IB tws Java has been discussed many times here on ET. What I notice is that with more than one IB TWS Java one minute charts open, that the CPU will almost peg out and every thing does not work timely. If I close out all IB TWS charts and only use QT charts with the IB feed, the CPU usage goes down to normal and all data is timely with no lags or delay.
Saico, there are a few things that can contribute to this. 1) Lower the stock table refresh rate to 0.05 on OPTIONS/Edit Preferences menu, MISC tab, bottom right 2) Are you watching 1 symbol or many? 3) do you have a large number shown below the bandwidth indicator in QT? and 4) there are some peculiarities in QT specific to IB. We just made some modifications to QT that should cut the lag to something like 0.2 sec at most. Email QT support and we will get you the version that has the update. Jerry Medved http://www.quotetracker.com/
Over several years, any differences I've noticed between eSignal and IB feeds always have come down in favor of IB. Some people will cry that TWS is built on java (as is ThinkorSwim, which is a fantastic app) but in my experience that is absolutely no problem. On the other hand I've seen some C++ apps, eSignal being one, that will occasionally hiccup.
I ssem to notice a difference between their historical data and quotes collected in realtime. I.e. there is a significant difference between the data when it is refreshed after a time. It's like IB' servers receive different, more ticks.
cmaxb, Are you referring to differences in: - price data - volume data - transaction counts? IM experience you miss the odd price extreme (just doesnt get into the price flow in real time); volume data can be "wrong" when your charting package is trying to construct it from a flow of prices and gets it "wrong" ... which is corrected in a backfill; and transaction data can suffer slight issues like that but backfill, not containing transaction data doesnt "fix" it. So IB data probably isnt as "accurate" as some of the other feeds. But they're not as accurate as CSI data either so they must miss the odd price/vol or transaction too. What it does have going for it is that it never seems to lag. So for real time trading its great. Not 100% accurate but extremely timely Cheers
Thunderdog, Kowboy and Dandxg nailed the slight problems with IB feed. A) Historical sometime does not jibe B) Use the feed only. Dont run IB charts I will add this one. IB feed is not true tick for tick. It is an aggregate. So if you trade tick charts this could be a concern, if you are trying to scalp ticks. Read this for better understanding. http://elitetrader.com/vb/showthread.php?s=&threadid=53426&highlight=ibs+tick+transmission Bottome line, you miss some data but not enough to make a difference. If you dont scalp sub 1 minute time frame ,you should be fine. Backup feed and broker is good idea however.
Hello Jerry, thanks for your advice! Yes, I'm watching only ES and I noticed that the number below the bandwidth indicator is very low, or even completely disappeared. The QT support recommended to download the latest beta version, but I hesitated here, because of possible flaws in that version. Saico
I'm talking about the difference between IB's own historical data, and their ticks coming in in realtime. I collect quotes (open, high, low, close for a time block, say one minute, then construct my own candle), and the candles look different from when you request history to overwrite realtime.
Thank you for your response. What you mentioned about IB's feed not being true tick-for-tick has been mentioned by other posters both here and in related threads. However, this raises a question, and perhaps a naive one on my part. How likely (or unlikely) is it that the High, Low and Close on 1-minute charts will be accurate? I trade exclusively off of 1-minute charts, and an accurate representation of High, Low and Close at this time interval is very important to me. Against this background, would I be expecting too much from IB's data feed?