IB; Biggest Possible Improvement?

[xiaohu]

Quote from kiwi_trader:

Good point. Onto that - I had a drift problem a few months back so I run Dimension 4 and it synchs every 15 minutes. The correction during that period never exceeded 15ms.

I'll run one when the afternoon session starts with 1 minute time corrections off the HK govt timeserver (I normally use a close Telstra timer thats slaved off their internal time server).


Edit: And thanks to def. Would appreciate their input when they get back. IM(H)O it would be more important for IB to improve the data quality by one of the measures suggested here than to, say, make the next couple of improvements to IB charts with "weak" data.

I think this suggestion is a great enhancement to IB feed, without it losing its advantageous.

But since you guys brought up the issue with timing, would it make sense to add in timestamp to the messages if they decide to make this change ?

 

[kiwi_trader]

Re the timestamps. I dont know whether that is needed or not.

My issue isn't that I need the stated time to be more correct. What I want is something we've got already + one other thing. So I think we'd all be better served by:
- real time no lag data (ib gets an A+ here)
- accurate highs and lows (ib gets a B or C here)

but we don't have to have timestamped packets to achieve that ... not saying that Ib shouldn't stamp them if it could be achieved without compromising what we have, just that its not needed for fast+accurate

Here's the picture with my clock updated every 30secs off hk govt ( variation +/- 0.01second so I'd go to faster local updates in preference next time ).

<img src="http://www.elitetrader.com/vb/attachment.php?s=&postid=1144185">

 

[jimrockford]

I have very often observed that IB's real-time data feed misses the high or low for the day, as well as for shorter time intervals. I think this is a much more important issue than time stamps, and also probably technically easier to solve, and that it would be best to avoid confusing the two issues. I think we will have a better chance of getting something if we just focus on highs and lows and forget about time stamps for now.

I would also like to mention that IB's line charts might be useful to people who want to see line charts of VWAPS, but for people like me, who want to see highs and lows, IB's line charts are almost useless. I previously discussed this in the the IB improvements and suggestions thread. The fix we are discussing in this thread, to real-time data generally, could also be applied to IB's line charts, and then they would be useful, like QuoteTracker's line charts.

 

[kowboy]

IB should adjust the historical charts for splits

 

[torel]

Here some more evidence of significiant differencies between real and backfill high/low/close prices for GOOG and SNDK (very liquid nasdaq stocks).

The four columns show for each bar
1. the timestamp of bar
2. diff high
3. diff low
4. diff close

Differences for a bit less liquid stocks are even bigger, making fine scalping hopeless.

 

[nononsense]

Quote from kiwi_trader:

In this thread the way that IB constructs the data for charting programs is described.

The downside of this approach is that when each tick is constructed it only includes the last price within the tick period (as low as 1/5th of a second). This means that a prior higher or lower price might occur but be ignored.

So your highs and lows are suspect for each bar especially if there are a lot of prices printed in a short period. I suspect this problem is biggest with markets like ES and HSI.

Some solve this problem by using another data feed which IMHO weakens IB's franchise with its customers. I would prefer to see IB improve their data feed. Hence the poll: would you like to see IB's data feed improved to include the highs and lows of each period?


My suggestion is that the data include another field. If the last price is the only price then thats all thats needed. If its not then the existing price field should include the high and the other field should include the low. The Sierra Chart, QT etc will still need a modification to include the improved data feed but will not be any worse even if they are not "fixed."

What do others think?

What does IB think?

The IB "missing tick problem" has been discussed AT NAUSEAM in the past. A little search would have given you lots of opinion on this, mostly cloud 7 style.

It is in 99.5+% of the cases a LOSERS PIPEDREAM. May make sense only for 0.5-% of the (oddball) traders.

I want IB to maintain TIMING ACCURACY better than its competitors and to prevent realtime stream DROPOUTS.

As it is clear from past experience that realtime stream dropouts appear to be related to backfill nonsense, I would insist on IB COMPLETELY SEPARATING feed communication channels for realtime datastreams vital to traders, from for backfill and other crowd pleasers.

 

[originalskunk]

Quote from nononsense:

I would insist on IB COMPLETELY SEPARATING feed communication channels for realtime datastreams vital to traders, from for backfill and other crowd pleasers.

No need to insist. Thats the way it works now. There is one socket connection for real time ticks and another socket connection to the history servers.

 

[kiwi_trader]

Thanks for the extra information torel.

Quote from nononsense:

I want IB to maintain TIMING ACCURACY better than its competitors and to prevent realtime stream DROPOUTS.

Ignoring your naturally offensive style and your desire to argue that black is white and nonsense is nononsense its good to see that we agree that:
- IB's datastream should be at least as timely as it is now (any future improvements would also be appreciated IB).
- IB's datastream should be as accurate as they can possibly make it while meeting the first requirement.

Def, Are the people who could comment on the high/low options discussed back from holiday yet please?

 

[kiwi_trader]

A good example occured today with the low 1 hour before close registered in charts as 16860 and the lowest on tws at 16832. This was later confirmed with futuresource.

15 minutes before close 16832 was reached again and the opportunity to trade the resistance wasn't available if you didn't know the prior low had been falsely reported in the datastream.

 

[nononsense]

Quote from kiwi_trader:

Thanks for the extra information torel.


Ignoring your naturally offensive style and your desire to argue that black is white and nonsense is nononsense its good to see that we agree that:
- IB's datastream should be at least as timely as it is now (any future improvements would also be appreciated IB).
- IB's datastream should be as accurate as they can possibly make it while meeting the first requirement.

Def, Are the people who could comment on the high/low options discussed back from holiday yet please?

Thank you for your comment shining with your naturally flattering style and your desire to argue extensively about cases where you obviously didn't see the difference between black and white or nonsense and nononsense.

I love guys of your kind! Keep on posting in between your kiwi snackies.