I think the ability to enter the market gradually over a period of days while it is correcting was really the missing link in my strategy. Monitoring of position information is critical to the portfolioâs performance. It really is a âmarket timerâ if one could ever exist. I understand that it is counterintuitive that being long should work in a falling market. It was not until I began back testing that I observed this phenomenon on a consistent basis. I think it has to do with the fact that the market âwantsâ to go up more than down. Otherwise, the market would eventually get to zero (sort of like the US dollar). I will just have to wait until the next bear market so see if it works for real. I actually have a couple short only strats in the works. However, the short rtm strat seems much more dependent on a down trending market. In back testing, it underperforms in a bull market. I started using it today, but if it underperforms, I will probably shelf it until the broader market breaks below its 200 day MA.
Hi Rol Does Amibroker execute your trades? I haven't found Amibroker/IB to be unstable, though I've only been trading automated with this combination for 6 months. Also, I only use Ami to send orders through to IB at the open. And that's where Ami/IB combination stops. All my buying is at the open, and then when a buy order is filled, my exit order becomes active. So I guess I don't really need Ami/IB to be that reliable as I'm not constantly sending orders through. In back testing I noticed when I became too choosy about my entries, it resulted in limited market exposure. There's nothing wrong with limiting market exposure. Its one of the ways to limit risk/drawdown. Although, if I go with my short strat in addition to long strat, I may find myself in the market all the time! My strategy also works on the short side but its just not good enough. So I eliminated it. So I trade long-only at the moment.
Total Net Profit $10,745.26 (Per Share) $0.15 Gross Profit $23,582.04 Gross Loss ($12,836.78) Profit Factor 1.84 Total Number of Trades 648 Percent Profitable 66.36% Winning Trades 430 Losing Trades 218 Avg. Trade Net Profit $16.58 Avg. Winning Trade $54.84 Avg. Losing Trade ($58.88) Ratio Avg. Win:Avg. Loss 0.93 Expectancy 0.28 Largest Winning Trade $429.64 Largest Losing Trade ($478.54) Max. Consecutive Winning Trades 15 Max. Consecutive Losing Trades 9 Total Shares/Contracts Held 69974 Total Commission $1,626.01 Return on Initial Capital 22.45% Annual Rate of Return 74.55% Buy & Hold Return -1.48% Return Retracement Ratio 3.57 Trading Period 3 Mths, 7 Dys, 5 Hrs, 50 Mins Max. Equity Run-up(Daily) $11,099.90 Date of Max. Equity Run-up 4/8/2011 15:00 Max. Drawdown(Daily) Value ($3,831.77) Date 2/23/2011 15:00 as % of Initial Capital 8.01% Max. Trade Drawdown ($475.02) I had to make some adjustments to result reporting. When considering open positions as closed for generating the equity curve, Tradestation miscalculates when a company changes its ticker. I sold WGOV back in January but the ticker recently changed to WWD so now TS calculates a nearly 100% loss on the position. I don't know how to correct it, other than to not consider open positions as closed. Maybe others TS users have noticed this problem. I recall this problem in the past, which is a reason I changed the view. I will still show dates as opposed to number of positions along the x-axis. My average hold on losers is 4 days, so after about a week passes the curve will reflect losses. What I will do is simply state my net open position P/L when posting equity curves. Currently it sits at (-$553). I also removed the few discretionary trades I have made ytd from reporting, to make the results more reflective of the true automated strategy performance.
I am going to be posting comparisons of my long, short, and long/short rtm strategies by year starting in 2005. I will also include the SP500 during each year. For the most part, the long strategy is what I am currently using for this journal. The short strategy is simply reversing the strategy rules of the long only rtm strategy to attempt to capture the opposite extremes of a move in stock price. Although I have experimented with a short rtm strategy in the past, this is my first in depth look and comparison with a long rtm strategy. This will be a learning experience for me, and I hope for others as well. While I realize there are probably many different mean reverting strategies with varying results, I am only making observations on the results of my strategies. I did not back test more than a year at a time because Amibroker does not increase maximum number of positions held as equity increases, but rather just increases each position size. Therefore, at some point liquidity becomes an issue and results become unrealistic. However, I would be increasing maximum number of positions held as equity increases. What I am observing is that the long only strategy is much more consistent than the short only strategy, regardless of overall market direction. The short only strategy depends much more on a down market to perform well, for example in March and April 2005. Then from May to July, the short only strategy did poorly. The long only strategy does not show this dependence on longer-term market direction, which is why I feel it is a survivable strategy, unlike my short only strategy. The long/short equity curve appears relatively smooth, but when compared to the long only strategy, one can see that it takes away profits. The conclusion I draw from these comparisons is that my rtm short strategy does not do as well as my rtm long strategy over the long term. I believe this is the case because the long only strategy is identifying a market dynamic that exists on the long side, but less so on the short side.
Here is a look at 2006. All three strategies had a maximum drawdown of about 12%. However, the short only strategy DD was more concentrated towards the beginning and the end of the year. It performed well between May to September, when the market turned down and went below its 200 dma. The annual return is well below the long only strategy return. The long strategy was more consistent and from May to September, rather than turning down, went sideways; then rose with the market for the remainder of the year, while the short strategy gave back gains. This is what I have observed with the long only strategy. It performs well in sideways and up markets, and avoids portfolio breaking drawdowns in declining markets.
For 2007, the long strategy continued its steady climb. The short strategy non-performed, except in mid July and October, when the market sold off. The long strategy had a 10% DD but was able to recover within a couple months. This is what I have noticed with it. Whenever it suffers a sizable DD, it recovers losses within a reasonable amount of time. This makes it a survivable strategy, in my opinion. The long only strategy appears unaffected by the market decline from October thru the end of 2007. This is likely because the average hold on a position is 2-3 days, thus the longer-term market trends become less relevant.
In 2008, the long strategy had a 30% DD in mid-October, during the strong market decline. I would consider this a "tail" for my strategy. It showed high volatily from September thru November, but managed to make back losses, and then finish the year strong. Even though the trend was down for the year, the short strategy did not capitalize until the strong decline occurred from September thru the end of the year. Again, the long/short strategy underperformed the long only strategy.
The long strategy had a sizable 23% DD into the March 2009 market lows. It then rallied off that and finished the year strong. The short strategy started out strong for the year, but then displayed little edge in a rising market.
For the final year, 2010, the long only strategy performed as expected. In the market downturn from May thru June, it trended sideways. The short strategy did well during this time and poorly the rest of the year. Until I come up with a better short strategy, I think I will remain using my long only strategy. Perhaps a trend following strategy, incorporating macroeconomics, would be a better short strategy than mean reversion.