Corrupt Exported Tick Ask Price on Split Adjusted Data (and other bugs)

[Reederoo]

I reported 2 bugs to platform support about 2 weeks ago and so far they have not been able to reproduce them. I am not doing anything very unusual and reproduced them myself on a new install on a new machine. I wonder if anyone else could follow the steps below and let me know if they also see the issues, or if it is really just me and my 2 computers that do this.

The bugs are:
1) Ask price of exported tick data on days preceding a split for instruments with split adjustment turned on is corrupt (exactly $0.51 above bid in the case of NUGT for data before 8/25/2016 - should be variable and between 2 and 4 cents above bid typically for this time period after split adjustment) See attached exported data.
2) The rightmost vertical scrollbar doesn't work on edit tab of historical data window after expanding to see tick data on right panel.

(this one was not reported to platform support yet - just discovered today)
3) Split event data is not downloaded along with price data when using download button on historical data window, which results in failure to adjust for splits in exported data. User must manually update split data through Tools > Instruments > NUGT > Splits & Dividends-> Update

Here is exacly what I did to reproduce these:

1) Fresh install of NinjaTrader 8.0.4.0 on a new PC (Windows 10) with all default choices and entered my Kinetick account information in order to download tick data. Enabled the Kinetick account.
2) checked that options->Market->Adjust for splits - intraday data is enabled.
3) Tools->export->Historical data->NUGT 8/22/2016-8/25/2016 Download tick data. Export same time period tick data.
4) checked exported file and found no split adjustment had been applied and no corrupt ask prices (just as in your case).
5) checked Tools > Instruments > NUGT > Splits and found 0 splits. After clicking "Update" found 2 splits (last one on 8/25/2016)
6) exported data again (historical data window still up from last time with same settings as last export).
7) checked exported file and found that prices are now adjusted for splits (5x smaller than before for all days before 8/25/2016).
8) Also found that exported data now has corrupt ask prices (exactly $0.51 above bid for all ticks)
9) clicked "Edit" tab of historical data window. Clicked Historical->NUGT->Last->Tick->August 2016->8/25/2016->7:00 AM (13675 Items). Then verified that the vertical scroll bar immediately right of 7:00 AM works fine. Tried the furthest right vertical scroll bar to the right of the tick data and found that it does not work.


So to summarize, I am able to replicate both bugs on a fresh install on a new PC with no custom settings other than what I listed above. In addition, this process has highlighted what I think is a 3rd bug, in that the splits for NUGT did not automatically update when I downloaded the tick data in the historical data window, which caused the export data process to fail to adjust for splits even though split adjustment was enabled. The user must manually update the splits as described above before split adjustment works properly. Shouldn't the splits get updated without manual intervention whenever data is downloaded for a security?

Any help in confirming these would be great so that platform support will be able to move on to fixing them instead of claiming they don't exist.

[Reederoo]

A More Recent Example using EES

The 120 day limitation for tick data makes my original example currently unreproducible, since the split for NUGT was on 8/25/2016. Instead, the issue can also be reproduced using EES with the recent 3:1 split on 2/6/2017.

I followed the same steps as originally posted exactly, except for I used EES instead of NUGT, and the download / export dates I used are 02/02/2017 to 02/08/2017. The Ask data is corrupt, with a price exactly $0.85 above the bid price for all ticks earlier than 02/06/2017, and after than has the expected variable spread between about $0.10 - $0.20. When split adjustment is turned off the ask data always has the expected variable spread, although also not adjusted for the split. See the attached exported data file.


I also checked that both other issues are still reproducable (failure to automatically download splits when downloading tick data, and broken vertical scrollbar).

In addition, I found a 4th issue. When updating splits for EES, the factor that was downloaded was 3.030303030303030, where it should have been 3 exactly. I manually corrected it to be 3 before exporting the data in the attached file, but isn't this incorrect behavior?


I'd appreciate it if anyone could test these and comment on all 4 issues. I can easily reproduce all of these on my main PC, and on a fresh install of NT8 on another Windows10 machine in about 15 minutes, so it should not be difficult for others to do the same.


Thanks,


Reed

[NinjaTrader_RileyS]

Reederoo, thanks for posting!

I have received your email and will be testing this today and will follow up with you as soon as I have analyzed this further.