QuantConnect Community

In order to deploy a live algorithm with QuantConnect, you need a prime subscription and, yes, you can run your algorithm using a margin account.
At the moment, for Equity, Options and Futurs trading, we only have support for Interactive Brokers.

Thanks again Alex. When you get a chance could you also answer the questions about costs of trading? I know this is not related to the technical stuff of Lean or QC but still it's important to make sure that Backtests are as accurate and close to reality as possible. For instance, a kind of estimate of charges and fees added to Dashboard would be really helpful. Things like tax, broker's trading fees, monthly fees, etc. At the end of the day, what we really care about is the profit/loss after all the calulations :)

Please take a look at the docs, Reality Modelling section:

These brokerage models set fees, fill models, slippage models and trading markets for a brokerage.

For taxes, you should ask an accountant. Some brokers also educate clients on this issue.

I should add that if your algo trades intensively, its performance may differ a lot from backtests without additional effort to model fills and/or slippage in backtests. Running on an IB paper trade account ASAP should be a goal if you're not doing something with infrequent trading costs (like monthly rebalancing).

Alex, thanks for the link. Somehow I missed that part.

@Petter,

This is exactly what I'm afraid of. I did a live test (QC's demo paper trade) and it's still working fine. Is that reliable? I mean, does QC's demo broker work with a real world spread or switching to real digital trades might give me very different results?

Its as reliable as you make it Patrick; its up to you to define slippage models and fee models for exotic assets. We account for the liquid assets but if you have a special case with wide spreads you should model this with a custom slippage model.

That's a good news! I'm mainly picking stocks in the top 100 price-volume list. I'm not an expert but I suppose this means that the speard should be very low on those? I mean with a simple SetHoldings() I should get a realistic buy/sell price in QC's live mode without customing anything, is that a correct assumption?

There is no magic yes/no answer to a general algorithm. You should do more research into spreads, slippage and then apply as you need for your case. Look at the spread and volumes for each of your symbols - what is the volatility for each? How often are you trading? Even with low spread highly volatile assets can move quickly (i.e. slippage) -- & test in live trading. 

If you're sensitive to this then you should build custom models for each asset you're trading and study the execution to see how well it fills at different times of day. 

When the trade is filled through the QuantConnect paper trading we use your slippage and fill models. When its filled through the brokerage paper trading we use the values the brokerage returns.