Measuring Risk
A few of you have asked what I mean when I say that I made "4 times my risk" in my previous post. I'll try and explain:
For each trade I risk a dollar amount. It's different for every trade. So, for example, I buy KEX today (not a recommendation) at 54.84. I see that on average it hasn't lost (or gained) more than $2.24 in a day over the last 2 weeks. I decide that I will sell the stock if it varies more than 2 times that amount in a day (2.24 X 2 = 4.48). I am willing to lose $4.48 per share on KEX.
My portfolio is $100,000 (as an example--I really have billions). I decide to risk 1% of my portfolio on each trade--or $1,000. For KEX that means I can buy 223 shares ($1,000/$4.48).
If KEX goes $4.48 against me, I sell and I lose what I risked or $999.04 or 1 times my risk. If KEX gains $8.96 I make 2 times my risk. I keep track of each trade and record how much I risked.
At the end of the month, I add the winners and losers and come up with a risk number for the entire month.
Here's are my stats for April: I had 14 winners that totaled 9.55 times my risk and 11 losers that totaled 4.96 times my risk.
Calculating risk this way can help you get a feel for your system while allowing you to "plan" your earnings. So, I know that if I can average 4 times my risk each month and I risk $1,000 on each trade, I can earn $4,000 per month with my system (theoretically of course).
Posts
2 comments:
I pulled in the historical prices for GHM, AEHR, ARO, AMED, WATG, ESL, WW, ARG and TDY (most of the Zweig) for 120 days and looked at what the lowest daily pct change was for each. Then I calculated -(2*ATR)/Price, which we'll call the 'risk divisor'. In each case the risk divisor was a more negative number than the most negative 1 day return. For GHM, -(2*ATR)/Price was -9.0% and the lowest 1 day return was -8.2% (on 14-Apr-08) for the past 120 trading days. In general, most Min daily returns were half the risk divisor (eg. ARG risk divisor is -5.3% and min 1d return in last 120d was -2.3%.
I also computed the avg daily return and max daily return for this group and interestingly the numbers were +0.8% and 9.7%, showing a group with positive expectation and positive skew, which is what you want. The average of the max was more than twice the average of the min daily returns (-4.1%).
Finally, I did a quick correlation matrix in Excel and see the stocks are not highly correlated, with the average correlation at 0.26 and only 6 pairs correlated between 0.4 and 0.5. That is great.
My efforts are just some late night spreadsheeting, but it gives me some comfort to spend more time thinking about SP's approach.
Following your example on KEX, I looked at AEHR. If you had $1,000 to risk (based on a $100k portfolio) the number of shares would be about 1,100 ($9,687) but that is 2.4% of the daily volume! Now there must be more than 50 people following Zweig-type screens so when everyone wants out it isn't gonna be pretty.
My view is that holding more than 2% of a stock's average daily volume is dangerous, particularly on Monday mornings. Any thoughts?
Post a Comment