More on Expectancy
I asked my brilliant wife to help me work out some issues I was having with calculating information for the two screens I wrote about in the previous post. She insisted that I needed a measurement of risk for each trade. I insisted that I didn't. She was right.
So I went back and re-calculated the expectancy for each screen and edited the post with more accurate information. What I learned in trying to figure things out is that the expectancy of each screen will measure the average risk of a trading system. For example, the Zweig RS 5 screen's expectancy is .18. So, if you make 100 trades you should average about 18 times the risk that you took on each trade over those 100 trades. The system produced 244 trades over 2007, so if you risked 1% of your portfolio value on each trade you should expect about a 43.92% return over the year.
I went back and applied a 10% stop loss to all the losing trades to even out some of the larger losses that probably would have been avoided with a stop loss. I know that a few of the gainers would be stopped out as well, but I didn't take that into account. A 10% stop loss shoots the expectancy up to .79. Nice. That brings an expectation of a 193% yearly return risking 1% per trade on 244 trades.
Using expectancy and risk analysis allows you to make better decisions about your system and your stop losses, while allowing you a better way to compare systems. This is the first time since I started trading that I have analyzed things in this way and it has really helped me get a better handle on what I am doing. I think that I can get to the point where my emotions and feelings don't control my trading behavior. I no longer will need to labor over the placement of stops, or obsess over finding the "perfect" system.
Posts
8 comments:
Scott,
Great information, but I have a question after these last two posts. You calculated based on buy on Monday sell on Friday. But, wouldn't you hold over the weekend and re evaluate based on the zweig screen update? I realize holding over the weekend in this market is risky, but I believe originally you said if you were following this mechanically the buying and selling would be done on Monday's after the weekend screening and evaluating.
thanks
I haven't thought of it this way.
I tend to over analyze everything and I think I need to just stick with the 2 screens I like best and buy everything on them. Trying to perfect the screens has cost me a ton of money. I would have done much much better over the last 18 months if I would have just bought what popped up on the screen instead of trying to modify the screen to death. If I can do half what the screens have done over the last 10 years then I will be very happy.
I really don't know why I try to cherry pick from the screen. Looking back before I started doing that proves my returns were better, I was not near as stressed out and I spent only a fraction of the time researching.
Scott, what is your main reasons for not just buying every stock (or nearly every one) on the Zweig screen and selling when it either drops off or you stop out (profit or loss)?
Pretty impressive results...just to clarify though, only the losers were removed and not those that ended up gaining? It might end up making a bit of a difference since the Zweig stocks can bounce around a lot during a single day since they are relatively small cap.
Also, if no more than 1% is risked per trade yet the portfolio is invested in only 5 stocks with 20% in each stock, wouldn't a 10% stop create a possible 2% loss? Just wondering...
Matt,
I based my information based on data that I collected over 2007. It isn't based on my actual trades. Personally, I would hold over the weekend, see if the stock passed the screen again, and adjust my stops if it did pass the screen.
Anonymous,
You make a great point and it's one that I've wrestled with. The reason I've tried to modify the Zweig screen is to tailor it to a smaller portfolio and to see if I could squeeze out a higher gain than the screen as a whole.
But you are totally correct. Just following the screen (even on a monthly basis) produces market trouncing results with very little trading costs, emotional turmoil, and time invested! But where's the fun in that!!?
Gary,
Great question! The cool thing about using expectancy (if I understand it correctly) is you don't have to use your entire portfolio (as I've been doing the last four years). It's based on your risk, so if I had a 100,000 portfolio, I'd buy 5 stocks at 10,0000 each with a 10% stop (1% of my portfolio). According to expectancy calculations, the return would be the same.
If you're going to use a stop you need to understand how it impacts all of your trades in the backtest and not just big losers. You might be surprised by how many trades that make 1-2% or lose the same actually experienced a 10% decline at some point during your hold period. A bunch of trades you're currently ignoring could end up being 10% losers with your stop.
Anonymous 2,
You are very right about stops. I tried to use a short enough time frame (one week) to eliminate some of the volatility. But I am sure that some trades made the type of swings you mention. I watch those stocks pretty closely and I don't remember a lot of big swings like that (but that doesn't mean anything because I usually can't remember what happened yesterday).
I don't have the time or ability to go back and figure out if any of those stocks made those types of swings. If anybody can figure out a way that doesn't take hours I would be interested.
For now, I just like seeing what kind of possibilities are out there. I like to dream.
Scott,
Good analysis, when you say your expentancy is 0.79, I presume are you saying that you make on average 0.79% per trade?
And when you say that the "return would be the same," you're saying that the % returned is the same, not the $ earned, yes? I would think that less money risked is less earned...
Gary,
Yep, that's the ticket. Now I'm trying to figure out that "sweet spot" between risk and reward.
Post a Comment