Announcement

Collapse
No announcement yet.

Partner 728x90

Collapse

NT8061 - Playback-MR takes wrong trades

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

    #16
    Hello sledge,

    Thank you for your response.

    Omitting BarsArray[0].LastBarTime, I am not seeing where you are making these comparisons nor where the difference is viewed.

    Can you confirm how you are seeing the differences? I understand you are looking at the Replay controller for the Replay time, correct?

    I look forward to your response.
    Last edited by NinjaTrader_PatrickH; 06-28-2017, 09:44 AM.

    Comment


      #17
      Originally posted by NinjaTrader_PatrickH View Post
      Hello sledge,

      Thank you for your response.

      Omitting BarsArray[0].LastBarTime, I am not seeing where you are making these comparisons nor where the difference is viewed.

      Can you confirm how you are see the differences? I understand you are looking at the Replay controller for the Replay time, correct?

      I look forward to your response.
      I'm using a 15 second ES chart.
      My original indicator uses Time [0]. My original strategy used Time [0] in a 10 minute window. I began to expand that to other times and stared to get these random trades that made no sense as they couldn't have been executed based on current conditions.

      I tracked it down to Time [0] randomly using some other Time series and returning incorrect times. I built the test indicator and strategy as proof.

      Time [0] should be 15 second increments. 09:30:00, 09:30:15, etc.
      Time [0] should not show me something from the past.

      Removing the AddDataSeries from the strategy shows Time [0] is correct always. (Or changing this to the main series also corrects the defect, ie ES 06-17/ES 06-17.

      In my posts I show that the Time [0] is even producing Time from previous days, which is definitely not part of replay.

      I even downloaded replay data IN NT8 from your servers to disprove it had anything to do with converted replay data from NT7.

      I'm not sure how to be more clear about this.
      If I must make a video tonight, I will.

      Like I said it is random, but it should happen almost daily.
      Attached Files

      Comment


        #18
        I have rewritten the code without LastBarPrice.

        You need data for ES 06-17 and 6E 06-17, or some other combination of ES/something else.

        I have a 15 second ES 06-17. I connect to market replay, change the date to 3/20/2017.
        I add the strategy to the chart.

        I run replay and then examine the log file after 4pm when it is closed.

        Notes:
        1) The strategy may self disable when trying to run to a file that isn't open (file closes a little after 4pm). This would mean that a date/time stamp outside of 9-4pm EST was seen and it tried to write it to a file.
        2) The indicator will print a second ^^^^^^^ when it gets back to the real time.
        3) I added Time[1], Time[2] in the prints, you can expand this further...but I suspect figuring out why the wrong series is used in the first place - would fix the rest of the series.

        There is a margin of error of 30 seconds or so in case no data is seen in ES. This is fine with me, the market doesn't always trade. But when I see a Time[0] from 6 minutes earlier, 1 hour earlier, or a day ago: something is going wrong behind the scenes.


        Code:
        						if (   (Time[0] > (prevTime.AddSeconds(30 ))
        							|| (Time[0] < (prevTime.AddSeconds(-30) )) ) )
        Attached Files

        Comment


          #19
          I have also tested version 3 of BarsArrayTest Strategy/Indicator on NT7. There are no issues there with Time[0] series being out of whack in the indicator.

          There is something that is structurally wrong in NT8 replay.

          I wonder if the bug was a result of NT8Beta 10 - when market replay became really slow (where # of charts/tab/strategies enabled became a serious problem and had to be limited).
          I think you all messed with the threading model to fix other issues. There were some corrections in a few releases later, but this thing is still out of whack.

          I'm going to try to test Beta9 and see what I get.
          Last edited by sledge; 07-01-2017, 11:13 AM.

          Comment


            #20
            Well, it was there in NT8 B9 also. So I must suspect then it has always been there. This might be part of all the random issues in replay from the start. I don't feel like going back to NT8B7/B8.

            I'm going to try another NT7 test just to make sure.


            NT8B9- 01-29-2014 does not belong in 02-14-2014 replay day.

            Code:
            ReplayTime=[B]20140204 [/B]*[B]105045 [/B]Close[0]=1743.75 High[0]=1743.75 Open[0]=1743.75 Low[0]=1743.75 * Times[0][0]=20140204 *105045
            ReplayTime=[B]20140204 [/B]*[B]105100 [/B]Close[0]=1743.75 High[0]=1743.75 Open[0]=1743.75 Low[0]=1743.75 * Times[0][0]=20140204 *105100
            ReplayTime=[B][COLOR="Red"]20140129 [/COLOR][/B]*[B][COLOR="Red"]145945 [/COLOR][/B]Close[0]=1768.5 High[0]=1768.75 Open[0]=1768.5 Low[0]=1768.5 * Times[0][0]=20140129 *145945
            ***^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^****
            ReplayTime=[B]20140204 [/B]*[B]105100 [/B]Close[0]=1743.75 High[0]=1743.75 Open[0]=1743.75 Low[0]=1743.5 * Times[0][0]=20140204 *105100
            ReplayTime=[B]20140204 [/B]*[B]105115 [/B]Close[0]=1744 High[0]=1744 Open[0]=1744 Low[0]=1744 * Times[0][0]=20140204 *105115
            ReplayTime=[B]20140204 [/B]*[B]105130 [/B]Close[0]=1744 High[0]=1744 Open[0]=1744 Low[0]=1744 * Times[0][0]=20140204 *105130
            Another replay shows it did not happen the next time.

            Code:
            ReplayTime=20140204 *105045 Close[0]=1743.75 High[0]=1743.75 Open[0]=1743.75 Low[0]=1743.75 * Times[0][0]=20140204 *105045
            ReplayTime=20140204 *105100 Close[0]=1743.75 High[0]=1743.75 Open[0]=1743.75 Low[0]=1743.75 * Times[0][0]=20140204 *105100
            ReplayTime=20140204 *105115 Close[0]=1744 High[0]=1744 Open[0]=1744 Low[0]=1744 * Times[0][0]=20140204 *105115
            ReplayTime=20140204 *105130 Close[0]=1744 High[0]=1744 Open[0]=1744 Low[0]=1744 * Times[0][0]=20140204 *105130
            Attached Files
            Last edited by sledge; 07-05-2017, 06:12 PM.

            Comment


              #21
              Hello sledge,

              Thank you for your response.

              Can you attach the full output in a text document to your response?

              I look forward to your response.

              Comment


                #22
                Originally posted by NinjaTrader_PatrickH View Post
                Hello sledge,

                Thank you for your response.

                Can you attach the full output in a text document to your response?

                I look forward to your response.
                Done.

                Note, 12:23:30 is NOT a problem, there appears to have been no trades or internet type disruption in ES for over 30 seconds. A lull in the market. It happens.

                Code:
                ReplayTime=20140204  122215 Close[0]=1750 High[0]=1750 Open[0]=1750 Low[0]=1750   Times[0][0]=20140204  122215
                ReplayTime=20140204  122330 Close[0]=1750 High[0]=1750 Open[0]=1750 Low[0]=1750   Times[0][0]=20140204  122330
                ***^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^****
                ReplayTime=20140204  122345 Close[0]=1750 High[0]=1750 Open[0]=1750 Low[0]=1750   Times[0][0]=20140204  122345

                Comment


                  #23
                  Originally posted by NinjaTrader_PatrickH View Post
                  Hello sledge,

                  Thank you for your response.

                  Can you attach the full output in a text document to your response?

                  I look forward to your response.
                  I updated on 7/5. Today is 7/11. Maybe you didn't get notice. Hopefully this goes through.

                  Comment


                    #24
                    Hello sledge,

                    Apologizes for the delay.

                    I will update this thread when I have information on this matter.

                    Thank you for your patience.

                    Comment


                      #25
                      Hello sledge,

                      Thank you for your response.

                      May we test your Replay data for the date "20140204"? We have been unable to reproduce the same behavior on our replay data or historical data playback.

                      You can send the data files over to platformsupport[at]ninjatrader[dot]com with 'ATTN: Patrick H' in the subject line and a reference to this thread in the body of the email.

                      I look forward to your response.

                      Comment


                        #26
                        Originally posted by NinjaTrader_PatrickH View Post
                        May we test your Replay data for the date "20140204"? We have been unable to reproduce the same behavior on our replay data or historical data playback..
                        I did more extensive testing this weekend and downloaded ES 09-17 and 6E 09-17 from 7/7 to 7/14, and it hit right away on 7/7.


                        https://www.screencast.com/t/qembiz2h <= 2 runs, each in different locations.

                        I switched to ES ##-## (how I stack my data, you may use ES 03-14/^VIX, and received issues on 1/31/2014 this time.


                        https://www.screencast.com/t/Y3zBvQ0tJyY (different location)

                        Run #3, now 1/3/2014 did not have a problem, but 2/3/2014 had a problem.

                        https://www.screencast.com/t/36w2NHSgan ( recording time limit reached)
                        https://www.screencast.com/t/t8XMQGVGRrRD ( cont )

                        I have attached new code which says in the output window when to look in a file.

                        I will send also send in the January 2014 data-Feb 2014 data via email.

                        Main ES chart - 15 second.
                        Attached strategy to chart in replay mode.
                        Attached Files

                        Comment


                          #27
                          Hello sledge,

                          Thank you for the updates and additional information.

                          I will test the data on my end.

                          Comment

                          Latest Posts

                          Collapse

                          Topics Statistics Last Post
                          Started by frankthearm, Today, 09:08 AM
                          7 responses
                          27 views
                          0 likes
                          Last Post NinjaTrader_Clayton  
                          Started by NRITV, Today, 01:15 PM
                          1 response
                          5 views
                          0 likes
                          Last Post NinjaTrader_Jesse  
                          Started by maybeimnotrader, Yesterday, 05:46 PM
                          5 responses
                          24 views
                          0 likes
                          Last Post NinjaTrader_ChelseaB  
                          Started by quantismo, Yesterday, 05:13 PM
                          2 responses
                          16 views
                          0 likes
                          Last Post quantismo  
                          Started by adeelshahzad, Today, 03:54 AM
                          5 responses
                          33 views
                          0 likes
                          Last Post NinjaTrader_BrandonH  
                          Working...
                          X