what criteria do you RTM guys use to determine when to cut your losses? When do you say "oh crap this is a legit trend and not just noise that will revert"?
I keep it simple, I buy weakness and sell strength. Sooner or later a weak stock shows some strength or a strong stock shows some weakness. I control risk by keeping position size small and trading multiple strategies.
I guess that is the idea. I am running this moment a test on SPY data using Price Action Lab and I will post the results if I find anything interesting.
dwpeters - most of the time this strategy works (more so for sector/industry and market etf's than individual stocks) however you MUST use a stop loss. Larry Connors says to buy weakness/sell strength w/ no stop which is suicide. I know this b/c I've done it before saying "the market is sooooo overbought/oversold now - it HAS to reverse". You can guess what happened. I suggest you place a stop 2 ATR (or whatever method you want) below low/above high if you're long/short to protect against trends that rip your face off. 1a2b3cppp made the excellent point - at some point you just have to admit you're wrong.
I appreciate the well intentioned advice but stops are not the only way to control risk. I've taken more than 1000 RTM trades, including long trades through the fall of 2008 and early 2009, and many thousands more backtested. I am not blindly following Connors or anyone else's advice. Risk control is critical but I am quite comfortable with my risk control strategy. And I trade based on probabilities, not whether I think I am right or wrong.
I finally found some time to run a few tests. The results are remarkable, at least it appears that way. I used a daily sample for SPY from 01/1997 to 02/17/2011. I divided the sample into 3 sub-samples: - search sample : 01/2000 - 12/31/2009 - validation sample: 01/1997 - 12/31/1999 - testing sample: 01/2010 - 02/25/2011 I had the software search for patterns with win rate > 80%, profit target = stop-loss = 2%. Then I run the results on the validation sample and I selected all the patterns with more than 8 trades that failed (pf < 1). This is contrary to common practice of selecting systems to perform well. Then, I used those patterns to develop of system. I changed long patterns to short and short to long. The system was backtested in the testing sample. Result: trading profit of about 17 points in SPY with 20 trades of average length 5 bars. Buy and hold gain was 22 points. The PAL results in the validation sample, i.e. the selected patterns, are shonw in the next post.
These are the PAL backtest results in the testing sample: Only 3 patterns survive in the test sample out of the 10 (pf > 1): I concluded that this is one of the most dynamic methods I have ever come accross and I thank goodgoing for his contribution. It evades common sense but it seems it works. I am going to concentrate on developing system this way, actually fading the signals most technical traders use.
This is the way to develop trading systems that go against the herd. Good work Bill. I'm glad you liked my approach.