I have gotten proficient at scalping stocks. By scalping, I mean intraday trades that last for a few minutes. Now, I would like to attempt to automate my trading. I have read on ET that some software (such as TS) does not perform well for scalping. Can anyone recommend software for backtesting and automating a scalping method? Norm
You might look into TradeOptimizer. They are only accepting new customers "by invitation only" and they are looking for people who make 40 to 100 trades a day manually so you might qualify. It has many features that fit this style of trading and it comes highly recommended (search on posts by Avalanche)
I've built my own software for automated trading, backtesting, etc. My feelings are it's very tough to backtest a scalping system unless you have the actual tick history - bar data will not provide any results I'd be comfortable with as you never know if you're dealing with the bid or ask, all you know is that an execution took place. I've automated day trading strategies for someone once that tested great fo him on TS, but I could not come close to replicating the results with my own backtesting based on tick-by-tick data, nor did they perform when attached to a live feed and "paper treaded" - so I told him him was not tradable. He found another programmer who told him otherwise, and the 2 of them quickly lost money when they went live. My point here is this - even if a platform will supposedly accomodate your back testing needs, if you're scalping and just taking ticks here and there, be very leery of any test results.
I agree 100%. Consider this: if you started with your software and laid a UI on top of it to set the parameters, plus added few more parameters as to not give away your specific strategy then you would have a "generalized" platform for automating that kind of trading, and it would require no programming. My impression was that TradeOptimizer is that kind of software, I don't know - it may not be oriented around tick-by-tick data which is too bad. I too am running a system on my own software so I don't care, but as I keep adding to it I wonder if I'm writing my own TradeOptimizer, and then I wonder why I didn't just pay the money and use that instead! It comes down to how much is your time worth. I just missed the boat about a year ago when TradeOptimizer was still allowing 30 day free eval's so I'll never know but it's always bugged me.
Do you know the free TRAIN system .. you can test it in realtime. <img src="http://img389.imageshack.us/img389/2576/chf0pz.jpg" border="0" alt=""><br /> http://www.elitetrader.com/vb/showthread.php?s=&threadid=54170
Historical bar data is usually just open/high/low/close over a certain time frame, not bid/ask tick. Maybe TS is no longer like this, but it was 3 -4 years ago. My historic tick data I built myself off a RT feed, by simply logging every tick. I understand that some people do in fact have historical bid/ask type data vs the standard o/h/l/c data, but that's not the norm from what I can see (I could be wrong). Agree or not, fact is that intraday systems that profitably could scalp a quarter or so on ES where easy to find on TS historical data/testing, but when ran against actual bid/ask ticks over the same period they did not perform nearly the same - nor did they when ran in real time. Others experience may not be the same...