| Pages:
1
..
14
15
16
17
18
..
98 |
cyrus68
Member
Posts: 171
Registered: 5-6-2017
Member Is Offline
Mood: No Mood
|
|
I tried WF in 48.24. I ran one WF. Upon completion, there were no curves and no reports. So I double-clicked it, whereupon the curves and reports
appeared. Only problem was that I had loaded the 30min dataset in "WF price data" field. But, it ran WF on the 25min dataset.
So, the basic question is: how do you select the dataset for which you want to do WF?
How do you select an alternative dataset to do WF for the same system? In your CL example, let's say you want to run WF on 30min and then on 31min.
What's the procedure? What are the steps?
|
|
|
admin
Super Administrator
Posts: 5060
Registered: 7-4-2017
Member Is Offline
Mood: No Mood
|
|
Quote: Originally posted by cyrus68  | I tried WF in 48.24. I ran one WF. Upon completion, there were no curves and no reports. So I double-clicked it, whereupon the curves and reports
appeared. Only problem was that I had loaded the 30min dataset in "WF price data" field. But, it ran WF on the 25min dataset.
So, the basic question is: how do you select the dataset for which you want to do WF?
How do you select an alternative dataset to do WF for the same system? In your CL example, let's say you want to run WF on 30min and then on 31min.
What's the procedure? What are the steps? |
WF will be opt price data if WF price data is blank.
Otherwise put your WF price data with the data you want.
If that doesnt work, time to do teamviwer.com
I critical question to ask. does WF on 30 min or 29,30,31 min bars give better out of sample results. Im looking into that now.
|
|
|
cyrus68
Member
Posts: 171
Registered: 5-6-2017
Member Is Offline
Mood: No Mood
|
|
I had 3 datasets loaded in "Opt price data": 25 30 35.
I had 30min loaded in "WF price data". What it did was to run WF on 25. According to your comments, it should have run WF on 30min.
If I leave "WF price data" blank, on which of 25 30 35 will it run WF? Seems it will grab the first dataset.
Also, I ran GSB with "Optimise price data" set to True. When I tried to verify, all the scores were 0/0, which is nonsense; based on previous trials.
It did, however, identify which frequency had been selected as optimal. In the previous build, the verification was done correctly and the graph
showed all the information.
I am following the rules and getting odd results.
If you have time, maybe we can do TV tomorrow morning.
|
|
|
admin
Super Administrator
Posts: 5060
Registered: 7-4-2017
Member Is Offline
Mood: No Mood
|
|
| Quote: Originally posted by cyrus68 | I had 3 datasets loaded in "Opt price data": 25 30 35.
I had 30min loaded in "WF price data". What it did was to run WF on 25. According to your comments, it should have run WF on 30min.
CORRECT
If I leave "WF price data" blank, on which of 25 30 35 will it run WF? Seems it will grab the first dataset.
IT COULD DO THIS IF OPTIMIZE DATA SET IS TRUE. IF ITS FALSE IT SHOULD NOT DO THIS
Also, I ran GSB with "Optimise price data" set to True. When I tried to verify, all the scores were 0/0, which is nonsense; based on previous trials.
I HAVE NEVER USED THIS FEATURE YET. MIGHT HAVE A BUG, I DONT KNOW
It did, however, identify which frequency had been selected as optimal. In the previous build, the verification was done correctly and the graph
showed all the information.
I am following the rules and getting odd results.
If you have time, maybe we can do TV tomorrow morning.
HAPPY TO DO THAT. EMAIL WHEN YOUR UP |
|
|
|
admin
Super Administrator
Posts: 5060
Registered: 7-4-2017
Member Is Offline
Mood: No Mood
|
|
| Quote: Originally posted by cyrus68 | I had 3 datasets loaded in "Opt price data": 25 30 35.
I had 30min loaded in "WF price data". What it did was to run WF on 25. According to your comments, it should have run WF on 30min.
If I leave "WF price data" blank, on which of 25 30 35 will it run WF? Seems it will grab the first dataset.
Also, I ran GSB with "Optimise price data" set to True. When I tried to verify, all the scores were 0/0, which is nonsense; based on previous trials.
It did, however, identify which frequency had been selected as optimal. In the previous build, the verification was done correctly and the graph
showed all the information.
I am following the rules and getting odd results.
If you have time, maybe we can do TV tomorrow morning. |
48.26 fixed the wf on 25 min, instead of 30.
|
|
|
admin
Super Administrator
Posts: 5060
Registered: 7-4-2017
Member Is Offline
Mood: No Mood
|
|
ts10
can anyone give comments if TS10 is ok for live trading? ie stable and no significant bugs?
|
|
|
Bruce
Member
Posts: 115
Registered: 22-7-2018
Location: Auckland - New Zealand
Member Is Offline
Mood: No Mood
|
|
I haven't migrated to it yet, as a general rule I wait until there's been a number of updates before upgrading.
|
|
|
Petzy
Junior Member
Posts: 73
Registered: 24-10-2017
Location: Sweden
Member Is Offline
Mood: No Mood
|
|
I tried TradeStation 10 for maybe 4-6 months ago believing it would have fewer bugs and that it would be a 64 bit application. Turns out it is
basically version 9.5 with a newer look, but more bugs and still a 32 bit application.
Switching to version 10 gives you nothing in my opinion.
If I had time to switch to anything it would be MultiCharts for the 64 bit and fewer bugs.
I use TradeStation 9.5 because of the data feed and the connection to broker. I would consider MultiCharts if the datafeed/broker were better. I would
not consider Tradestation 10.
Regards,
Peter G
|
|
|
eastpeace
Junior Member
Posts: 19
Registered: 15-9-2018
Member Is Offline
Mood: No Mood
|
|
Hello everyone,
I have 2 questions.
1, I have get some strategies, all the entry type are "buy/sellshort 1 contracts this bar on close".
How to set that as "buy 1 contracts next bar at market"?
2, I have only MC, no TS.
Norm2 function compile with errors,
“MultiCharts Standard does not allow using the following reserved words: DefineDLLFunc, External, External Method. The words is available in
MultiCharts Gold.”
So, I have to comment the first line out.
External: "TradeStationUtilsRc7b.dll", {success}bool, "WriteFileCpp", {path}string, {text}string, {delayFileWrite}int;
Does that matter?
|
|
|
eastpeace
Junior Member
Posts: 19
Registered: 15-9-2018
Member Is Offline
Mood: No Mood
|
|
Hello everyone,
I have 2 questions.
1, I have get some strategies, all the entry type are "buy/sellshort 1 contracts this bar on close".
How to set that as "buy 1 contracts next bar at market"?
2, I have only MC, no TS.
Norm2 function compile with errors,
“MultiCharts Standard does not allow using the following reserved words: DefineDLLFunc, External, External Method. The words is available in
MultiCharts Gold.”
So, I have to comment the first line out.
External: "TradeStationUtilsRc7b.dll", {success}bool, "WriteFileCpp", {path}string, {text}string, {delayFileWrite}int;
Does that matter?
|
|
|
admin
Super Administrator
Posts: 5060
Registered: 7-4-2017
Member Is Offline
Mood: No Mood
|
|
| Quote: Originally posted by eastpeace | Hello everyone,
I have 2 questions.
1, I have get some strategies, all the entry type are "buy/sellshort 1 contracts this bar on close".
How to set that as "buy 1 contracts next bar at market"?
2, I have only MC, no TS.
Norm2 function compile with errors,
“MultiCharts Standard does not allow using the following reserved words: DefineDLLFunc, External, External Method. The words is available in
MultiCharts Gold.”
So, I have to comment the first line out.
External: "TradeStationUtilsRc7b.dll", {success}bool, "WriteFileCpp", {path}string, {text}string, {delayFileWrite}int;
Does that matter?
|
1) Currently only This bar at market supported. Other entry types are in the pipeline. Next bar at market buys at the open, which is later. The
earlier the order the better.
2) Norm2 is legacy GSB code. It shouldnt be needed. The dll was just a faster way to write files, and I dont think GSB uses it. It should not have
been included with the pla
|
|
|
eastpeace
Junior Member
Posts: 19
Registered: 15-9-2018
Member Is Offline
Mood: No Mood
|
|
Thank you, admin.
I will do some sim trade for test.
Is the MaxBarsBack number always 500? Why? Can it be reseted?
And 2nd, the data have 3 parts, 40% for training, 50% for test, 10% for validation. But the graphs above have no validation part showing. And the
PF-V,NP/DD-V are blank. -S are equal to -SV.
|
|
|
admin
Super Administrator
Posts: 5060
Registered: 7-4-2017
Member Is Offline
Mood: No Mood
|
|
| Quote: Originally posted by eastpeace | Thank you, admin.
I will do some sim trade for test.
Is the MaxBarsBack number always 500? Why? Can it be reseted?
And 2nd, the data have 3 parts, 40% for training, 50% for test, 10% for validation. But the graphs above have no validation part showing. And the
PF-V,NP/DD-V are blank. -S are equal to -SV. |
maxbars should never be changed. Otherwise you may get GSB<>TS
Im unclear what graphs you mean. My response is a guess to what you say. When doing Nth tests, I do 100% test, as we then have every second day un
seen data.
|
|
|
eastpeace
Junior Member
Posts: 19
Registered: 15-9-2018
Member Is Offline
Mood: No Mood
|
|
What I am mean is that,
Even when I divide the data into 3 sections, training, test, validation.
But the generated strategies's equity curve just show training and test part.
The validation stats in the below table are also empty.
|
|
|
admin
Super Administrator
Posts: 5060
Registered: 7-4-2017
Member Is Offline
Mood: No Mood
|
|
| Quote: Originally posted by eastpeace | What I am mean is that,
Even when I divide the data into 3 sections, training, test, validation.
But the generated strategies's equity curve just show training and test part.
The validation stats in the below table are also empty. |
As a test put a higher percent. Check dates on left side is set to all
|
|
|
eastpeace
Junior Member
Posts: 19
Registered: 15-9-2018
Member Is Offline
Mood: No Mood
|
|
Is it right?
The training part is negative in equity curve. But the stats below are positive.
Attachment: Login to view the details
Attachment: Login to view the details
|
|
|
admin
Super Administrator
Posts: 5060
Registered: 7-4-2017
Member Is Offline
Mood: No Mood
|
|
Sorry, but im not clear what your asking. Weren't you looking for validation stats?
|
|
|
kelsotrader
Junior Member
Posts: 29
Registered: 16-2-2018
Location: Tapanui - New Zealand
Member Is Offline
Mood: No Mood
|
|
I have moved this Question.
I have a several queries that as yet have not been able to find answers for.
1: The file named GSB.pdb that accompanies beta upgrades: Is in necessary to replace existing files of the same name, and what does it do.
2: When switching from a testing strategy in TS to a live or demo system what should the lookback bars or days, months be set to.
3: I have noticed that on one strategy that I traded live that the number of trades being executed on TS do not correspond with trades when TS is
asked to be in backtesting mode. While others are almost exact. What could be the cause of the variation.
4: When using TS analyzer and bringing up a trade list it cannot be exported in XML format for importing into Advanced portfolio analyser.
Is there a solution to exporting trades so Portfolio Analyser can pick them up.
Answers are from Peter.
1: It is required for crashes
2: 500
3: Unknown
4: Have you tried Portfolio Analyser.
|
|
|
eastpeace
Junior Member
Posts: 19
Registered: 15-9-2018
Member Is Offline
Mood: No Mood
|
|
This is a new question, not about the validation part showing or stats.
The problem is that earning in the training part according to the equity curves looks ugly, but the stats in the below table are positive. Which one
is right?
|
|
|
admin
Super Administrator
Posts: 5060
Registered: 7-4-2017
Member Is Offline
Mood: No Mood
|
|
| Quote: Originally posted by eastpeace |
This is a new question, not about the validation part showing or stats.
The problem is that earning in the training part according to the equity curves looks ugly, but the stats in the below table are positive. Which one
is right? |
double click on a row, and see if the results are correct. You shouldnt normally see a curve where training is negitive, as the training filters
prevent this. Unless a double click fixes this, send me your teamviewer details
|
|
|
eastpeace
Junior Member
Posts: 19
Registered: 15-9-2018
Member Is Offline
Mood: No Mood
|
|
hello,sir
I see the connection.
I think the teamviewer is ok now. Please try again. Thank you
|
|
|
ProbTrader
Junior Member
Posts: 16
Registered: 3-7-2018
Member Is Offline
Mood: No Mood
|
|
Hello all
So I just started to (try to) transfer some of my systems from GSB to TS. I am running into an issue with the TS code however. For the system below, I
want to exit every trade at 2200 hrs (local GMT +1 time). It works perfectly fine on GSB. But on TS, the code generated by GSB, does not work
flawlessly. The trades that are entered after 22:00 (eg at 22:20), are exited on the same bar/day. In GSB, this is handled correctly (given my
inputs), ie all trades are exited at 22:00, either the following day if the trade is entered before 00:00, or the same day if the trade is entered
after 00:00.
Question: How could I get TS to replicate this exactly? Just cutting and pasting the code as is does not seem to work. And as I am a complete
illiterate in EL, I don't know how to adjust the code to accomplish this. Could somebody please be kind and let me know how the code needs to be
adjusted in order to create the identical trade signals in TS as I am getting in GSB. Thanks in advance.
TS code:
////Contract's Session Close
If (TimeHms >= 220000) Then
Begin
BuyToCover this bar on close;
Sell this bar on close;
End;
SetExitOnClose;
// Exit Minutes
If BarDateTime.ELDateTimeEx >= tradeExitTime Then
Begin
BuyToCover this bar on close;
Sell this bar on close;
tradeExitTime = 1000000000;
End;
// Trading Dates
If Not ((dateYmd >= 20000101 And dateYmd <= 20181231)) Then
Begin
BuyToCover this bar on close;
Sell this bar on close;
End;
// Result and decision
result = ((GSB_Norm2(TrueRange of Data(i1Data), 11, 100) of Data(i1Data) * i1Weight) * ((GSB_Norm2(GSB_FastK(i2stockLength) of Data(i2Data), 11, 100)
of Data(i2Data) * i2Weight) * ((GSB_Norm2(GSB_Decycler(i3cutoff) of Data(i3Data), 11, 100) of Data(i3Data) * i3Weight) * (GSB_Norm2(GSB_Highest(High,
i4length) of Data(i4Data), 11, 100) of Data(i4Data) * i4Weight))));
result = IFF(AbsValue(result) > zs, result, 0);
decision = GSB_Decision(result, 1, entryParams);
sfResult = GSB_Norm2(GSB_CloseOverPrevCloseD of Data(iSFData), 11, 100) of Data(iSFData) * iSFWeight;
sfResult = IFF(AbsValue(sfResult) > zs, sfResult, 0);
sfDecision = GSB_Decision(sfResult, 2, sfEntryLevel);
// Entry-filter check
If (dateYmd > beginDate And dateYmd <= endDate)
And ((dateYmd >= 20000101 And dateYmd <= 20181231))
And (Not (TimeHms >= 220000 And TimeHms <= 220001)) Then
Begin
// Buy/Sell
If decision = 1 And sfDecision = 1 Then
Begin
Buy("Long entry") 1 contracts this bar on close;
tradeExitTime = GSB_CalcDateTime(BarDateTime.ELDateTimeEx, exitMinutesValue);
End
Else If decision = -1 And sfDecision = -1 Then
Begin
SellShort("Short entry") 1 contracts this bar on close;
tradeExitTime = GSB_CalcDateTime(BarDateTime.ELDateTimeEx, exitMinutesValue);
End;
End;
|
|
|
admin
Super Administrator
Posts: 5060
Registered: 7-4-2017
Member Is Offline
Mood: No Mood
|
|
| Quote: Originally posted by ProbTrader | Hello all
So I just started to (try to) transfer some of my systems from GSB to TS. I am running into an issue with the TS code however. For the system below, I
want to exit every trade at 2200 hrs (local GMT +1 time). It works perfectly fine on GSB. But on TS, the code generated by GSB, does not work
flawlessly. The trades that are entered after 22:00 (eg at 22:20), are exited on the same bar/day. In GSB, this is handled correctly (given my
inputs), ie all trades are exited at 22:00, either the following day if the trade is entered before 00:00, or the same day if the trade is entered
after 00:00.
Question: How could I get TS to replicate this exactly? Just cutting and pasting the code as is does not seem to work. And as I am a complete
illiterate in EL, I don't know how to adjust the code to accomplish this. Could somebody please be kind and let me know how the code needs to be
adjusted in order to create the identical trade signals in TS as I am getting in GSB. Thanks in advance.
TS code:
////Contract's Session Close
If (TimeHms >= 220000) Then
Begin
BuyToCover this bar on close;
Sell this bar on close;
End;
SetExitOnClose;
// Exit Minutes
If BarDateTime.ELDateTimeEx >= tradeExitTime Then
Begin
BuyToCover this bar on close;
Sell this bar on close;
tradeExitTime = 1000000000;
End;
// Trading Dates
If Not ((dateYmd >= 20000101 And dateYmd <= 20181231)) Then
Begin
BuyToCover this bar on close;
Sell this bar on close;
End;
// Result and decision
result = ((GSB_Norm2(TrueRange of Data(i1Data), 11, 100) of Data(i1Data) * i1Weight) * ((GSB_Norm2(GSB_FastK(i2stockLength) of Data(i2Data), 11, 100)
of Data(i2Data) * i2Weight) * ((GSB_Norm2(GSB_Decycler(i3cutoff) of Data(i3Data), 11, 100) of Data(i3Data) * i3Weight) * (GSB_Norm2(GSB_Highest(High,
i4length) of Data(i4Data), 11, 100) of Data(i4Data) * i4Weight))));
result = IFF(AbsValue(result) > zs, result, 0);
decision = GSB_Decision(result, 1, entryParams);
sfResult = GSB_Norm2(GSB_CloseOverPrevCloseD of Data(iSFData), 11, 100) of Data(iSFData) * iSFWeight;
sfResult = IFF(AbsValue(sfResult) > zs, sfResult, 0);
sfDecision = GSB_Decision(sfResult, 2, sfEntryLevel);
// Entry-filter check
If (dateYmd > beginDate And dateYmd <= endDate)
And ((dateYmd >= 20000101 And dateYmd <= 20181231))
And (Not (TimeHms >= 220000 And TimeHms <= 220001)) Then
Begin
// Buy/Sell
If decision = 1 And sfDecision = 1 Then
Begin
Buy("Long entry") 1 contracts this bar on close;
tradeExitTime = GSB_CalcDateTime(BarDateTime.ELDateTimeEx, exitMinutesValue);
End
Else If decision = -1 And sfDecision = -1 Then
Begin
SellShort("Short entry") 1 contracts this bar on close;
tradeExitTime = GSB_CalcDateTime(BarDateTime.ELDateTimeEx, exitMinutesValue);
End;
End;
|
try changing this from
Entry-filter check
If (dateYmd > beginDate And dateYmd <= endDate)
And ((dateYmd >= 20000101 And dateYmd <= 20181231))
And (Not (TimeHms >= 220000 And TimeHms <= 220001)) Then
to
Entry-filter check
If (dateYmd > beginDate And dateYmd <= endDate)
And ((dateYmd >= 20000101 And dateYmd <= 20181231))
And (Not (TimeHms >= 220000 )) Then
|
|
|
ProbTrader
Junior Member
Posts: 16
Registered: 3-7-2018
Member Is Offline
Mood: No Mood
|
|
Thanks Peter.
The solution you posted takes out the possibility of entering and exiting on the same bar by excluding all trades that may otherwise be entered
between 22:00 and 00:00. Excluding these trades doesn't solve the issue unfortunately. The system would now miss out any potential trades between
22:00 and 00:00, which is not the intention and is also not what GSB does.
Solving by excluding trades will obviously become an even bigger factor if the close chosen would be at say 17:00 (in which case no trades would be
entered between 17:00 and 00:00, which is neither the intention nor desireable).
Appreciate your help resolving this in a way that makes the trade rules/logic identical in GSB and the EL code it is generating.
|
|
|
ProbTrader
Junior Member
Posts: 16
Registered: 3-7-2018
Member Is Offline
Mood: No Mood
|
|
To further clarify (with relation to above script)...
Current GSB logic:
If enter at < 22:00 on DAY 0, then exit at 22:00 on DAY 0
If enter at > 22:00 on DAY 0, then exit at 22:00 on DAY 1
If the auto-generated EL code for the above is run in TS, this TS logic and trades will not be identical.
Thus, some sort of adjustment is needed the TS code.
Thanks again.
|
|
|
| Pages:
1
..
14
15
16
17
18
..
98 |
![[*]](images/davis/default_icon.gif){kind=icon}
