology is showing 250 wf, not 1000.
There is a limit on the free cloud of 10?? or so wf at a time, so 1000 is going to take a long time.
(In the past if a user did 1000 wf, it would take over the entire free public cloud at the expense of all other gsb users
Also when I use the cloud, most gsb free workers will be killed.
The last week I was on leave, so cloud resources were high, but low the last 48 hours. (better now as im not using it)
Your confusing wf status with building system status. The manager or worker resource status have no correlation with WF.
So like the wf that show 39 to 40 % are doing something, but if there workers have been killed, they will eventualy be resubmitted.
If you dont have a I9 (or fast i7) good to hire one for $10 a day from me.
I bought 5 more servers in the last few months.
|
|
|
Sten
Junior Member
Posts: 35
Registered: 25-10-2019
Member Is Offline
|
|
Hi Peter,
Thanks, I removed 750 systems and did WFO for the remaining 250 ones. I also added dual Xeon server to my workers pool.
I was able to finish the Methodology process, however, it looks I'm doing something wrong:
Unlike in the documentation, I see some performance degradation from D to C stats (-3.5%). And a big degradation from F to E (-51.6%). So on Out of
Sample period strategies performance degrades significantly.
I think I was careful enough to implement Optimization settings described in the "Proving the Methodology" and "Updates" sections. It is possible I
misunderstood some items, or there are differences in the settings described in the documentation and in the latest GSB version.
I attached my optimization settings:
Attachment: Login to view the details
The biggest difference that I use IQFeed data for my process as I do not have access to TradeStation data. From my experience many workflows do not
work as good on IQFeed continuous futures data when ported from TradeStation.
So, I do not know whether I do something wrong in GSB, or it is just the result of data coming from different data provider. It would help if you look
at my optimization settings.
----
Another observation - when I run wf_stats macro and then look at Graph for some strategy I see something like this on the chart:
The big gap in P/L on the chart. I think this is the result of Trading Dates Mode being set to NoTrd after macro finishes. If I change Trading Dates
Mode to All, the chart looks ok. But I wonder if this disrepancy on the chart breaks statistics (for example, does GSB calculates NetProfit correctly
for Dates = NoTrd case?).
I tried to modify the wf_stats macro and change the date in the latest GlobalDates command from 12/31/2100 to 02/28/2018 so that to eliminate this
2019 OOS data completely. But then D/C and E/F statistics become even worse. So it looks like this OOS 2019 part of the P/L curve is not responsible
for the stats degradation.
|
|
|
admin
Super Administrator
Posts: 5060
Registered: 7-4-2017
Member Is Offline
Mood: No Mood
|
|
| Quote: Originally posted by Sten | Hi Peter,
Thanks, I removed 750 systems and did WFO for the remaining 250 ones. I also added dual Xeon server to my workers pool.
I was able to finish the Methodology process, however, it looks I'm doing something wrong:
Unlike in the documentation, I see some performance degradation from D to C stats (-3.5%). And a big degradation from F to E (-51.6%). So on Out of
Sample period strategies performance degrades significantly.
I think I was careful enough to implement Optimization settings described in the "Proving the Methodology" and "Updates" sections. It is possible I
misunderstood some items, or there are differences in the settings described in the documentation and in the latest GSB version.
I attached my optimization settings:
The biggest difference that I use IQFeed data for my process as I do not have access to TradeStation data. From my experience many workflows do not
work as good on IQFeed continuous futures data when ported from TradeStation.
So, I do not know whether I do something wrong in GSB, or it is just the result of data coming from different data provider. It would help if you look
at my optimization settings.
----
Another observation - when I run wf_stats macro and then look at Graph for some strategy I see something like this on the chart:
The big gap in P/L on the chart. I think this is the result of Trading Dates Mode being set to NoTrd after macro finishes. If I change Trading Dates
Mode to All, the chart looks ok. But I wonder if this disrepancy on the chart breaks statistics (for example, does GSB calculates NetProfit correctly
for Dates = NoTrd case?).
I tried to modify the wf_stats macro and change the date in the latest GlobalDates command from 12/31/2100 to 02/28/2018 so that to eliminate this
2019 OOS data completely. But then D/C and E/F statistics become even worse. So it looks like this OOS 2019 part of the P/L curve is not responsible
for the stats degradation. |
Hi Sten,
I can see numerous causes for this.
The biggest being 6 entry modes enabled, not just the entry cross dual level
Your verification needs to be 30,29,31 (vip 30 is first)
What I cant see in your settings, is entry level weights -10,10,1 in parms and wfparpms. The field doesnt exist when I use your settings
autoNth must also be set to all. This has been moved from app settings to opt settings in a recent build.
All of this is going to help a lot.
What you see in the graph was ok
 
|
|
|
Sten
Junior Member
Posts: 35
Registered: 25-10-2019
Member Is Offline
|
|
Hi Peter,
Thanks for the tips!
I changed verification to be 30,29,31. autoNth = All.
And set "Entry Level Mode" to SignedPower, "Entry Level Value: 4".
Also I found I did not set "Entry Params. 2" to "-10;10;1" for Params and WF Params. Fixed this. On the other hand I did have "Entry Params" set
properly.
As for "Entry Modes". It looks like there is some incompatibility in settings in trial version and full build that you use. I only have "Cross" Entry
Mode enabled in my settings (there is no CrossOver or "entry cross dual" in the trial build, so I hope "Cross" is the correct setting).
Anyway, I am going to rerun the process. Let's see if these changes help to produce more robust strategies.
|
|
|
admin
Super Administrator
Posts: 5060
Registered: 7-4-2017
Member Is Offline
Mood: No Mood
|
|
| Quote: Originally posted by Sten | Hi Peter,
Thanks for the tips!
I changed verification to be 30,29,31. autoNth = All.
And set "Entry Level Mode" to SignedPower, "Entry Level Value: 4".
Also I found I did not set "Entry Params. 2" to "-10;10;1" for Params and WF Params. Fixed this. On the other hand I did have "Entry Params" set
properly.
As for "Entry Modes". It looks like there is some incompatibility in settings in trial version and full build that you use. I only have "Cross" Entry
Mode enabled in my settings (there is no CrossOver or "entry cross dual" in the trial build, so I hope "Cross" is the correct setting).
Anyway, I am going to rerun the process. Let's see if these changes help to produce more robust strategies.
|
Hi Sten,
trial users has a few less options. This is due to either the feature being experimental - and not decided if its a good idea, or just to keep things
simpler, or to stop users using options that likely are not helpful unless they know what they are doing.
Dual cross is my favorite setting, followed by single level cross, followed by cross.
There is however little difference between them all. Hope to see you as a purchaser soon
|
|
|
Daniel UK1
Member
Posts: 470
Registered: 4-6-2019
Member Is Offline
|
|
Hi Peter, any possibility on getting more info about exit modes ,https://trademaid.info/gsbhelp/Exitmodes.html
I have all these as part of testing in my set process when starting with a new market.. however as it is now, i have no real idea of what they
actually do or mean.
I can read in the link to the manual above about the result of each for an individual market, however i would be interested in understanding what they
actually do.
If you could kindly perhaps include an explanation and not just a formula, of each mode, it would be greatly appreciated.
Thanks
|
|
|
engtraderfx
Junior Member
Posts: 98
Registered: 15-10-2018
Member Is Offline
Mood: No Mood
|
|
Hi Peter, I am updating thru RM & finding no change in root directory c:\GSB but all the exe files are now in cloud directory C:\GSB\Cloud Workers.
Is that where they all appear now? Maybe i should delete & reinstall?
Thanks Dave
|
|
|
admin
Super Administrator
Posts: 5060
Registered: 7-4-2017
Member Is Offline
Mood: No Mood
|
|
| Quote: Originally posted by engtraderfx | Hi Peter, I am updating thru RM & finding no change in root directory c:\GSB but all the exe files are now in cloud directory C:\GSB\Cloud Workers.
Is that where they all appear now? Maybe i should delete & reinstall?
Thanks Dave |
Ive seen another user with this issue, and not sure why.
check c:\gsb is excluded from your anti virus program.
No need to delete, you can just rename the folder to rm.bad
there is a new rm out a few minutes ago. Bug fix but I dont think it will fix your issue.
Worst case make a batch file to copy the manager from \worker folder to \gsb
(or copy it manually)
|
|
|
Daniel UK1
Member
Posts: 470
Registered: 4-6-2019
Member Is Offline
|
|
| Quote: Originally posted by admin | | Quote: Originally posted by engtraderfx | Hi Peter, I am updating thru RM & finding no change in root directory c:\GSB but all the exe files are now in cloud directory C:\GSB\Cloud Workers.
Is that where they all appear now? Maybe i should delete & reinstall?
Thanks Dave |
Ive seen another user with this issue, and not sure why.
check c:\gsb is excluded from your anti virus program.
No need to delete, you can just rename the folder to rm.bad
there is a new rm out a few minutes ago. Bug fix but I dont think it will fix your issue.
Worst case make a batch file to copy the manager from \worker folder to \gsb
(or copy it manually) |
Yes i have the same issue, but mine ends up in WORKERfolder below GSB, not in GSB CLOUD WORKER, they have alwasy ended up in main GSB folder, where i
believe they should end up? my path in RM is correct, so it does not make any sense... i have solved it just by copying over managers and scripts tp
the main GSB folder... but it would be nice to understand why, one day it just changed  Anyway, works as it should in the end.
|
|
|
admin
Super Administrator
Posts: 5060
Registered: 7-4-2017
Member Is Offline
Mood: No Mood
|
|
| Quote: Originally posted by Daniel UK1 | | Quote: Originally posted by admin | | Quote: Originally posted by engtraderfx | Hi Peter, I am updating thru RM & finding no change in root directory c:\GSB but all the exe files are now in cloud directory C:\GSB\Cloud Workers.
Is that where they all appear now? Maybe i should delete & reinstall?
Thanks Dave |
Ive seen another user with this issue, and not sure why.
check c:\gsb is excluded from your anti virus program.
No need to delete, you can just rename the folder to rm.bad
there is a new rm out a few minutes ago. Bug fix but I dont think it will fix your issue.
Worst case make a batch file to copy the manager from \worker folder to \gsb
(or copy it manually) |
Yes i have the same issue, but mine ends up in WORKERfolder below GSB, not in GSB CLOUD WORKER, they have alwasy ended up in main GSB folder, where i
believe they should end up? my path in RM is correct, so it does not make any sense... i have solved it just by copying over managers and scripts tp
the main GSB folder... but it would be nice to understand why, one day it just changed Anyway, works as it should in the end. |
Under rm update field, check gsbmanger folder is set to c:\gsb
put logging (under settgins) to true
days to remove files 7
verbose true
then on next update, send me the log files (folder log)
That way I can look into what happens
|
|
|
Sten
Junior Member
Posts: 35
Registered: 25-10-2019
Member Is Offline
|
|
| Quote: Originally posted by admin | | Quote: Originally posted by Sten | Hi Peter,
Thanks for the tips!
I changed verification to be 30,29,31. autoNth = All.
And set "Entry Level Mode" to SignedPower, "Entry Level Value: 4".
Also I found I did not set "Entry Params. 2" to "-10;10;1" for Params and WF Params. Fixed this. On the other hand I did have "Entry Params" set
properly.
As for "Entry Modes". It looks like there is some incompatibility in settings in trial version and full build that you use. I only have "Cross" Entry
Mode enabled in my settings (there is no CrossOver or "entry cross dual" in the trial build, so I hope "Cross" is the correct setting).
Anyway, I am going to rerun the process. Let's see if these changes help to produce more robust strategies.
|
Hi Sten,
trial users has a few less options. This is due to either the feature being experimental - and not decided if its a good idea, or just to keep things
simpler, or to stop users using options that likely are not helpful unless they know what they are doing.
Dual cross is my favorite setting, followed by single level cross, followed by cross.
There is however little difference between them all. Hope to see you as a purchaser soon |
Hi Peter,
After I rerun the process I've got the following results:
It became better, but still far from what is described in the documentation. F/E stats -15.3% drop concerns me. As far as I understand we should see
an improvement here.
My optimization settings are attached for the reference, but I think I took into account all previous suggestions, and I should be pretty close to the
Methodology from the documentation.
Overall, it is extremely hard to evaluate the trial version and see if GSB works for me, in particular, for the following reasons:
- Methodology process is described in two separate sections in documentation ("Proving the methodology" and "Updates") with some parts being outdated
and another parts overriding what is described previously. There is no plain single Methodology document;
- Some parameter names differ in trial build and in full version (and from what is described in the documentation);
- Trial version does not include some parameters described in the Methodology;
- Optimization Settings file with all Methodology settings is not provided with the trial build (that would simplify the process a lot);
- The software is not completely stable (for example, after I run 250 multithreaded WFO's on my workers + cloud, most of them complete withing the
night, but 1-3 WFO usually remain in the Running state forever, and even if I stop them and run WFO again for these systems, they enter Running state
and never finish);
- As a professional developer I see lot's of bugs in GSB (progress bar works somewhat strange, for example);
- Also product documentation is not easy to read and understand.
So, I like the idea behind GSB, in particular that it is able to do WFO for the generated systems to assess their robustness. But, at least for now,
I'm not able to properly evaluate GSB and prove it can generate some useful systems for me given the IQFeed historical data that I have. At least, I'm
not able to replicate the results from the documentation.
I already own AdaptTrade Builder which I find not being able to generate systems that I consider robust enough to trade for real money, so I'm a bit
careful with GSB evaluation.
I am going run a few more tests and try to produce better systems, and then continue to monitor GSB development. But for now I am not convinced to
invest a solid amount of money into buying the full version.
Attachment: Login to view the details
|
|
|
LucaRicatti
Junior Member
Posts: 16
Registered: 12-10-2019
Member Is Offline
|
|
| Quote: Originally posted by Sten |
Hi Peter,
After I rerun the process I've got the following results:
It became better, but still far from what is described in the documentation. F/E stats -15.3% drop concerns me. As far as I understand we should see
an improvement here.
My optimization settings are attached for the reference, but I think I took into account all previous suggestions, and I should be pretty close to the
Methodology from the documentation.
Overall, it is extremely hard to evaluate the trial version and see if GSB works for me, in particular, for the following reasons:
- Methodology process is described in two separate sections in documentation ("Proving the methodology" and "Updates") with some parts being outdated
and another parts overriding what is described previously. There is no plain single Methodology document;
- Some parameter names differ in trial build and in full version (and from what is described in the documentation);
- Trial version does not include some parameters described in the Methodology;
- Optimization Settings file with all Methodology settings is not provided with the trial build (that would simplify the process a lot);
- The software is not completely stable (for example, after I run 250 multithreaded WFO's on my workers + cloud, most of them complete withing the
night, but 1-3 WFO usually remain in the Running state forever, and even if I stop them and run WFO again for these systems, they enter Running state
and never finish);
- As a professional developer I see lot's of bugs in GSB (progress bar works somewhat strange, for example);
- Also product documentation is not easy to read and understand.
So, I like the idea behind GSB, in particular that it is able to do WFO for the generated systems to assess their robustness. But, at least for now,
I'm not able to properly evaluate GSB and prove it can generate some useful systems for me given the IQFeed historical data that I have. At least, I'm
not able to replicate the results from the documentation.
I already own AdaptTrade Builder which I find not being able to generate systems that I consider robust enough to trade for real money, so I'm a bit
careful with GSB evaluation.
I am going run a few more tests and try to produce better systems, and then continue to monitor GSB development. But for now I am not convinced to
invest a solid amount of money into buying the full version.
|
Hi Stan,
I just endend my trial period and decided to purchase GSB and apply for a Tradestation account (got Multicharts but they need to fix data2 bug).
I had poor results with trial version and had difficulty understanding the method like you, but I liked so much the robustness-stress idea behind GSB
that I choose not to give up and asked Peter for help.
Here the results of my - assisted by Peter - last build with trial version:
PS: First time in this business i met someone dedicated like Peter.
|
|
|
| Pages:
1
..
27
28
29
30
31
..
98 |
![[*]](images/davis/default_icon.gif){kind=icon}
