IB's Data Feed is Useless With ButtonTrader, NinjaTrader, ect...
- Thread starter version77
- Start date
Quote from roberk:
I knew someone would call me on thisIn fact if I was writing on an esig thread I might have reversed the phrase and said esig seems as reliable as IB for data.
I thought it was a fair comment since esig is one of the most popular feeds. I am no real fan of esig, but it has access to many markets and most software can use it for datafeed. I just wish it was as bug free as a data feed as buttontrader is as a frontend.
To me it feels like we are still in pioneer times and a few years from now there will be major improvements in data feeds.
Hi roberk,
you put this very nicely!
Be good,
nononsense
Def, I think that's the key issue here ..."automated". Look at the subject header again. Many of these IB front-end programs have clever features like automatically trailing stops behind the last price, or only executing a stop-loss order if Bid Size is the half the value of Ask size. Things you cannot immediately see with the naked eye. If your auto trailing stop-loss doesn't move because IB's snapshot feed missed a key tick, then you'll be unhappy. The point is that these products are way ahead of the IB snapshot datafeed that is coming into them.
It's not a matter of anything being "wrong" or IB being a bad broker. The issue is that times are changing. People are gravitating to these 3rd party front-end products in greater numbers, because of their user-friendliness and their intelligent and clever order entry/trade management features (by comparison, TWS is very basic). But these products need a real tick-by-tick datafeed to exploit their full potential.
Whether IB can in fact do this, or whether it even wants to try, is another matter. Just recognise that there is a danger that IB gets stuck in the past as technology and consumer desires overtake you. All it might take is for another cheap reliable broker to put out a quality tick-by-tick feed with one of these interfaces and watch people flock.
Right now, the only products with sophisticated and clever order entry abilities with a reliable tick-by-tick feed are the ones costing hundreds of dollars a month like X-Trader. But there is a blossoming market for the smaller retail trader who now has inexpensive access to some of these 'intelligent' order entry capabilities (via Ninja, Button, Futures Trader et al) thanks to IB making it's API freely available. I'm just saying that getting a tick-by-tick datafeed into these products could be a smart business decision for IB. But that's for management to decide I guess
Quote from John Lydon:
Def, I think that's the key issue here ..."automated". Look at the subject header again. Many of these IB front-end programs have clever features like automatically trailing stops behind the last price, or only executing a stop-loss order if Bid Size is the half the value of Ask size. Things you cannot immediately see with the naked eye. If your auto trailing stop-loss doesn't move because IB's snapshot feed missed a key tick, then you'll be unhappy. The point is that these products are way ahead of the IB snapshot datafeed that is coming into them.
It's not a matter of anything being "wrong" or IB being a bad broker. The issue is that times are changing. People are gravitating to these 3rd party front-end products in greater numbers, because of their user-friendliness and their intelligent and clever order entry/trade management features (by comparison, TWS is very basic). But these products need a real tick-by-tick datafeed to exploit their full potential.
Whether IB can in fact do this, or whether it even wants to try, is another matter. Just recognise that there is a danger that IB gets stuck in the past as technology and consumer desires overtake you. All it might take is for another cheap reliable broker to put out a quality tick-by-tick feed with one of these interfaces and watch people flock.
Right now, the only products with sophisticated and clever order entry abilities with a reliable tick-by-tick feed are the ones costing hundreds of dollars a month like X-Trader. But there is a blossoming market for the smaller retail trader who now has inexpensive access to some of these 'intelligent' order entry capabilities (via Ninja, Button, Futures Trader et al) thanks to IB making it's API freely available. I'm just saying that getting a tick-by-tick datafeed into these products could be a smart business decision for IB. But that's for management to decide I guess
Hi Johnny,
I posted earlier that I would love to see IB go tick by tick. However, whether IB, X-trader or Y-trader, they are all faced with the same technological situations: many ticks, many users, lots of bandwith. How to tackle this: install heavier servers, pay for faster communication links -> cost goes up, user number goes down -> you get all your ticks.
Because the crowd goes crazy about cheap front ends - many crazes come and go in the speculation business - a lot of guys are now faulting people like IB for this. They don't seem to grasp much of the underlying reality. Let these go to "better" places. If you know how to make money with this the much stiffer costs will easily pay for themselves! No "cheap crowd pleasers" exist for X-trader though!
Furthermore, some posters pointed out that data leaving the exchanges is not necessarily "tick by tick". See references to CME and EUREX in previous posts!
One day everybody will have tick by tick. I'm sure. Other crazes will reign by that day!
nononsense
>How to tackle this: install heavier servers, pay for faster communication links -> cost goes up, user number goes down -> you get all your ticks.<P>IB could still save bandwidth and transmit the latest data (rather than the one sec snapshot) by changing their protocol.
Solution? Send an update immediately after the bid/ask price level changes. Send an update immediately after a print goes off at a different price.
Volume data should be bucketed. So that a front end using tick bars will have accurate volume by price data.
You still aren't transmitting the expensive tick by tick data, yet you are transmitting changes immediately after they occur (rather than once a second) which is what Button Trader needs.
You could put an upper limit on the number of updates that could be sent per second (like 5 per sec as a max). This would be done to limit the peaks the system would handle.
The bandwidth would have higher peaks but on average might not be that much more than what IB already has.
Solution? Send an update immediately after the bid/ask price level changes. Send an update immediately after a print goes off at a different price.
Volume data should be bucketed. So that a front end using tick bars will have accurate volume by price data.
You still aren't transmitting the expensive tick by tick data, yet you are transmitting changes immediately after they occur (rather than once a second) which is what Button Trader needs.
You could put an upper limit on the number of updates that could be sent per second (like 5 per sec as a max). This would be done to limit the peaks the system would handle.
The bandwidth would have higher peaks but on average might not be that much more than what IB already has.
def
Sponsor
you have some valid points which I didn't consider. perhaps we'll rethink the issue over time as technology improves (it wasn't too long ago a 96K modem was a luxury) and we are confident that our speed doesn't suffer. for now the writers of the 3rd party should use the stops and stop limits of the IB servers when possible and hope that we'll be able to accommodate them over time.
I understand that going tick-by-tick may not be possible right now given bandwidth constraints and agree that the speed of IB snapshot quotes is very good. We can but wish. But it's food for thought. What if TradeStation goes with the model of charging for a tick-by-tick feed and slaps one of these high-brow interfaces into the equation? IB management has gotta keep their eye on the ball I think.
Your last point is off the mark though. Using IB server stops is possitively the last thing any user of these products should do (unless there is no native stop order available on that exchange). Remember that these products can only access the snapshot feed you send out. It is this snapshot feed that they use to determine when a stop should be moved, not your "internal" feed which they have no access to. As such there is no merit to keeping stops on IB servers as pertains to this issue.
Even though sending the tick data is the cleanest and easiest
way in the long run. IB could send the price range, basically
the high and the low since the last snapshot. This would
go a long way toward making it look like a true tick feed.
This would require very little additional bandwidth and
would not bog down in fast markets since it's still a snapshot
feed.
way in the long run. IB could send the price range, basically
the high and the low since the last snapshot. This would
go a long way toward making it look like a true tick feed.
This would require very little additional bandwidth and
would not bog down in fast markets since it's still a snapshot
feed.