Page 1 of 1

Issues with CSI

Posted: Tue Feb 08, 2011 5:37 am
by CyTrend
Hello guys I am assuming some of you have CSI can some of you please vouch for their quality? I have been trying to figure out exactly how their perpetual future works and had difficulty getting a clear ans from them. Further to that they seem to not have the expiry dates for some of the future contracts and they have and have told me they have difficulty getting them. I can get them for free at CME.com. For instance the expiry date is non-existent for SB( sugar). I felt that I was asking very trivial questions and getting not very satisfactory answers. Has anyone else felt the same way? Can anyone suggest any other good automated solution for the data, that also has good continuous contracts?

Tx,
Cy

Re: Issues with CSI

Posted: Tue Feb 08, 2011 8:25 am
by AFJ Garner
CyTrend wrote:Hello guys I am assuming some of you have CSI can some of you please vouch for their quality?

Tx,
Cy
Name me an alternative vendor with a sophisticated front end to build continuous contract data in a flexible and efficient manner and I too will be grateful. So far, CSI seems as good as it gets without building your own software.

Posted: Tue Feb 08, 2011 9:35 am
by CyTrend
this is what i am realising so far myself just wanted to put it out there to see the response... I guess many ppl do it themselves. Apparently bloomberg is going to start rolling out a backtesting platform to match CQG

Posted: Tue Feb 08, 2011 9:44 am
by AFJ Garner
CyTrend wrote:this is what i am realising so far myself just wanted to put it out there to see the response... I guess many ppl do it themselves. Apparently bloomberg is going to start rolling out a backtesting platform to match CQG
I was not aware that CQG provided a "back testing" platform. Or at least not in the ultra sophisticated sense which the fortunate Trading Blox user has come to love and appreciate.

Posted: Wed Aug 31, 2011 5:47 am
by adamant
The bllomberg backtesting platform is very limited compared to blox (though for an amateur like me, more than sophisticated enough). It comes in two versions. If you are a BB user and interested, type STDY (GO) for access to these tools.

Re: Issues with CSI

Posted: Mon Sep 10, 2012 1:25 pm
by sunyata
AFJ Garner wrote: Name me an alternative vendor with a sophisticated front end to build continuous contract data in a flexible and efficient manner and I too will be grateful. So far, CSI seems as good as it gets without building your own software.
I just bought historical data only through CSI and they are telling me that I need to have all the data in the form I want it saved to my computer in 30 days, at which time I lose access to Unfair Advantage and its ability to create continuous contracts. Is this true from yours and others' experience? I have to rush to figure out when I want to roll everything and after then, I still can't test if these are optimal because I will no longer have access to UA. I guess I need to purchase ongoing data access, but I don't need that right now since I am in the design stages.

Seems pretty shitty to me, since I bought the data already. :(

Posted: Mon Sep 10, 2012 2:04 pm
by sluggo
I just bought historical data only through CSI and they are telling me that I need to have all the data in the form I want it saved to my computer in 30 days, at which time I lose access to Unfair Advantage
Well, one possibility is: you could do what I used to do with Technical Tools data back when Ed Seykota ran the company. Save the data as raw contract files, i.e., one file per contract. For products like NYMEX crude oil, which have a contract every single month, that's 12 contracts per year times 30 years = 360 files, just for crude oil. Do the same for every product you plan to backtest or think you may plan to backtest. It's a lot of files. Fortunately CSI provides an Application Programming Interface ("API") to let you automate a lot of this stuff. Search the TB Roundtable forum, I think a couple of members here have contributed API code here, which you could use for ideas or as a starting template.

Then write your own plain, unadorned, not-fancy, very-few-options continuous contract "backadjuster" software. It reads the raw contract files, splices them together using whatever splice method(s) YOU like, performs the price-adjustment-at-splice-point YOUR way, and emits continuous contracts suitable for backtesting and/or trading.

The good news is, you only need to do Part I before the 30 day window closes. You only need to save the data as individual raw contract files. Then, later, you can create and fiddle with your continuous contract generator at leisure, whenever you like. There's no time limit, there's no deadline.

Posted: Mon Sep 10, 2012 2:36 pm
by sunyata
sluggo wrote:Fortunately CSI provides an Application Programming Interface ("API") to let you automate a lot of this stuff. Search the TB Roundtable forum, I think a couple of members here have contributed API code here, which you could use for ideas or as a starting template.
Thanks for your response. I thought I might have to do this. It suprised me that CSI did not allow continued use of its "backadjuster" after purchasing the data, but whatever. I do what I need to do. :wink:

I was a little confused by the part of your answer quoted above. In UA, I see how I can download individual contracts as you mention without use of an API. Maybe US is the API you are referring to.

I'm still learning what all this stuff is and what all this stuff does.

Thanks!

Posted: Mon Sep 10, 2012 7:24 pm
by sluggo

Posted: Tue Sep 11, 2012 8:42 am
by sunyata
Thanks Sluggo. You're a huge help.

Posted: Wed Sep 12, 2012 2:28 pm
by sunyata
Just wanted to provide an update: was able to download the individual contracts I needed through UA without using a script that accessed its API. When you purchase historical data only, CSI automatically cues UA to start the download process the first time you open UA.

Now to program a "Concatenater."

Posted: Thu Oct 04, 2012 7:33 pm
by sunyata
I've noticed that the default CSI fields for the data in the "Futures" folder that TB utilizes are "OHLCVINU" where VI are "Total Volume" and "Total Open Interest" respectively. My intention is to build my own continuous contracts and the only criterion I know right now to use to determine when to roll is volume: when volume depletes to an unsatisfactory level I roll. If this is the case, then instead I would need to use "vi," that is, "little v" and "little i," which designate "Contract Volume" and "Contract Open Interest" respectively, correct? I am unsure because I don't understand why TB would specify "big V" and "big I" since they would overstate the available liquidity. This seems important because TB compares these values to the Global Parameter setting for minimum volume to determine whether to trade or not based on available liquidity.

I might be missing something. Help would be appreciated. Thanks.

Posted: Thu Oct 04, 2012 7:57 pm
by sluggo
Just to be clear, you want to hire someone to help you set up your CSI Unfair Advantage software? If so then perhaps the Trader's Roundtable forum named

Simulation and Testing Platforms / Data Providers and Other Non-Testing Software

might be a good spot to place your advertisement.

Posted: Thu Oct 04, 2012 8:02 pm
by sunyata
No, I don't want to hire someone. Sorry, I guess I was unclear. I want to understand why TB would specify "VI" for the default CSI data fields when it uses these fields to determine if liquidity exceeds the minimum volume threshold. It seems to me that "vi" would be more appropriate. I will use my understanding of this to build continuous contracts from the correct set of individual contracts that I've downloaded.

Thought maybe someone might be able to shed some light on this.

Posted: Thu Oct 04, 2012 9:38 pm
by LeviF
Whether right or wrong, I will say that I have opted to replace VI with vi.

Posted: Fri Oct 05, 2012 4:04 am
by AFJ Garner
Maintain a separate portfolio in CSI called Test. In situations such as these create different varieties of continuous contract and observe the differences, if any. Create separate monthly contracts and compare them to the continuous contracts.

Posted: Fri Oct 05, 2012 7:21 am
by sunyata
Thanks for the feedback fellas. I've downloaded individual and continuous contracts with VI and vi fields, just in case. My 30 days is up today, so I had to make sure to get it right, otherwise I'd have to shell out another $500 to access the UA database.

Posted: Fri Oct 05, 2012 8:26 am
by sunyata
As expected, the contracts that use "vi" have lower levels of volume and open interest as compared to those that use "VI." Continuous contracts that use "vi" show a radical decrease in v and i around expiration, which to me seems to best mimic real trading experience.

CSI changed SWI to SMI?

Posted: Sat Oct 06, 2012 9:25 pm
by hopewell168
It is for Swiss Market Index futures. I noticed the symbol change when I realized I am missing one update, but there is no clear message. Is there a better way to monitor changes by CSI? Thanks.