swhitney has contributed to 21 posts out of 18307 total posts
(0.11%) in 4,458 days (0.00 posts per day).
20 Most recent posts:
Does server time have precedence over an L1 update?
In otherwords, is it possible to print L1 updates whose stamp time is later than the last posted server time update?
CBOE Put/Call Ratio PCRATIO.Z [INDEX]
Can you tell me what CBOE P/C Ratio this is?
Upon visiting the CBOE site and downloading the variou XLS files available there through todays date, I found that none of the CBOE PC Ratios ending values correspond with the value provided by your symbol PCRATIO.Z[INDEX].
Please tell me what this ratio is, and/or what would be involved in getting the IQFeed to supply the actual three PC Ratios provided by and at the CBOE website on this page: http://www.cboe.com/data/PutCallRatio.aspx
In all cases the terminating value does not correspond with the terminal value in the PCRATIO.Z Index from the servers. Further, looking back down the data chains and comparing them, it is clear that the data returned by PCRATIO.Z does not correspond to the published data at CBOE.
The CBOE splits its PC Ratios into to subsets, as well as provides a combined set. The splits are for equity and index options.
As of 12-23-05 the breakdown for terminal values [Daily] is as follows:
Your terminating value for the index as of 12-23-05 is .88.
Please explain what your data set is derived of, and if-when-how we can get the actual data published by CBOE into the IQFeed stream.
Edited by swhitney on Dec 25, 2005 at 04:01 PM
The Barclays Daily USDMXN data needs a bit of tweaking....
Can anyone comment on how long it takes to turn a history request intraday?
If I call the history server, I am getting ET's of up to 8 seconds for 10 days of minute bars on ES(eMini S&P)...
I have been poking around on my code to speed things up but nothing seems to improve the turnaround time...
If I request the last 10 days of minute data (in single minutes)...it takes up to 8 seconds to report back the entire block, at which point I launch the next request in the que...
Somthing must be wrong...
It should not take 8 seconds to get 10 days worth of 1 minute bar data.
Edited by swhitney on Dec 17, 2005 at 07:42 PM
I buffer the incramental callbacks until i see an ENDMSG then kick it to the que cruncher and the entire block is sent as a text chunk...to the listners..
Edited by swhitney on Dec 17, 2005 at 07:44 PM
I do have an account at IB and am very familiar with the API...
Where in the API is this information available?
Contract Details does not give you the exchange information
Being able to exclude non regular session data would also be nice but I know thats asking to much.
I second this request.
And further, we should be able to draw X number of bars from the history as well.
The nature of charting is that it is fractal in nature. It is in reality time indepenedent.
That is to say, if I showed you two charts, one at 15 minutes, and another that was daily or hourly, you couldn't tell what the time frame was unless I told you.
That having been said, as I maintain my data locally, I want to keep X bars of data in storage, not X days.
X days of data would requires some sort of parsing work to figure out how to get and keep X bars of data, taking into consideration overnight sessions, market hours and most importantly interval.
It would be helpful to have both a call by date range, and a set number of bars, and an ideal planet would allow combining both together.
Sorry for not posting back sooner. It is did not reoccurr.
It only happened on the day of the rollover and the symbol was a wild card.
I will double check tonite...but I have not seen it since.
Notice that the missing day is the day before the roll or the Wednesday before the second Thursday...ES,NQ,YM etc
Incidentally I have noticed inconsistent behaviour using wildcard futures symbols and have since abandoned them all together.
I wonder if this had somthing to do with the rollover...
I have checked this all the way back to the Socket Async CallBack...
And the result is the same.
It is as if the day(12/07/05) doesnt exist other than the midnight print...
An XML based service would be a great service to have...
There has to be someone out there doing it...
Or if not there should be...
I would pay for it.
My charting app works perfectley fine during the day and into the the evening when requesting historical data...
But about midnight EST I lose the 2nd to last day of data...
I am using asynch sockets and the data output you see here is the data chunk returned.
I parsed out the open price and the date part for clarity..
Is it me?
interval 15 minute
1255.00, 2005-12-08 00:45:00
1255.00, 2005-12-08 00:30:00
1254.50, 2005-12-08 00:15:00
1255.25, 2005-12-08 00:00:00
1265.25, 2005-12-07 00:00:00
1265.25, 2005-12-06 23:45:00
1265.25, 2005-12-06 23:30:00
1265.25, 2005-12-06 23:15:00
1265.25, 2005-12-06 23:00:00
1265.75, 2005-12-06 22:45:00
1265.75, 2005-12-06 22:30:00
1265.75, 2005-12-06 22:15:00
.......many more bars
I was wondering why the market hours are not in the security fundamental or summary message...but its not unusual.
IB TWSAPI contract details request doesn't include the market hours either.
Why aren't the market hours included in the fundamental message?
How are other programmers dealing with this?
Maintaining data of market times seems the only solution at this point.
Any suggestions? from anyone?
Are you using threads? (you should be)
And if so are they synched? (they have to be)
IQ will hog a single threaded application message que when the data is flying.
Sounds like you may need to set up a blocking to squeeze your outbound message in and through.
Just a thought..
The return fields on lookup HX HT for BidSize and AskSize are always 0.
Any chance at all of getting this data?
If not then why have them in the return string at all?
Edited by swhitney on Nov 22, 2005 at 12:49 PM
Just to be clear...
I don't care if its packaged up in an L2 style block...
I am more than happy to consume and process the data via simple L1 stream.
All I would like is to see the 'outside' posts.
If we know that the tick type==bid || ask && !trade, we can craft the rest of it up on our end.
The game is rigged for the mkt makers, brokers and specialists.
This tidbit of information, no matter how abstract, is very useful for tape readers.
I would also like to see the depth of market and outside bid/ask postings for CME commodities such as ES, NQ etc...
Right now I can import that from the IB TWS API...But Jay has indicated you have the data on your side...
I am by heart a tape reader-programmer...
The dodges, head fakes pull backs and reposts are key bits of data for my style of programing-trading.
PLEASE GIVE US CME MKT DEPTH AND OUTSIDE BID-ASK UPDATES
It would be very very very helpful to be able to tag or key a request message for historical data, and have that key returned in the ENDMSG string.
I buffer the return messages and process them in a different thread than the socket connection.
If the ENDMSG had my key in it, I could broadcast it, and the listening processing unit will goto work on the buffer pack whose key matched the return key transmission. Because the transmission was complete, I can then lock down the buffer pack in the processing thread for saftey.
Just a thought...
PLEASE GIVE US A USER IDENTIFIED RETURN KEY IN ENDMSG
The Summary Message and Update Message Format descriptions have a couple of errors in the Notes column for Items 11(Bid) and 12(Ask).
For both pages Item 11(Bid) Note reads:
"The lowest price a market maker or broker is willing to pay for a security"
It should read :
"The highest price a market maker....(yada yada)
For both pages Item 12(Ask) Note reads:
"The lowest price a market maker or broker is willing to pay for a security"
It should read : "(yada yada)......willing to sell the security
Sounds like your missing a reference object...
Make sure all implicit or explicit external references are present or consistent with your path declarations and pointers...