NT8 COT Indicator Problem?

[trubolotta]

I have been working on a custom COT based indicator for my son to meet his requirements. In the course of development, I have found some discrepancies between my COT indicator and the NT8 stock COT indicator. I have checked the data I'm using from CFTC against the data used by the stock indicator and it is the same. I'll give two examples of my findings based on the OpenInterest reports.

1) The CFTC data is not always complete and there can be significant gaps. In my indicator, those show as breaks in the graph where no data is plotted. In the stock indicator, those gaps have been filled in. I see no explanation in the c# code for the stock indicator of how those gaps are managed. I think it would be useful to explain how the gaps are handled.

2) I'll illustrate the next problem by example. I used OpenInterest data for testing. In my indicator, the CFTC data with an effective date of Tuesday, December 15, 2020, the latest available as of this morning (12/28/2020) is matched to the previous Friday, December 11, 2020. It is my understanding that the CFTC data, usually effective on a Tuesday applies to the previous week. In the stock COT indicator, the data for December 15 is matched against Friday, December 18. This shows up when I compare graphs as a one week shift in all the data points in the graphs.

Is this right? Am I mistaken?

[NinjaTrader_Kate]

Hello trubolotta,

Thank you for your post.

The built in COT indicator will plot the reported data forward until the next update of COT data has been received. The reason it's plotted forward is so that the data will be available on all time frames, otherwise it would just be a single plot on Tuesday with nothing in between.

COT data is released each Friday at 3:30pm and contains data from Tuesday of that week. So I would expect data from last Tuesday and Friday to be the same. Since the COT data projects forward until new data is available, the current COT data will plot forward until Friday when it is released, then it will be updated for this Tuesday going forward.

Please let us know if we may be of further assistance to you.

[trubolotta]

I see what you are saying but I question whether or not that is giving an accurate accounting. My understanding is that the data from the previous week is compiled into a file on the following Tuesday but not released until the following Friday. Apparently, it goes through some kind of review process while being compiled and before being released. While it is being released on a Friday, the data always applies to the week before, usually ending on the previous Friday. Since I'm using the OnBarUpdate event on weekly charts for calculations, I am matching the as-of-date of the data, usually a Tuesday, to the Time series date of the previous Friday, which is the end of the week for which the data actually applies. It seems to me that taking last weeks values and projecting them forward is not accurate. Does that have any negative consequences? I'm not the analyst so I've booted the observation over to my son.

[NinjaTrader_Kate]

Hello trubolotta,

Thank you for your reply.

The COT report provide a breakdown of each Tuesday's open interest as noted on the publicly available CFTC page regarding COT here:

https://www.cftc.gov/MarketReports/CommitmentsofTraders/index.htm

So, it's not a compile of the data for the whole week, it's a snapshot of where things are at that moment on Tuesday. Technically that means the data is only accurate for that specific Tuesday. The only reason it plots forward is so that users can see the data on any time frame. Otherwise, it would only really be visible on a weekly chart, which arguably is how it should be viewed, but many people still wish to see the values on smaller time frames.

Please let us know if we may be of further assistance to you.

[trubolotta]

Thank you. That makes it clear I was misinformed. Fortunately, I have only one line of code to change to make it right.