Two GLOBEX issues

[saxon987]

#1) There is a data inconsistency in NQ H5 for this morning.

Look at the data from 9:25am CT to about 10:00am CT. The RT data that I load (eSig) appears to match what the CME shows. However, the static data that I downloaded (eSig) is different. Perhaps this was caused by the big sell-off in NQ that occured at 9:30am CT.

#2) I trade and use eSig RT data for 3 Globex currencies: Euro/Fx (6E), Canadian Dollar (6C), and Japanese Yen (6J). For some reason the loading time for the EuroFx data is about 3 times as long as for the other 2...or for ANY other Globex contract that I have seen...even though I am loading the same number of bars for each. Why would Euro/Fx be slower than the others? Does this indicate a possible problem with your feed from the CME?

Thanks,
saxon987

[ScottJ]

Hi,

ReL #1, please post this static data you are referencing so we can compare to the data on our servers.

Re #2, our intraday bars are built from tick data. If you compare the volume from the 3 contracts referenced, you'll see that that 6E contract has 2-3 times the volume of the other two, hence more data to sort through to build bars.

Thanks.

[saxon987]

Data mismatch

ScottJ,

Here are the data sets (1 min bars). I have starred the lines where there are discrepancies in price and/or volume.

Unfortunately, I am now also receiving the (presumably) erroneous set via the Realtime feed as well, in my latest load. But I saved these sets previously.


Realtime Data

NQ H5 Time Open High Low Close Volume
1/3/2005 9:26:00 AM 1621.5 1622.5 1621 1622.5 2276
1/3/2005 9:27:00 AM 1622.5 1622.5 1619 1619.5 3380 +++
1/3/2005 9:28:00 AM 1620 1621 1619 1620 1679 +++
1/3/2005 9:29:00 AM 1620.5 1621 1620 1620 1921 +++
1/3/2005 9:30:00 AM 1620 1621 1618.5 1618.5 3354 +++
1/3/2005 9:31:00 AM 1618.5 1619 1616 1616 3696 +++
1/3/2005 9:32:00 AM 1616 1616.5 1614.5 1615.5 2836 +++
1/3/2005 9:33:00 AM 1615.5 1617.5 1615 1616.5 3314 +++
1/3/2005 9:34:00 AM 1616.5 1616.5 1614.5 1615 2185
1/3/2005 9:35:00 AM 1615 1615.5 1613 1613.5 2364
1/3/2005 9:36:00 AM 1613.5 1613.5 1608 1608.5 5904 +++
1/3/2005 9:37:00 AM 1608.5 1610 1607.5 1608.5 3487 +++
1/3/2005 9:38:00 AM 1608.5 1608.5 1605.5 1606.5 3471 +++
1/3/2005 9:39:00 AM 1607 1607.5 1605 1607 3687 +++
1/3/2005 9:40:00 AM 1607 1609 1606.5 1609 2908 +++
1/3/2005 9:41:00 AM 1608.5 1609.5 1608 1609.5 2289



Static Data

NQ H5 Time Open High Low Close Volume
1/3/2005 9:26:00 AM 1621.5 1622.5 1621 1622.5 2276
1/3/2005 9:27:00 AM 1622.5 1622.5 1619 1620 3610 +++
1/3/2005 9:28:00 AM 1620 1621 1619 1620.5 1782 +++
1/3/2005 9:29:00 AM 1621 1621 1620 1620 1614 +++
1/3/2005 9:30:00 AM 1620 1621 1619.5 1619.5 2753 +++
1/3/2005 9:31:00 AM 1619.5 1619.5 1616 1616 4348 +++
1/3/2005 9:32:00 AM 1616.5 1616.5 1614.5 1615.5 3353 +++
1/3/2005 9:33:00 AM 1615.5 1617.5 1615.5 1616.5 3412 +++
1/3/2005 9:34:00 AM 1616.5 1616.5 1614.5 1615 2185
1/3/2005 9:35:00 AM 1615 1615.5 1613 1613.5 2364
1/3/2005 9:36:00 AM 1613.5 1613.5 1607.5 1608 7360 +++
1/3/2005 9:37:00 AM 1608.5 1610 1607 1607.5 3869 +++
1/3/2005 9:38:00 AM 1607.5 1607.5 1605.5 1607 4046 +++
1/3/2005 9:39:00 AM 1607 1607.5 1605 1607 3736 +++
1/3/2005 9:40:00 AM 1607 1609 1606.5 1609 2513 +++
1/3/2005 9:41:00 AM 1608.5 1609.5 1608 1609.5 2289

[JayF]

Could you describe the process you are using to download or export the static data you reference? Is this through the Advanced Charts' Data Export tool? Through a 3rd party software? Through our API?

I just used the Advanced Charts' Data Export tool, and compared the exported data to the data I see in the Advanced Chart. It seems to match.

[saxon987]

Jay,

Check ALL your servers. Reference data from the exchange. I've done all I can do.

saxon987

[JayF]

We were able to find differences after checking multiple servers against each other. We are researching into the cause of this, and will reply with further findings.

[JayF]

There was indeed discrepancies during that time across some of our servers. This was actually a known and much more vast problem ~6 months ago, prior to us starting a synchronization process that brings all the Tick Servers to have the same data sets. After recently analysis, it was determined that the synchronization process was not keeping up under heavy market conditions, and the data would slip out of sync across servers.

I'm happy to say that we are already in process of releasing a new generation of servers that take advantage of many of the recent advances in microprocessors. With this upgrade, the servers will be able to stay nearly perfectly synchronized.

We expect the last of these servers to rolled out later in the week (as early as tomorrow.)