Announcement

**NinjaTrader_Lance** · 08-07-2013, 10:43 AM

While I'm not able to run the script on my end due to a failed write, I think there may be a misunderstanding about how bar data is referenced. With a multi series script OBU() will possibly be called twice at the same time. It will then evaluate each bar series in the order they were added. Your script doesn't appear to account for times when the time stamps are equal.

While this may not be the issue you're referring to I want to note, It would also be possible for one series to start having ticks on the new bar before the lower volume instrument had a tick placed on its new bar.

If you're looking to more monitor the time at which Kinetick (or any other datafeed) is sending the data consider using OnMarketData() http://www.ninjatrader.com/support/h...marketdata.htm

**NutFlush** · 08-07-2013, 11:12 AM

Originally posted by NinjaTrader_Lance View Post

While I'm not able to run the script on my end due to a failed write, I think there may be a misunderstanding about how bar data is referenced. With a multi series script OBU() will possibly be called twice at the same time. It will then evaluate each bar series in the order they were added. Your script doesn't appear to account for times when the time stamps are equal.

While this may not be the issue you're referring to I want to note, It would also be possible for one series to start having ticks on the new bar before the lower volume instrument had a tick placed on its new bar.

If you're looking to more monitor the time at which Kinetick (or any other datafeed) is sending the data consider using OnMarketData() http://www.ninjatrader.com/support/h...marketdata.htm

Hi Lance,

is the ZIP file corrupt or what do you mean by "failed write"? I just redid the export of the script, but I can also paste the code here if you prefer.

I am aware that the OBU() is called for each tick and each symbol seperately. What I check with my code is when there are ticks for a bar that was already completed on one symbol, but the equivalent bar of the other symbol still receives ticks (although the ending time of the bar has already passed so those ticks should be sorted into the next bar). In these unexpected cases Time[0] of the "late" symbol is smaller than the bar time stamp of the other symbol (which was already increased to the next minute).

There is no check for an equal or bigger time stamp, as that would be expected "normal" behaviour.

Thanks,
NutFlush

Attached Files

CompareBarTimeOfTwoSymbols.zip (2.2 KB, 7 views)

**NinjaTrader_Lance** · 08-07-2013, 11:30 AM

I get the message

**NT** Enabling NinjaScript strategy 'tkCompareTimingOfTwoTickStreams/9159d0a4cce7438c9fdda3e94ea1d575' : On starting a real-time strategy - StrategySync=SubmitImmediately SyncAccountPosition=False EntryHandling=AllEntries EntriesPerDirection=1 StopTargetHandling=PerEntryExecution ErrorHandling=StopStrategyCancelOrdersClosePositio ns ExitOnClose=True/ triggering 30 before close Set order quantity by=Strategy ConnectionLossHandling=KeepRunning DisconnectDelaySeconds=10 CancelEntryOrdersOnDisable=True CancelExitOrdersOnDisable=True CalculateOnBarClose=False MaxRestarts=4 in 5 minutes
### 2SymbolsBarTimeSyncTest initialized on symbol: ES; PC-Time at Home: 07.08.2013 11:17:29.154
>>> Failed to write to log file: C:\2SymbolsBarTimeSyncTest_ES_SPY.txt
**NT** Error on calling 'OnStartUp' method for strategy 'tkCompareTimingOfTwoTickStreams/9159d0a4cce7438c9fdda3e94ea1d575': A required privilege is not held by the client.

But this is besides the point.

It's now my understanding that you're saying this one line should never be true correct? If so let me know and I can test the simpler script on my end.

(BarsInProgress == 0 && Times[0][0] < Times[1][0])

**NutFlush** · 08-07-2013, 11:54 AM

Originally posted by NinjaTrader_Lance View Post

It's now my understanding that you're saying this one line should never be true correct? If so let me know and I can test the simpler script on my end.

(BarsInProgress == 0 && Times[0][0] < Times[1][0])

What a strange error message...

But you were at least able to open the script, right?

Yes, this term should never become true (or only at times where the causing tick was delayed a little bit, but certainly not for more than a second):

if ((BarsInProgress == 0 && Times[0][0] < Times[1][0]) || (BarsInProgress == 1 && Times[1][0] < Times[0][0]))

Since any of the two symbols could be "late" we need to check both ways.

Thanks for looking into it!
NutFlush

P.S. I took a look at the error message and it looks like the strategy fails to write the log file to C:\. Maybe you need to change the path in the code to a place where you have writing rights? Or simply deactivate that part of the code and just look at the output window.

**NinjaTrader_Lance** · 08-08-2013, 08:39 AM

Sorry for the delayed response. I had forgotten a key point to the way OBU() is called with multi-instrument series.

Looking at this snippet

if(BarsInProgress == 0 && Times[0][0] < Times[1][0])

This shouldn't trigger true when dealing with a multi series script with the same instrument. However, in this case we are dealing with a multi instrument script. It's possible BIP 0 gets called several times without BIP 1 ever getting called or vice versa. Several minutes could go by leading to the results you're seeing.

For example BIP 0 is the ES BIP 1 is the NQ
The ES has multiple bars go by without the NQ receiving a tick
When the NQ does receive its next tick the time stamp of that is now behind that of the ES. It could be behind by a few seconds or several minutes depending on how much time passed without a tick.

**NutFlush** · 08-08-2013, 09:47 AM

Originally posted by NinjaTrader_Lance View Post

Sorry for the delayed response. I had forgotten a key point to the way OBU() is called with multi-instrument series.

Looking at this snippet

if(BarsInProgress == 0 && Times[0][0] < Times[1][0])

This shouldn't trigger true when dealing with a multi series script with the same instrument. However, in this case we are dealing with a multi instrument script. It's possible BIP 0 gets called several times without BIP 1 ever getting called or vice versa. Several minutes could go by leading to the results you're seeing.

For example BIP 0 is the ES BIP 1 is the NQ
The ES has multiple bars go by without the NQ receiving a tick
When the NQ does receive its next tick the time stamp of that is now behind that of the ES. It could be behind by a few seconds or several minutes depending on how much time passed without a tick.

Hi Lance,

please forgive for saying this, but it appears you don't understand the problem I described in my posts so far.

Let me explain by using your example:
1. ES and NQ both receive a tick at 10:00:15, so the current bar's time for both is 10:01.
2. From then on NQ doesn't receive any more ticks, while ES does.
3. ES receives a tick at 10:03:11, so ES is now in bar 10:04 while NQ hasn't closed bar 10:01 yet, right?
4. NQ receives a tick at 10:03:13. What it should do is close bar 10:01 and then put the new tick into bar 10:04 as the FirstTickInBar. So that Times[1][0] is 10:04 (NQ) and Times[0][0] is 10:04 (ES).

Now here's the problem: The new tick is still put into the unfinished bar 10:01 (not 10:04). As a consequence Times[1][0] is still 10:01 (NQ) while Times[0][0] is 10:04 (ES).That's what if(BarsInProgress == 1 && Times[1][0] < Times[0][0]) is checking and then calling a "follow-up" tick.

Could I make myself clear now?

Originally posted by NinjaTrader_Lance View Post

When the NQ does receive its next tick the time stamp of that is now behind that of the ES. It could be behind by a few seconds or several minutes depending on how much time passed without a tick.

It is my understanding that the current bar's time of the instrument that receives a new tick should not be behind the current bar's time of the other instrument (if both instruments are on the same time interval, of course)!

Cheers,
NutFlush

**NinjaTrader_Lance** · 08-08-2013, 11:12 AM

Yes it is your step 4 that is where the difference occurs. This is because the time that is returned with be the previous bar's time stamp at the start of the OBU() (the point when you're checking it). Once the OBU() finishes the time will be updated. This is why you only see this happen in real time but would not replicate it on historical data.

If you want to monitor the time of the current tick instead of the bar please use OnMarketData() http://www.ninjatrader.com/support/h...marketdata.htm

This is a useful read as well: http://www.ninjatrader.com/support/h..._are_built.htm

Let me know if I can further assist.

**NutFlush** · 08-08-2013, 11:33 AM

Originally posted by NinjaTrader_Lance View Post

Yes it is your step 4 that is where the difference occurs. This is because the time that is returned with be the previous bar's time stamp at the start of the OBU() (the point when you're checking it). Once the OBU() finishes the time will be updated. This is why you only see this happen in real time but would not replicate it on historical data.

If you want to monitor the time of the current tick instead of the bar please use OnMarketData() http://www.ninjatrader.com/support/h...marketdata.htm

This is a useful read as well: http://www.ninjatrader.com/support/h..._are_built.htm

Let me know if I can further assist.

Hi Lance,

what you write is correct under the assumption that COBC is set to true.

But I was always talking about COBC being set to false. That's why I said in the example above:
4. NQ receives a tick at 10:03:13. What it should do is close bar 10:01 and then put the new tick into bar 10:04 as the FirstTickInBar.

But FirstTickInBar is not going to true while all those "follow-up" ticks are coming in.

Please try my strategy for yourself (XIV & SPY make a good pair) and you will quickly see plenty of "follow-up" ticks.

Thanks,
NutFlush

**NinjaTrader_Lance** · 08-08-2013, 11:52 AM

Hello,

It does not matter if COBC is set to true or false. This is expected behavior and my tests were run with COBC set to false. You will see this happen in both cases.

Time[0] on a minute chart is not going to have second granularity.

If you want to monitor the time of the current tick instead of the bar please use OnMarketData() http://www.ninjatrader.com/support/h...marketdata.htm

**NutFlush** · 08-08-2013, 12:41 PM

Originally posted by NinjaTrader_Lance View Post

Hello,

It does not matter if COBC is set to true or false. This is expected behavior and my tests were run with COBC set to false. You will see this happen in both cases.

Time[0] on a minute chart is not going to have second granularity.

If you want to monitor the time of the current tick instead of the bar please use OnMarketData() http://www.ninjatrader.com/support/h...marketdata.htm

Sorry Lance,
I don't mean to sound rude, but could you get a colleague to take a look at my posts? Somehow I fail to get my point across, but I don't know how else to explain the problem that you on the other hand fail to see.

Please take a look at the example of my very first post (last code area) which I repost here. As measured by my PC clock there were ticks coming in over a time span of 90 seconds that all got sorted into the same 1 minute bar. This cannot be expected behaviour, right?

Code:

#FFF# First tick in bar! [COLOR=DarkOrange]BIP: 1[/COLOR]; [COLOR=Purple][B]BarTime[/B][/COLOR]_Home-Time: 07.08.2013 [COLOR=Purple][B]18:00[/B][/COLOR]:00.000; PC-Time at Home: 07.08.2013 17:59:01.863; price: 168,83$ 
#???# Follow-up tick after other symbol started new bar! BIP: 0; BarTime_Home-Time: 07.08.2013 17:59:00.000; PC-Time at Home: 07.08.2013 17:59:08.318; price: 27,19$ 
#???# Follow-up tick after other symbol started new bar! BIP: 0; BarTime_Home-Time: 07.08.2013 17:59:00.000; PC-Time at Home: 07.08.2013 17:59:08.321; price: 27,19$ 
#FFF# First tick in bar! [B][COLOR=SeaGreen]BIP: 0[/COLOR][/B]; [COLOR=Purple][B]BarTime[/B][/COLOR]_Home-Time: 07.08.2013 [B][COLOR=Purple]18:00[/COLOR][/B]:00.000; PC-Time at Home: 07.08.2013 [COLOR=Blue][B]17:59:24.703[/B][/COLOR]; price: 27,19$  

// both symbols started the [B][COLOR=Purple]18:00[/COLOR][/B] bar, but XIV sent the first tick very late at 17:59:24.703  

#FFF# First tick in bar! [COLOR=DarkOrange]BIP: 1[/COLOR]; [B][COLOR=Purple]BarTime[/COLOR][/B]_Home-Time: 07.08.2013 [B][COLOR=Purple]18:01[/COLOR][/B]:00.000; PC-Time at Home: 07.08.2013 17:59:59.833; price: 168,82$  

// SPY moved to the [B][COLOR=Purple]18:01[/COLOR][/B] bar  

#???# Follow-up tick after other symbol started new bar! [B][COLOR=SeaGreen]BIP: 0[/COLOR][/B]; [COLOR=Purple][B]BarTime[/B][/COLOR]_Home-Time: 07.08.2013 [COLOR=Purple][B]18:00[/B][/COLOR]:00.000; PC-Time at Home: 07.08.2013 18:00:55.808; price: 27,18$ 
#???# Follow-up tick after other symbol started new bar! BIP: 0; [B][COLOR=Purple]BarTime[/COLOR][/B]_Home-Time: 07.08.2013 [B][COLOR=Purple]18:00[/COLOR][/B]:00.000; PC-Time at Home: 07.08.2013 18:00:55.811; price: 27,18$ 
#???# Follow-up tick after other symbol started new bar! BIP: 0; [B][COLOR=Purple]BarTime[/COLOR][/B]_Home-Time: 07.08.2013 [B][COLOR=Purple]18:00[/COLOR][/B]:00.000; PC-Time at Home: 07.08.2013 18:00:55.814; price: 27,18$ 
#???# Follow-up tick after other symbol started new bar! BIP: 0; [COLOR=Purple][B]BarTime[/B][/COLOR]_Home-Time: 07.08.2013 [B][COLOR=Purple]18:00[/COLOR][/B]:00.000; PC-Time at Home: 07.08.2013 18:00:55.817; price: 27,18$ 
#???# Follow-up tick after other symbol started new bar! BIP: 0; [B][COLOR=Purple]BarTime[/COLOR][/B]_Home-Time: 07.08.2013 [B][COLOR=Purple]18:00[/COLOR][/B]:00.000; PC-Time at Home: 07.08.2013 18:00:55.821; price: 27,18$ 
#???# Follow-up tick after other symbol started new bar! BIP: 0; [B][COLOR=Purple]BarTime[/COLOR][/B]_Home-Time: 07.08.2013 [B][COLOR=Purple]18:00[/COLOR][/B]:00.000; PC-Time at Home: 07.08.2013 18:00:55.824; price: 27,18$ 
#???# Follow-up tick after other symbol started new bar! BIP: 0; [B][COLOR=Purple]BarTime[/COLOR][/B]_Home-Time: 07.08.2013 [B][COLOR=Purple]18:00[/COLOR][/B]:00.000; PC-Time at Home: 07.08.2013 18:00:55.827; price: 27,18$ 
#???# Follow-up tick after other symbol started new bar! BIP: 0; [B][COLOR=Purple]BarTime[/COLOR][/B]_Home-Time: 07.08.2013[B][COLOR=Purple] 18:00[/COLOR][/B]:00.000; PC-Time at Home: 07.08.2013 18:00:55.830; price: 27,18$ 
#???# Follow-up tick after other symbol started new bar! BIP: 0; [B][COLOR=Purple]BarTime[/COLOR][/B]_Home-Time: 07.08.2013 [B][COLOR=Purple]18:00[/COLOR][/B]:00.000; PC-Time at Home: 07.08.2013 18:00:55.832; price: 27,18$ 
#???# Follow-up tick after other symbol started new bar! [B][COLOR=SeaGreen]BIP: 0[/COLOR][/B]; [COLOR=Purple][B]BarTime[/B][/COLOR]_Home-Time: 07.08.2013 [B][COLOR=Purple]18:00[/COLOR][/B]:00.000; PC-Time at Home: 07.08.2013 [B][COLOR=Red]18:00:55.834[/COLOR][/B]; price: 27,18$ 

// more  than 90 seconds after starting the 18:00 bar (at [COLOR=Red][B]17:59:24.703[/B][/COLOR])   

#FFF# First tick in bar! [B][COLOR=SeaGreen]BIP: 0[/COLOR][/B]; [B][COLOR=Purple]BarTime[/COLOR][/B]_Home-Time: 07.08.2013 [COLOR=Purple][B]18:01[/B][/COLOR]:00.000; PC-Time at Home: 07.08.2013 [B][COLOR=Blue]18:00:55.867[/COLOR][/B]; price: 27,18$  

// finally XIV moves also to the [B][COLOR=Purple]18:01[/COLOR][/B] bar

I apologize if I sound unfriendly in any way, which I don't mean to! I would gladly accept that I am wrong at some point, but I need to be certain first that my problem was understood.

Thanks for your time!

NutFlush

**NinjaTrader_Lance** · 08-08-2013, 02:52 PM

We've looked into this a bit further and ran this test using OnMarketData()

Code:

protected override void OnMarketData(MarketDataEventArgs e)
   {
    if( BarsInProgress == 1)
    {
     if (e.MarketDataType == MarketDataType.Last) 
     {
      if(e.Time != DateTime.Now)
      {
       Print("BIP: " + BarsInProgress + " ( " + Instrument.FullName + " ) ");
       Print("Computer Time: " + DateTime.Now);
       Print("Trade Time: " + e.Time);
       
      
      }
     }
     
    }
   }

There does appear to be an issue with XIV erroneously sending delayed prints

We are currently working with DTN to resolve this issue and will get back to you once I have more info.

**NinjaTrader_Lance** · 08-08-2013, 03:00 PM

Hello,

Please send me your log and trace files for today so that I may look into what occurred.

You can do this by going to the Control Center-> Help-> Mail to Support.

Please reference the following ticket number in the body of the email: 896578 ATTN Lance

Please also include your Kinetick ID

I look forward to assisting you further.

**NutFlush** · 08-08-2013, 03:47 PM

Originally posted by NinjaTrader_Lance View Post

We've looked into this a bit further and ran this test using OnMarketData()

Code:

protected override void OnMarketData(MarketDataEventArgs e)
   {
    if( BarsInProgress == 1)
    {
     if (e.MarketDataType == MarketDataType.Last) 
     {
      if(e.Time != DateTime.Now)
      {
       Print("BIP: " + BarsInProgress + " ( " + Instrument.FullName + " ) ");
       Print("Computer Time: " + DateTime.Now);
       Print("Trade Time: " + e.Time);
       
      
      }
     }
     
    }
   }

There does appear to be an issue with XIV erroneously sending delayed prints

We are currently working with DTN to resolve this issue and will get back to you once I have more info.

Hi Lance,

thanks for looking into that further. An error in the data stream would explain the strange occurences. Please tell DTN that more symbols than just XIV might be affected. Actually SPY which I used as the second symbol also showed "follow-up" ticks.

I will send you the log files tomorrow morning.

Cheers,
NutFlush

Topics	Statistics	Last Post
Issue with Stop Loss by alifarahani Started by alifarahani, Today, 09:40 AM	4 responses 21 views 0 likes	Last Post by alifarahani Today, 02:41 PM
Is it possible to have a drawing, such as a line, cycle through different colors? by gentlebenthebear Started by gentlebenthebear, Today, 01:30 AM	3 responses 16 views 0 likes	Last Post by NinjaTrader_Jesse Today, 02:41 PM
New Y-Axis Column by PhillT Started by PhillT, Today, 02:16 PM	2 responses 7 views 0 likes	Last Post by PhillT Today, 02:33 PM
DrawObjects Backtesting Issue by Kaledus Started by Kaledus, Today, 01:29 PM	3 responses 11 views 0 likes	Last Post by NinjaTrader_Jesse Today, 02:24 PM
Getting Error While Trying To Login by frankthearm Started by frankthearm, Yesterday, 09:08 AM	14 responses 47 views 0 likes	Last Post by NinjaTrader_Clayton Today, 02:19 PM

Announcement

Partner 728x90

comparing bar time on 2 symbols: ticks not matching

comparing bar time on 2 symbols: ticks not matching

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Latest Posts