Home >> Support Board

Support Board

Date/Time: Thu, 02 May 2024 10:18:16 +0000

How to Post a Message for Support

Bid/Ask Depth Bars Study: Differences Between New (>= 1949) and Old Version

View Count: 3013

[2019-08-22 19:32:43] #1 [Go To First Post]
User41727 - Posts: 124
I am running the aforementioned study on equities with the SC data feed, i.e. I can only use top of the book quotes. I am seeing large discrepancies between the values that the old implementation was returning compared to those calculated by the new one to the point where both are regularly telling me the opposite at the same point in time. I have ASCIL code that relies on those values, and on backtesting those signals with the new version, the results are significantly worse, bordering on useless. If I re-install a pre-1949 version, the results are as expected again even when using historical depth data that had been collected with a newer version. Hence, the issue must be with the way the new calculation is performed, because the data is clearly still there. Therefore, could you please at least document the exact nature of the changes in the depth bar ratio calculation, so as to enable me to understand what is going on, and maybe find a workaround for this issue? Or, ideally, add a "legacy" calculation mode to the study that performs as the old version would have.

[2019-08-23 22:12:56] #2 [Go To First Post]
User41727 - Posts: 124
This is an extremely important issue for me. Could I please get a response to this?

[2019-08-26 21:16:29] #3 [Go To First Post]
User41727 - Posts: 124
Bump. I really need an answer to this.

[2019-08-27 05:02:38] #4 [Go To First Post]
Sierra Chart Engineering - Posts: 104368
What we can say is that the new implementation of Bid and Ask Depth Bars is very accurate. What we will do is set up a test to validate that. We cannot take the time to compare it to an older version. Once we validate that it is working properly according to documentation that is all that we can do. We will try to get this done in the next few days. Also in the latest versions of Sierra Chart Bid and Ask Depth Bars does now work with one level of market depth. In other words the best bid and ask data. Therefore, could you please at least document the exact nature of the changes in the depth bar ratio calculation, The calculations are identical. The difference really would be in the frequency of the calculation which would be more in the newer versions but in the case of the best bid and ask data only, there should not be any difference. So we have to look into that. Sierra Chart Support - Engineering Level Your definitive source for support. Other responses are from users. Try to keep your questions brief and to the point. Be aware of support policy: https://www.sierrachart.com/index.php?l=PostingInformation.php#GeneralInformation For the most reliable, advanced, and zero cost futures order routing, change to the Teton service: Sierra Chart Teton Futures Order Routing Date Time Of Last Edit: 2019-08-27 05:06:06

[2019-08-27 08:24:00] #5 [Go To First Post]
NotIll - Posts: 118
I am having the same problems with the latest update. I can not get my head around what changed. It is as if the market changed the way it works after the update(which in this case as the topic starter tested is not the case).

[2019-08-27 11:36:50] #6 [Go To First Post]
NotIll - Posts: 118
Please look at the image I attached. It is the same market and same data, yet the newest version just changes the bid ask data. what is going on here? The difference on the big up thrust is also very much telling. It cannot be that an update does this to actual market data. I have looked and the changes happend after 1941 which is the same what User41727 has found. please reverse this. Date Time Of Last Edit: 2019-08-27 13:19:09


SC BID ASK DEPTH.png / V - Attached On 2019-08-27 11:35:51 UTC - Size: 199.12 KB - 482 views

[2019-08-27 15:41:34] #7 [Go To First Post]
User41727 - Posts: 124
What we can say is that the new implementation of Bid and Ask Depth Bars is very accurate. I don't doubt that. The values for the ratio calculation seem to be as expected with the new version based on the bid size to ask size ratio, but they are still often enough notably different from what the old version would calculate. I have looked into this a bit more, and I was using "Historical Market Depth - Last" as "Source Data" with the old version. From a cursory glance, the values calculated by the old version seem to be identical to the new version when using "Time & Sales" instead. Unfortunately, the documentation was never particularly clear on what exactly using the historical market depth as data source actually meant, and how exactly it differs from using T&S when only L1 quotes are available anyway. Maybe you could shed some light on that? In any case, I would greatly appreciate it if you could implement a way for the new study version to behave just as the old one did with the depth data. @NotIll It is as if the market changed the way it works after the update This is exactly the sense that I get. I have rolled back to 1943, which is still using the old study implementation, and seems to be working as expected. Date Time Of Last Edit: 2019-08-27 15:45:50

[2019-08-27 16:01:48] #8 [Go To First Post]
NotIll - Posts: 118
@User41727 How do you select ''time & sales'' in the new version in the indicator? In the new version when in ''settings & inputs'' of the indicator the option: ''source data'' to select time and sales as source was no longer there. Or am I missing something here?

[2019-08-27 16:15:03] #9 [Go To First Post]
User41727 - Posts: 124
I meant that for the old version. In the new one the option is not there anymore, and based on my observations I would assume that that version always operates in T&S mode, which is really the issue here. From what I can gather, in that mode the study will calculate the ratio of bid and ask size for the best bid and offer only at the time the respective bar closes. If you use "market depth" as source, the study appears to be taking additional data into account, possibly the observed bid and ask size at other levels, but that is the part that I am not clear on. If you switch the old version to T&S, you should see the same values as with the new version.

[2019-08-27 16:25:07] #10 [Go To First Post]
NotIll - Posts: 118
Right I understand what you are saying. And from that I would assume that it should be fixable.

[2019-08-27 16:32:27] #11 [Go To First Post]
User41727 - Posts: 124
I would assume so, because when running the old version on depth data that was collected by a new SC version, the calculations are correct again, which would suggest that the data required for the old-style calculation is still there. I'm not sure if there is an ACSIL API to access that data. If there is, it should be possible to re-implement the old version ourselves, but I'm hoping that SC might help us out here.

[2019-08-27 16:55:56] #12 [Go To First Post]
NotIll - Posts: 118
I will be patient.

[2019-08-27 21:52:11] #13 [Go To First Post]
Sierra Chart Engineering - Posts: 104368
In the case of when using CQG also be aware of this change at this version: https://www.sierrachart.com/index.php?page=doc/Whats_New.php#SCVer1959 We are also going to be running a detailed test over the next couple days with Bid and Ask Depth Bars to validate its accuracy. This involves just simply logging the underlying data and doing the calculations manually. Sierra Chart Support - Engineering Level Your definitive source for support. Other responses are from users. Try to keep your questions brief and to the point. Be aware of support policy: https://www.sierrachart.com/index.php?l=PostingInformation.php#GeneralInformation For the most reliable, advanced, and zero cost futures order routing, change to the Teton service: Sierra Chart Teton Futures Order Routing

[2019-08-27 21:58:05] #14 [Go To First Post]
User41727 - Posts: 124
Did you see what I wrote about this? It's not a question about accuracy of the new version. The issue is that that version always behaves like the old version would have with data source "Time & Sales", and does not have an equivalent to using data source "historical market depth", which causes the discrepancies.

[2019-08-28 06:03:52] #15 [Go To First Post]
NotIll - Posts: 118
+1 for post #14. Just and example: It is easy to see market participants accumelating large quantities of orders on the limit in the old version, just by eye. this is the behaviour we want to be able to see for lets say SP500(/ES) and is known behaviour. This example is one of the things I was looking for when using the indicator and the idicator did its job perfectly. What I am saying is that the indicator was correct for a trader as it was, because it showed known market behaviour traders are looking for. Traders could make money just be a simple chart and the old version of the indicator. thats what we want right? That is gone now. Please consider the posts of User41727. The data is still available and works for the new version, just the option is missing now. We are not asking for anything new and we are alsno not asking for you to get rid of the newer version. They could live happily side by side. Also I use CQG, but the indicator is acting correct when using the old version. It becomes near useless with the new version(this would have happend with any data supplier). So the data we are getting trough CQG is not causing any problems with the old version of SC. Date Time Of Last Edit: 2019-08-28 06:11:28

[2019-09-05 12:23:50] #16 [Go To First Post]
NotIll - Posts: 118
Kindly bumping to know if the ''legacy'' suggestions(old version and new version side by side) of the topic starter is beeing considerd. Keep up the good work!

[2019-09-09 15:12:14] #17 [Go To First Post]
User41727 - Posts: 124
Any news on this topic? A response from support would be greatly appreciated.

[2019-09-09 15:25:23] #18 [Go To First Post]
Sierra Chart Engineering - Posts: 104368
Yes we are going to follow up as soon as we can. Not sure if we can get to this today, but hopefully by the end of the day tomorrow . We are still doing testing. Sierra Chart Support - Engineering Level Your definitive source for support. Other responses are from users. Try to keep your questions brief and to the point. Be aware of support policy: https://www.sierrachart.com/index.php?l=PostingInformation.php#GeneralInformation For the most reliable, advanced, and zero cost futures order routing, change to the Teton service: Sierra Chart Teton Futures Order Routing

[2019-09-09 15:29:05] #19 [Go To First Post]
User41727 - Posts: 124
Again, did you see my explanation? This is not about accuracy! It's about a feature that was present in the previous version, but is no longer in the current one.

[2019-09-09 15:47:40] #20 [Go To First Post]
Sierra Chart Engineering - Posts: 104368
Yes we do plan to respond to that. Understand, that your questions are very complex to answer. And we cannot be expected to take these complex questions on demand and give immediate answers. It could literally be weeks before we can get to even answering them. And that is reasonable. We have enormous amounts of work to do. We have to balance user questions with development. It is far better, sometimes that we just don't answer and focus more on development and get long-overdue work done. It is not good when we take years to get certain task done, and we are wasting time on this board day after day dealing with all kinds of garbage. We are looking at this, and we are going to analyze what you have said, and we will determine the best action and we may very well put back the old functionality that uses the historical market depth data as an option. Sierra Chart Support - Engineering Level Your definitive source for support. Other responses are from users. Try to keep your questions brief and to the point. Be aware of support policy: https://www.sierrachart.com/index.php?l=PostingInformation.php#GeneralInformation For the most reliable, advanced, and zero cost futures order routing, change to the Teton service: Sierra Chart Teton Futures Order Routing Date Time Of Last Edit: 2019-09-09 15:48:04

[2019-09-09 16:16:57] #21 [Go To First Post]
User41727 - Posts: 124
What you have to understand is that if you do not at least acknowledge what I wrote that I cannot be sure that you are even taking these things into consideration. And this is a critical issue for me to the point where newer versions of the platform are effectively broken for my purposes. Hence, if I cannot be sure that you at least take this seriously, I need to find another solution. So, a quick: "We get what you are saying, and we have put it on our agenda" would be more than helpful. For the time being, I don't think there is a way to access the data store in the .depth files manually via ACSIL, is there? Because if there is, I might be able to solve this on my own.

[2019-09-18 20:09:18] #22 [Go To First Post]
Sierra Chart Engineering - Posts: 104368
For the time being, I don't think there is a way to access the data store in the .depth files manually via ACSIL, is there? Yes there is. Refer to: Market Depth Data File Format Although this may not be very efficient. Sierra Chart Support - Engineering Level Your definitive source for support. Other responses are from users. Try to keep your questions brief and to the point. Be aware of support policy: https://www.sierrachart.com/index.php?l=PostingInformation.php#GeneralInformation For the most reliable, advanced, and zero cost futures order routing, change to the Teton service: Sierra Chart Teton Futures Order Routing

[2019-09-20 15:06:08] #23 [Go To First Post]
Sierra Chart Engineering - Posts: 104368
We recommend using version 1989 or higher and see if you notice any difference. We have verified the calculations are accurate as of this version. If you are still not satisfied with the new study we can put back the old study, for use only with the historical market depth data. However, not using the time and sales data. This will be the fastest solution rather than us going through and examining differences between the old and new version of Bid and Ask Depth Bars. Sierra Chart Support - Engineering Level Your definitive source for support. Other responses are from users. Try to keep your questions brief and to the point. Be aware of support policy: https://www.sierrachart.com/index.php?l=PostingInformation.php#GeneralInformation For the most reliable, advanced, and zero cost futures order routing, change to the Teton service: Sierra Chart Teton Futures Order Routing Date Time Of Last Edit: 2019-09-20 15:08:34

[2019-09-20 15:16:50] #24 [Go To First Post]
User41727 - Posts: 124
If you have made no changes to how the study calculates its values in recent version, then I can tell you right here and now without even testing it, that this does not solve my problem. So, please restore the old version of the study.

[2019-09-21 00:05:34] #25 [Go To First Post]
User41727 - Posts: 124
I can definitely confirm that 1989 still behaves as the recent post-1949 versions in terms of the values. The calculation may be accurate, but it is simply the wrong way to perform the calculation for my purposes. This will be the fastest solution rather than us going through and examining differences between the old and new version of Bid and Ask Depth Bars. The differences are pretty simple really: The current implementation only allows the user to use the last quote (via T&S) for the calculation. The old version gave us three options: 1) use last quote (via T&S) 2) use maximum seen depth at each level up to the specified maximum of levels 3) use last seen depth at each level up to the specified maximum of levels. The latter two options are now missing, and that is the issue.

To post a message in this thread, you need to log in with your Sierra Chart account: