ActiveTick Quotes That Are Missing from Polygon Data

I am comparing the historical tick data from ActiveTick and from Polygon.
I noticed that ActiveTick has some quotes that are missing from Polygon.
Don't know if Polygon was missing quotes, or if ActiveTick was adding fictional ones.

Does anyone have another data source to look at, to see if your data confirms that from ActiveTick or from Polygon?

Thanks.

Quotes from ActiveTick for IBM on 12/7/2020:

timestamp bid ask sizes exchanges
Q 103026:251 126.29 126.32 2 1 Q X
Q 103029:994 126.29 126.32 2 2 Q Q <-------- this one is not found in Polygon data
Q 103030:660 126.29 126.32 1 2 Q N
...... ......
Q 104129:962 126.7 126.72 1 1 N Q
Q 104129:986 126.7 126.72 2 1 N Q
Q 104129:995 126.7 126.72 3 1 Q Q <-------- this one is not found in Polygon data
Q 104130:003 126.71 126.72 2 1 N Q
Q 104130:004 126.71 126.73 2 2 N P

 

I believe both are correct. Polygon simply ignores BBO from the same exchange.

 

Polygon does not omit quotes where bid_exch == ask_exch, if this is what you mean. Examples below (bid_exch == ask_exch == 12), from Polygon data:

786070801|104116.446212000|_|104116.446423808 126.72 126.75 4 2 12 12 1~81 1
787771901|104129.949740544|_|104129.949773056 126.72 126.74 3 1 12 12 1~81 1

 

Hmm in that case I'm not sure. Probably a representative from Polygon can help you out here.

 

It would be awesome if you could report back after you have spoken with polygon. I am quite interested in the outcome of this discrepancy. Sorry I can't help, I don't use neither data vendor.

 

I fully reported everything to Polygon. They said they would look into it. But I never heard back. I can understand. They just pass data along. They do not have another data source to compare against.

 

So this I believe comes down to NBBO appendages in the actual packets. We only send NBBO ticks which have an NBBO appendage with content in them. BBOs which do not have an appendage do not get broadcast. Different providers can vary on this. ( see pg 46 of https://www.ctaplan.com/publicdocs/ctaplan/CQS_Pillar_Output_Specification.pdf for NYSE specification )

We take the same approach as NYSE TAQ data in determining which is an "NBBO" quote, which we do QA checks against.

Hopefully this answers your question

 

Hi Quinton,

Thanks for getting back. I wish things were that simple. Below, I listed the complete sequence of NBBO ticks from both vendors for SREpA on 12/4/2020, at around 10:04AM. Notice that "missing" quotes improved markets. So this cannot be simply a difference in handling by vendors. One of these vendors has technology issues. Either Polygon dropped ticks or ActiceTick inserted non-existent ones.

The reason I sound very confident is that all such discrepancies have a very clear pattern. For each ticker symbol I looked, such discrepancies started with hhmmss:99* stamps (e.g. 993 in the example below). Then they become hhmmss:994, then hhmmss:995, as the day progresses. I will have the a partial list of such discrepancies below as well, for this day, for SREpA. So this proves, all but certain, that this is a technology issue.


~~~ Complete Polygon Sequence ~~~
quoteID timeStamp@Exch timeStamp@SIP bid ask sizes exchs condCodes
692771501|100405.503591168|_|100405.503623936 103.52 103.72 20 10 11 10 1~81 1
695662901|100415.905467648|_|100415.905503232 103.52 103.74 20 7 11 19 1~81 1
695663901|100415.912047360|_|100415.912081408 103.52 103.7 20 10 11 10 1~81 1
704115201|100449.444697856|_|100449.444733440 103.53 103.71 1 7 10 19 1~81 1
710917401|100513.115634176|_|100513.115668992 103.53 103.69 1 7 10 19 1~81 1

~~~ Complete ActiveTick Sequence ~~~
Q 100405:503 103.52 103.72 20 10 P N
Q 100415:905 103.52 103.74 20 7 P Z
Q 100415:912 103.52 103.7 20 10 P N
Q 100429:993 103.53 103.7 1 10 N N <--- missing in Polygon (this improves the bid)
Q 100449:444 103.53 103.71 1 7 N Z
Q 100459:993 103.53 103.69 1 10 N N <--- missing in Polygon (this improved the offer)
Q 100504:116 103.53 103.71 1 7 N Z <--- missing in Polygon
Q 100513:115 103.53 103.69 1 7 N Z

~~~ Partial list of "missing" quotes (notice the migrating milli-sec stamp patterns) ~~~

hhmmss.xxx bid ask bSz aSz
100359993 103.55 103.72 1 10

100429993 103.53 103.7 1 10

100459993 103.53 103.69 1 10

100504116 103.53 103.71 1 7

100529993 103.55 103.73 1 10

100659993 103.57 103.73 2 10

100729993 103.57 103.68 2 10

100859993 103.55 103.74 2 10

101029993 103.61 103.79 1 10

101059994 103.61 103.81 1 10

101129994 103.61 103.78 1 10

101159994 103.61 103.75 1 10

101229994 103.61 103.76 1 10

101245637 103.61 103.78 1 7

101259994 103.61 103.71 1 10

101329994 103.56 103.67 1 10

101359994 103.54 103.67 1 10

101429994 103.45 103.63 1 10

101459994 103.48 103.63 1 17

101529994 103.48 103.62 1 17

102429994 103.29 103.58 10 17

102459994 103.31 103.57 10 17

102559994 103.33 103.57 10 17

102629994 103.34 103.6 10 17

102659994 103.32 103.59 10 17

102729994 103.3 103.56 10 17

102829994 103.37 103.65 10 17

102859994 103.4 103.69 10 10

102929994 103.44 103.69 10 10

102959994 103.48 103.75 10 10

103059994 103.53 103.78 10 10

103129994 103.54 103.78 10 10

103159994 103.48 103.73 10 10

103229995 103.39 103.68 10 10

103259995 103.39 103.67 10 10

103329995 103.41 103.67 10 10

103429995 103.39 103.67 10 10

103459995 103.42 103.69 10 10

103529995 103.42 103.68 10 10

103559995 103.44 103.73 10 10

103629995 103.44 103.69 10 10

103659995 103.4 103.65 10 10

103729995 103.35 103.62 10 10

103759995 103.31 103.59 10 17

103859995 103.39 103.66 10 10

103929995 103.42 103.68 10 10

104029995 103.43 103.68 10 10

 

Hmm. If I remember correctly, TAQ data does not include NBBO when the BBO comes from the same exchange.

To Polygon defense however, there are a lot of quotes from local exchanges as well. Consolidating and pick which quotes to show up as BBO are complicated (Not as simple as choosing best bid/ask price). So it's unlikely that either ActiveTick or Polygon miss/add non-existent quotes. The difference is in the approach of choosing which quotes to show as NBBO.

 

Isn't that like saying that "it depends on how each of us define traffic rules, for some red means stop for others Red means go"? In the end of the day standards serve a purpose and the purpose is defeated if we don't adhere to those standards. What constitutes NBBO is perfectly defined in its finest grained details. The only question here is whether a data vendor understands and adheres to those precise definitions or not.