EVERYONE's software will be "wrong" if you use IB's datafeed for tick charts simply because IB's datafeed is NOT tick-by-tick, and IB explicitly tells you so. Its not a matter of our algorithm, yours, or someone else's. It is simply impossible to extract EXACT ticks from IB datafeed. You can only do best approximation.
IB data feed
- Thread starter simki
- Start date
EVERYONE's software will be "wrong" if you use IB's datafeed for tick charts simply because IB's datafeed is NOT tick-by-tick, and IB explicitly tells you so. Its not a matter of our algorithm, yours, or someone else's. It is simply impossible to extract EXACT ticks from IB datafeed. You can only do best approximation.
Could anyone recommend a data vendor that does provide real time Tick BY Tick Data? I was thinking of using Ensign with the IB feed. Traders here have recommended using this combo. Would this be a complete waste of time for a trader who uses Time & Sales data for both entry and exits? Any insight and first hand knowledge would be helpful.
Thank you
Thank you
Personally, I use IQFEED with everything else. It is exceedingly good and I am very sensitive to every tick depending on how I trade on a given day. What IB lacks in tick accuracy, they gain in load efficiency. What I mean, as was explained elsewhere, is that IB is snapshot oriented. The weakness is as described here. The advantage is when the market is hyper (ie. really moving along or cascading through stops). A design that takes a 300ms snapshot will not lag whereas a feed that is designed to pump out updates for each tick will lag given the number of updates being pushed out from the server to it's connected clients. With snapshot, IB sees the same load being pushed out regardless of whether the market is stagnate or racing... It is a tradeoff. Because of this, I have both as partners. If things are racing, I switch to monitoring IB data since it always able to keep up without lagging. When things are slow, I am on IQFEED. FOMC announcements tend to be the best time for noticing this disparity...
Quote from Andrew Kirillov:
3. I believe your example and general understanding is not valid, because Total volume will be always changed for each new tick and you always know tick size of a trade = Current Total volume â Previous Total Volume.
Specifically, the above quote appears to the be source of the misunderstanding. If you see:
13:00:00 TotVolume=13350
13:00:01 TotVolume=13360
you have no way to know if this represents 1 trade of 10 contracts, 10 trades of 1 contract each, or something in between. This IS by design. It's not supposed to be a tick feed. myTrack (www.myTrack.com) provides tick-by-tick, AFAIK.
3. I believe your example and general understanding is not valid, because Total volume will be always changed for each new tick and you always know tick size of a trade = Current Total volume â Previous Total Volume.
Specifically, the above quote appears to the be source of the misunderstanding. If you see:
13:00:00 TotVolume=13350
13:00:01 TotVolume=13360
you have no way to know if this represents 1 trade of 10 contracts, 10 trades of 1 contract each, or something in between. This IS by design. It's not supposed to be a tick feed. myTrack (www.myTrack.com) provides tick-by-tick, AFAIK.
Smalltrader,
Everything depends on exactly how you trade. This (not infrequently repeated) discussion just raises the issue of how IB generates "IB Ticks." To optimize internet traffic they don't send a data packet for every true tick but bundle them and send a maximum of 5 packets per second (it could be 2, I dont recall, but within reasonable numbers of displayed items it certainly doesnt vary).
This has a disadvantage: the tick count is not "true"
This has an advantage: on fast movers like HSI IB doesnt get lag in its datafeed like esignal so you get "now data"
I use tick charts but my decisions are based on IB ticks. If I used non-IB ticks my rules would need to allow for the difference. So, IB ticks which give me an activity based chart with a minimum granularity of 200ms work well for me on HSI.
Volume is a different discussion and actually varies from exchange to exchange (in my understanding) depending on whether they report true volume.
But THE REAL TRADERS QUESTION is "can I develop a set of rules around this known data stream that achieves my trading objectives?" IMHO
Thanks for your responce. I use the Tape (Times & Sales) for both my entry and exit. I am in and out in seconds to minutes. I need a new data feed so my real question is how would using the IB Data along with Ensign charting work for my style of trading? I dont want any lag on the tape especially when the market is moving fast and I want to be able to see trade by trade on the tape. If I use IB & Ensign will I have what I need? If not what data package would you recommend? I have used e-Signal and I an currently using Instaquote. I will not return to e-Signal and Instaquote dosen't provide data for the Dax & ER2.
Thank you
Thank you
Quote from smalltrader35:
Kiwi's answer is about as accurate a description you'll most likely find, as it does indeed depend on whether you can adjust your style/method to adjust for the data discrepancies for how you want to trade.
Iâve tracked Ensign side-by-side using, IB, DTN, and eSignal real-time data simultaneously, and the resulting charts are almost always different, both on tick and intra-day. Furthermore, you can mark each of the charts, and the next day the charts are not exactly the same as the day before. Therefore, if your trading style and experience level depends on a tick by tick play, truth is, none of the typical low-end retail data providers are going to meet your needs. But for the average ET type trader, either will work well as long as you know the differences and adjust accordingly.
Another disadvantage, or advantage depending on how you look at it, is that Ensign uses DTN for long-term backfill and all but a few minutes for tick backfill. Therefore, unless you save each and every tick from IB, your long-term indicators are going to be skewed when you backfill. In addition, with Ensign/IB, you have to build a file for each and every combination of instrument/time frame, which I find a PIA.
As for Time and Sales, IB didnât use to provide it, and charting vendors simply interpolated it. With the newer IB upgrades, Iâm not sure how this is currently being calculated.
Ensign is a powerful charting package better suited for custom and/or proprietary set-ups, so if all you require is T&S or standard charting, itâs probably overkill for your needs. It seems many use it for no other reason than they can create âcoolâ psychedelic looking charts, which makes them believe they can trade better. Kids!!!That said, for basic charting needs, QuoteTracker is hard to beat.
Other than high-end data providers, there will be differences between all typical retail data feeds. Even though there are many who successfully (?) use tick charts with IB, it's not really tick by tick data, and shouldnât be treated as such. Therefore, as Kiwi correctly pointed out, you must adjust your method with the data being used.
Unless you are using precise candle patterns or other exact entry methods, IB data will work for what it is intended for, that is, free data. Personally, when a trader gets to the level of putting serious money on the line, paying for a reliable and accurate data feed is a given, if for nothing more than a cross reference.
Ensign offers a free one week (?) trial. In addition, dacharts.com has discussed the IB issue to death.
http://www.dacharts.com/faq/Comparable-tick-charts-by-feed.htm
Hope this helps.
st
Has anybody as of late done a timing accuracy check of IB tick data against CME or CBOT nightly published T&S? Don't bother about missing ticks. How accurate are those you got? (This requires recording ticks together with a precision timing source).
Any other information on data feed accuracy?
Thx
Any other information on data feed accuracy?
Thx
IB's TWS does not necessarily report each tick
FYI, and this comes from the IB Help Desk, TWS does not necessarily report each tick.
It only reports each data CHANGE in bid,ask, and last price/size.
SO, if lastPrice = 1.1177 and lastSize = 1, unless the NEXT price and size <> 1.1177 and 1, THEY WILL NOT BE REPORTED.
In simple terms, if the current price/size = LAST price/size, this is not reported as the data 'never changed'.
This being the case, any cumulative volume calculations based on the TWS API's will not coincide with the volume data.
And from my own experience, this discrepancy can be/is HUGE.
I have a suspicion that this is why one might think that the various charting apps mentioned might not be handling the real time volume so well.
It is isn't the APP.
Its the DATA FEED . . .
FYI, and this comes from the IB Help Desk, TWS does not necessarily report each tick.
It only reports each data CHANGE in bid,ask, and last price/size.
SO, if lastPrice = 1.1177 and lastSize = 1, unless the NEXT price and size <> 1.1177 and 1, THEY WILL NOT BE REPORTED.
In simple terms, if the current price/size = LAST price/size, this is not reported as the data 'never changed'.
This being the case, any cumulative volume calculations based on the TWS API's will not coincide with the volume data.
And from my own experience, this discrepancy can be/is HUGE.
I have a suspicion that this is why one might think that the various charting apps mentioned might not be handling the real time volume so well.
It is isn't the APP.
Its the DATA FEED . . .
You must calculate volume bars from TWS VOLUME events (not LAST_SIZE event). These report cumulative volume. If you do this then volume bars are fairly accurate and fairly closely correspond to backfill data. Jerry Medved has pointed this out several times and he is quite right about it.
Charting vendors that havn't worked this out by now should lift their game.