Hello, I was wondering if anyone has backtested, or seen results of backtesting, on the breakout theory such as Gary B. Smith's. (Out of trendline/resistance on strong volume). I think it would be interesting to see the results..
This might be somewhat related to what you are looking for. www.trendfollowing.com/whitepaper/Does_trendfollowing _work_on_stocks.pdf
In five years of backtesting I've never found a breakout system for equities that's worth placing my bets. The statistical probabilies are on your side when selling into strength and buying into weakness.
Ive had the same problem trying to test the turtle system, as an example. Apparently you need sophisticated entry position sizing, namely scale in, which not all reasonably priced software has, or is easy to use. Ive heard as much that with random entries its possible to build a trading system if the scale in method is well designed. As for position sizing a fixed amount at the entry without scale in, youll have little or no luck with getting the results of the turtles or other breakout system designers.
Has anyone done or found any research on breakouts as it relates to time frames? For example, is a breakout on a 15 min time frame more likely to succeed or fail than one on a 5 min basis? My true interest is in high frequency time frames so daily, weekly, monthly time frames are not terribly applicable. My group has worked on a few 'breakout' algos on high frequency time frames that observe multiple inputs. Anyone have any research on 'most successful' indicators, price, volume, trade frequency etc. ?
========= Yes, dont trade 15 minutes much but ; 5 minutes = noise & even Don Bright doresnt like a 5 minute chart. If you read all his threads , using something better than 5 minute charts. And the originator elitetrader may want to look past '' breakout theory''; namely a bull market , like several year old=higher highs ,higher lows , higher closes. And volume =clearly secondary to price.
Auto, you can backtest different timeframe and indicators on different timeframes using amibroker. I havent yet tried it in the short timeframes yet, but it looks interesting http://www.amibroker.com/docs/MTFIndicators.html