Right, and my point, which you are repeatedly missing, is that based on that one post, you have no idea whether he uses a news feed or not. You have no information on his system whatsoever other than it was translated from his manual trading and that it was built up from .NET. You assume that he is only using price. He may even be trading based on the weather forecast for that day. Who knows? That certainly wouldn't be TA and since he didn't reveal any detail about his system at all, we can't say whether it's TA or not. We can assume, but we all know what they say about ASSuming!
As far back as in 1993. All custom hardware/software with broker. If you want to use indicators in a statistically significant manner you should be prepared to accept large drawdown. If you optimize, you will have a random system. If you are looking for a trading system via random combinations of indicators you are doing manually you should know that at least one million traders before you tried them already. http://www.elitetrader.com/vb/showthread.php?s=&postid=3370846#post3370846
I've seen you flame other people in the past and, though I generally agreed with you, found it a bit harsh. In this instance, you're making a mighty fool of yourself. Yes, looking at price chart history, or volume, or price along with no indicators is still TA. I WASN'T DOING ANY OF THAT! It doesn't look at past price points to make trades. It uses no historical data. Such systems are clearly outside of your awareness and comprehension, and while you will argue that it's impossible, that surely I "must" be looking at some past data because then what it is relying on, you're, erm. Wrong. I'll leave it at that.
Thanks, Kohanz, for seeing how presumptive someone is being. There's another option for how my system could have been trading that I haven't seen considered. It wasn't weather or data unrelated to the stocks it traded, and it wasn't "price" in the traditional sense. Definitely not anything that could be considered TA, by any definition.
1. I basically used the same process - had a good manual system and wanted more precision in the execution without the need to hover over the console all day 2. I adapted the manual system to TS / EL and produced two variations, each under 8 lines of code - one for SPY and another for individual stocks. Both are proving out nicely in forward testing. Both are based on indicators. PF's range from 1.5 with 55% win rate for the SPY version which goes long and short, to 3.5 going long only, with 45-50% win rate, depending on the price action of the given stock. 3. Go where no one else is. You haven't tried everything.
Actually, it was just level I data. Price, size for each market center. No time and sales, so you can't call it price action/price/TA; besides the trades that were people taking the other side of my robot's traders/orders, if no other trades went off, my bot would still trade (based on the quotes), hence it's certainly not based off of TA.