Running live in Interactive Brokers

Hi Andrés, 

I don't use live trading via QuantConnect myself, but I would like to tell you something about it anyway: I strongly advise you to test everything on your IB Paper account. In particular, you will be able to answer your questions on a paper account yourself and you will also feel better and more secure because you don't have to rely on the testimony of a stranger. 

Just link it to your IB Paper Account and let it run. You can disconnect and reconnect from the Trader Workstation using Ctrl+Alt+F. The question regarding the selected cash balance and position size should appear immediately after the first order is executed. 

I hope someone with experience in live trading via QuantConnect can tell you more. 

Thank you Arthur Asenheimer for responding.

I will go ahead and do what you proposed.

If I may ask, do you do live trading? if so, why don't you use Quantconnect and what do you currently use?

Thank you again

I do live trading but I am currently using the native IB Python API and the ib_insync package. I will shift to QuantConnect after I am familiar enough with the API. 

Sorry if this is unrelated to the original topic, but I am wondering Arthur Asenheimer if you could comment on the pros/cons of using the native IB Python API versus Quantconnect for live trading? I do not see any functionality to private message else I would have asked there.

Hey Dahui

I can share a bit my personal experience with you here. 

Native IB API: 

• Alternative data
• the speed is of course faster as you have only your needed class or method called, especially if you use gateway not workstation,
• but the tradeoff is you will have to retrieve data and debug harder as there is not much processed datasets. If you use fundamentals, you will know what I mean. Depends on your view, the overall time used to develop IB API algo could be just as long.
• One special thing about IB API is that it will restart everyday to clear cache, so you have to schedule every business day to run your algo as it will be disconnected everyday.
• Susceptible to crash when local computing resources are used for other applications and network problem

QC:

• Preset datasets ready to use, like indicators, fundamentals
• Also some alternative data
• Many preset implementation methods like alphas, portoflio construction models, execution models, risk management models, etc.
• Cloud running, much lower risk of local resources insifficient and network problems to crash
• Much friendlier backtest
• When you run into something, you can always ask for support ;)
• Slower speed as running lots of modules, you will have to purchase strong computational power for huge universe/datasets running in very tight timeframe.

I could think of these by now only lol, hope it helps.

Cheers
Louis

I've tried live on IB through QuantConnect (QC). Problem for me was that every time I logged into IB's app it wanted to kick my live algo running on QC. 

QC is still useful for me because I want to program in Python and use QC's fundamental data.

I think you might need to run your algo on a standalone IB account you don't plan to log into while your QC algo is running.

Hi Dahui, 

Louis has already mentioned many good points that I can confirm. 

Just two more points I'd like to add: 

• With QC you can easily switch between backtesting, live trading and research. This is not the case if you build everything yourself using the IB Python API. Of course, live trading via the native IB API would be cheaper (in terms of infrastructur costs), as you can use a small Raspberry Pi as server. But with QC development is easier in my opinion and you also have technical support. 
• If your algorithm needs data for a large universe of +100 securities, then you have to book a booster package. These are then extra costs you can avoid using QC data stream for live trading. 

Hi Ton86,

you can create 2 users to access the 1 account in IB. so can simultaneously login to TWS and also run the algo live.

Cheers,

Pranava