Tuesday, May 13, 2008

Looking for Mr. Goodbar - Part 2

Today we look at the some of implications of bar size, again using Dr. Brett's watchlist.Attached are the backtest results of the same simple RSI system (with no stops) for 5, 10, 15 and 30 minute bars. Since I'm really only interested in finding the best technical performers, I'll cull results of the top five stocks in each time frame to see if any pattern emerge.
All tests range over 500 bars, so the 5 minute bars extend 6.4 days, while the 30 minute bars extend 38.4 days.
Top 5 stocks cumulative trade profit results are as follows:
5 minute: $2095
10 minute: $3417
15 minute: $4516
30 minute: $7473
Interestingly, in each of the four time frames 4 out of 5 stocks are the same: SLB,OXY,COP and CVX (all energy stocks). Also of note: the last place performer in each time frame was AIG.
The results of this little (and I emphasize little) test sample support several concepts I utilize in my daytrading.
1. 5 minute bars provide the biggest risk reward payoff for short term traders. Had I added trailing stops and/or break even stops, results would be even more impressive, but time is limited here in blogdom and I am deferring discussion of stops for several weeks until I get rolling with the new blog format.
2. Certain sectors display technical analysis correlations better than others. When I first started using TradeStation about 10 years ago, INTC was the poster child for technical analysis correlation. Simple systems based on RSI, stochastics, MACDs, linear regressions, Darvas Box- it didn't matter- they all made money with INTC. The situation has changed today with the dominance of program trading, order flow management (masking), backchannel trading (Pipeline) and any number of other factors that make the old workhorse algorithms ineffective. (also see The Game)
3. Finding a trading edge to improve your ROI is not always where you expect to find it. It should could as no surprise that some of the most successful prop trading firms seek out new traders with backgrounds in mathematical modelling, game theory and forensic statistics. Understanding probability analysis is crucial to a trader's success, and whether done on the fly or with the aid of sophisticated real time algorithms, every successful retail trader I know is very adept in pattern recognition and statistics.
Tomorrow, we'll further explore the technical alignment possibilities of various sectors as reflected in ETF performance. Keep in mind that this current series is a work in progress and, just as in trading, it's best not to have too many expectations about what's going to happen next.

No comments: