Jay, I would definitely consider a move back to a moving average a pullback. This is exactly what I am trying to enter. Perhaps I should have posted a chart to make it clear what I'm trying to do. It doesn't bother me if the strategy is being used on ES or Bonds or Live Cattle or whatever. Nor do I care whether it's being used on Quarterly or Monthly or 15 minute or 5 minute charts. To me, the idea(s) should work in all of these environments. If you set the moving average or trendline to mark the trend you are planning to enter/follow, and come up with a way to time the entry and place an exit stop, what's the difference which chart you use it on? So, I would be happy to hear how you go about using fibonacci, trendlines, and MA's to enter pullbacks. That is, if you want to share the specifics. Banker
I attached a chart where I marked some pullbacks in trend. The markings are a little crude to say the least. Hopefully it will give an idea of what I mean by pullback though. Banker
Actually, the market you trade makes a pretty big difference. Some markets, even on a daily or longer basis, are just too damn choppy to use MA's effectively imo. My use of trendlines, fibs, and MA's is totally subjective and depends on the market at the time. My favorite is using a pullback to a High/Low MA channel. I got this from Larry Williams years ago, and have tweaked it very little. I only use this during smooth market action though (which I eyeball to determine). Well established channels are another favorite. If its not an obvious channel though, I won't use it. If nothing else seems obvious to me, I use fibs. I only use the 50% and 62%. I love these on an intraday basis for ES when a large range has been set early. Uncanny how well the 50% line works around mid day in trending markets. Obviously a self fulfilling thing like trendlines. I also fade obvious support and resistance. This is a reversal strategy however. And finally, I fade nearly every gap of size I see in ES, but not in bonds or euros (euros don't actually gap, they run). Jay
It is interesting how you define pullbacks. Here is another definition was inspired by the more successful systems that have been contributed to date on the wealth-lab site, thanks for the great work everyone form there! The most effective systems share a common timing technique ... they attempt to take advantage of minor extreme price movements away from the norm. We've seen several methods of detecting these extreme price movements: linear regression, extreme RSI, ... Here is a description of the system from the web site and some screenshots below... The Glitch Index system detects deviations from the norm by measuring how far prices move above and below a Detrended Simple Moving Average. The resulting indicator is called the "Glitch Index" and equals the percentage excursion that prices have made above or below the Detrended SMA. The implication is that prices will move back to the norm and we can profit by taking advantage of the temporary deviation. The basic system rule used here is to buy when the Glitch Index goes below -2% and sell when it swings back up above 2%. There's one additional rule that I added that I found helps perfect entry timing. The system doesn't buy if the highest Glitch Index value within a 30 bar period is > 5. This prevents the system from entering when prices snap back from an extreme oversold level. The coloring of the Glitch Index bars helps to illustrate what the system is doing. Green bars indicate GI > 5. We won't buy within the next 30 bars even if we get an entry signal. Dark Green GI indicates > 2 (our sell zone). Red bars indicate areas that are within the "no buy" zone, while dark red bars indicate our buy zone. This system was originally written for 5 minute bar charts, but showed good results on daily charts too. Play around with the De-Trended SMA and with the periods. I think there's more we can gain from this. To see the system at work follow this link and key in any symbol: http://www.wealth-lab.com/cgi-bin/WealthLab.DLL/editsystem?id=2361
Since I like the system and I think the idea deserves more recognition I have tested the system on the 30 stocks that the Dow30 contains. I used daily bars, 5 years of history and 5% of the overall capital for each signal whenever it appeared. My starting capital was $100.000. Attached you find the equity curve and here is part of the Performance Report: Long + Short Starting Capital $100,000.00 Ending Capital $133,420.92 Net Profit $33,420.92 Net Profit % 33.42% Annualized Gain % 5.96% Exposure 38.02% Number of Trades 541 Avg Profit/Loss $61.78 Avg Profit Loss % 1.17% Avg Bars Held 18.11 Winning Trades 354 Winning % 65.43% Gross Profit $122,958.76 Avg Profit $347.34 Avg Profit % 5.99% Avg Bars Held 12.07 Max Consecutive 11 Losing Trades 187 Losing % 34.57% Gross Loss ($89,537.84) Avg Loss ($478.81) Avg Loss % -8.00% Avg Bars Held 29.44 Max Consecutive 7 Max Drawdown ($29,944.26) Max Drawdown % -22.75% Max Drawdown Date 7/23/2002 Wealth-Lab Score 12.12 Profit Factor 1.37 Recovery Factor 1.12 Payoff Ratio 0.75 Sharpe Ratio 0.66 Ulcer Index 5.84 Wealth-Lab Error Term 3.71 Wealth-Lab Reward Ratio 1.61 Luck Coefficient 4.64 Pessimistic Rate of Return 1.25 Equity Drop Ratio 0
Jay, Of course it makes a big difference what market you trade. In order to execute a pullback in trend strategy, there must be readily identifiable trend. What I was driving at, is that if there is a system that works in a trending environment, that system should work in a different trending environment. I doubt if it will do so good in chop. As an example, I used to follow trends on intraday charts. When I moved to daily charts, I didn't have to design a new trend following system. Well, whatever. Good luck to you, Banker
Volker, Like I said in the original post, I'm not interested in swing trading or scalping. I think this Glitch approach would be better suited for a different thread. Good luck to you, Banker
Well, if I know how to define a traend and know when it begins and ends then I don't need anything else. That is why I like to test on many symbols/markets and a long period and I also like parameter stability. BTW, the system that I refer too in the former postings was published in 5/2001.
Volker, Any thorough dictionary should have a definition of trend for you. I think you will need to know more than what a trend is, and when it begins and ends. I don't believe that any system will work across all markets at all times. Good luck to you, Banker
I guess I misinterpreted your definition. In fact I believe it does not matter where the pullback happened or how far it goes? This pullback goes below some kind of MA, but it could easily be changed to just touch the SMA. So why do you see a difference?