What would you choose to develop: IB's api or Ninjatrader8 ?

P

pstrusi

I have a decent experience working with NT7 connected to IB, pretty fine. Now platforms continue to evolve and I'd like to know technical experience of people in the deep know which platform to develop an Algo from scratch.

Thanks
 
IB API all the way. I don't want to be dependent on someone else's software over which I have no control.
I've heard enough bitching along the lines: "hey xxx doesn't work any more ever since I took that update"

Yes, it's more work but I've built up my own abstraction layer on top of IB's API that I understand and control.
Also I'm a crack programmer :D
 
stevegee58 said:
IB API all the way. I don't want to be dependent on someone else's software over which I have no control.
I've heard enough bitching along the lines: "hey xxx doesn't work any more ever since I took that update"

Yes, it's more work but I've built up my own abstraction layer on top of IB's API that I understand and control.
Also I'm a crack programmer :D
More...

I appreciate your opinion, it's pretty logical what you say. However I'm not a professional programmer, so given my time table needs, more hard technical work means to me being behind schedule, that's why I consider nice platforms like NT. Just for curiosity: do you code professionally ?
 
pstrusi said:
I appreciate your opinion, it's pretty logical what you say. However I'm not a professional programmer, so given my time table needs, more hard technical work means to me being behind schedule, that's why I consider nice platforms like NT. Just for curiosity: do you code professionally ?
More...
I don't want to come across as bashing NT. It seems like a good product and in your case it may very well be the way to go.

I'm a professional programmer and have been for 30+ years. That's why my first thought is always to "do it yourself".

Regarding latency requirements: Unless you can find hard evidence that NT actually adds measurable delay to market data and order entry it's probably not worth chasing performance. You'd have to expend maybe 10X the amount of coding work for maybe a 10% improvement. The TWS API itself is not a low latency interface. You'll get lowest latency using FIX and a co-located server.
 
So, obviously, I must ask: do you program as freelancer some times? Perhaps is a scenario that I should consider.

Regardless, thank you for your opinions
 
stevegee58 said:
IB API all the way. I don't want to be dependent on someone else's software over which I have no control.
I've heard enough bitching along the lines: "hey xxx doesn't work any more ever since I took that update"

Yes, it's more work but I've built up my own abstraction layer on top of IB's API that I understand and control.
Also I'm a crack programmer :D
More...
I've got news for you, you're dependent on someone else over which you have no control when you build to IB's API. After you trade with them for a while and realize how truly awful they are in a number of areas, you still stick with them for months longer because you locked yourself with the sunk cost of your time developing to their API and you don't want to loose all that code. Which doesn't work with anyone else. Believe me, I've lived that nightmare. If you must use IB, use the FIX solution and at least you'll be able to port it to any other professional level broker with minimal refactoring. And optimally get your quotes from somewhere else, FIX isn't very amenable to that piece.
 
Last edited:
Very interesting opinion, so that left me wondering: how much harder is it to program in Fix vs Api ?
I imagine there're freelancers for this too.
 
You must log in or register to reply here.
Back
Top