Here are the issues I have with QC, in the order of how important they are to me.

1) The IDE is terribly slow. It takes up 100% of a core of my processor, and quickly gets slower and slower. After 10-15 minutes it slows to a crawl, as if it has a memory leak. Refreshing the IDE makes it go away, and then this process repeats itself. I'm using Firefox, and have the same issue on XP, Windows 8, and my Nexus cellphone.

2) The framework poorly manages memory. I constantly get "Runtime Error: Execution Security Error: Memory Usage Maxed Out - 1024MB max." It happens very randomly, but simply increasing the test period beyond 3-4 months almost always causes it. Having a large number of trades in a small time period also seems to cause it. This makes backtesting painful as I have to break it up into small time chunks. Sometimes I can re-run the exact same code and it doesn't happen the second time. This inconsistency really concerns me for when I move to live trading.

3) Backtesting is unacceptably slow. Before moving to QC, I would write my own backtests using entirely custom software in C. I would run it on my 10 year old dual core XP laptop. An example is a program that tested every single RSI value from 1 to 100, for 20 years of daily data, for 50 stocks, and I'd have the results in under 2 minutes. With QC it can take minutes to do just a few months of data for a single RSI value on a single stock. Combine this with issue #2 and it will take hours to do what should take minutes.

4) Only 4 days of warmup for an algorithm that uses second data is not enough. My algorithms can sometimes goes nuts until it has ~10 days worth of data. Even though it's a profitable algorithm, it won't work on QC for just this reason.

5) The profit graphs use a line to fit the daily closes of a stock. Any serious trader would, at minimum, use daily candles and more likely 15 minute candles. If my algorithm trades on the second time frame, a daily line chart is useless for debugging.

6) There is no available ask/bid data. This is a MUST for anyone who wants to write an HFT, trade the spread, or trade less liquid stocks.

7) Can't trade options. Also, options have huge spreads, so you'd need ask/bid data as a must for backtesting.

8) No matter how good or bad my algorithm is, it always says it's "better than 45% of the community". Possibly a bug?

Thanks. I really think QC has potential, so please keep up the good work.

Author