Hello, I've enjoyed reading Journals and thought I would post one for a while. A very short background, I've been trading for the past 8 years and derive most of my income from it. I got started because after 25 years of working, they decided to eliminate my department (Chemical Engineering). Rather than start out with a new company, I took my "bronze" parachute and my interest in stocks and gave trading a go. Luckily, I did very well, actually too well, at first. But at least I got off to a good start but the early success gave me some bad habits. I've until recently traded in options and IShares. Early this year, I did not like the trend in my results, making money got harder, spreads got wider, and getting fills became very difficult. So, I decided to try an idea that had been kicking around in the back of my head for a couple of years. Simply put, the idea was rather than pick what I wanted to trade and find a method that was profitable, I wanted to see if the opposite would work. I picked a fairly simple entry pattern and added 7 exits and called it a system. I then started screening it against stocks that I selected in various ways. No particular screen seemed to be better than another, so I started to look at all stocks not in the SP500 index, trading above $5/share and trading at least 250k shares per day. A long list and it took months to screen it. The good news was I found many stocks that showed profitability using my method. By May I was trading over 100 stocks and sometimes had positions in 60. This was just too much to keep track of. I whittled the list down to 60 recently and am now out of the developmental phase and into more or less routine trading. So, I'll start by posting this past week's results: NOTE: I will not disclose the stocks I trade as they are the key to the system. Week of 8/18/2003 Monday: 11 wins 8 losses, Lost $580 Tuesday: 10wins 11 losses, Profit $310 Wednesday: 9Wins 4 losses, Profit $1060 Thursday: 11wins 4 losses, Profit $1660 Friday: 14wins 19 losses, Profit $240 Overall 55wins 46 losses, Profit $2690 In addition, I have 32 open positions, 18 shorts 14 longs that at the close on Friday had a net profit of about $1700. 25 were showing a gain. DS
What are you using to screen the stocks, and how do you decide if a stock fits into your criteria? (Based on how your rules worked with it on a historical basis?) And can you elaborate on the 7 exits? (Seems like quite a few...) Thanks for the interesting thread.
You can't talk about stocks, but how about holding periods, anything regarding your trading methodology? The journal will be somewhat dry if you can't.
I use Multex Investor to screen. The screen (I call it my NotSPX screen) criteria are: 1. Not a member of the SP500 2. Price>5 3. Trades over 250000 shares per day. The screen gave 934 candidates. A stock meets criteria if: On the initial test (Tradestation) it shows about 50% wins, a profit factor of at least 2 (total profits/total losses) , a RINA index of at least 25, and an equity curve that looks smooth. If it looks promising, I'll look at it further. I like to see at least 50 trades over my 18 month range of data. There are of course exceptions as I see a couple of recently issued stocks performing very well for me. For a while I was screening IPOs in the aftermarket (starting 6 months after IPO date) and the results were very promising, but with so little IPO activity the last few years the number to pick from is small, but the hit rate was very high compared to the non SPX screen. After the initial screen, I optimize a few variables. I borrow one of Stridsman's ideas and optimize some of the exit parameters using random entries and then export the results to EXCEL to find the best one to use. Exits-actually it is 8 but I usually use 6 or 7. They are: Stop Loss around 5% Maximum Adverse Excursion again around 5% Breakeven Stop-if the stock shows a profit of X% or more during a trade, I enter a breakeven stop. Parabolic Stop-I use this classic Wilder study for exit only with the parameters of around 0.02 AFstep and 0.2 AFMax. I've found this is the most important stop or at least it is the one that triggers most often. A dynamic profit target, but I don't use this one much. It rarely triggers even where I use it. A traditional trailing stop. A Chandelier exit, usually around 6 atrs. A DMI cross exit. DS
Holding periods vary some. I use 15, 30, and 60 minute bars, depending on the stock. Overall, losers exit around 10 bars or less and winners last around 20 bars. I've written about the exits earlier, The entry is fairly simple, consisting of a moving average type cross of a short term average and longer term average. I don't use simple moving averages although in my initial work that is what I used. I have since starting using a DSP filter (see Ehlers book Rocket Science for Traders) for the short average and a triangular volume weighted average for the signal line. For some stocks, I use two triangular VWAs and pre-process the data with a Fisher transform to get rid of excess noise. It seems the first 15-30 minutes of trading has a lot crazy swings that seem to be toned down some via the Fisher transform. The Fisher version works better with the really small stocks. Both of these entries have 2 filters, I consider proprietary in nature. What I've found is there are periods when price action is favorable for my style and times when it is not. These filters help. If I recall correctly, my profit factor went up a lot (from around 2.5 to about 4) and time in the market went down (75% to 45%) after I started using it. The average profit per trade increased a lot too. One quick work re. money management, I take individual postions that are about 3% of my trading capital. This means most of the time I am taking odd lots ( I think 132 shares for example is an odd lot and often take postions under 100 shares with 40 being the fewest I can recall.) DS
Seems complicated but actually it is fairly simple. Some of the math I use may be unfamiliar to you like the Fisher transform and Ehlers DSP , but in essence it enters on a moving average crossover (with a couple of filters) and uses 6-8 exits all of which are usually found in software packages. The hard part is sitting at the computer for weeks screening almost 1000 stocks over 3 different time frames. Today's results: Monday, 8/25/2003 9 wins for $1390 10 losers for $490 Net profit $900.00 DS
When I was using AIQ the Fisher transform was the only indicator that I bothered to program. Very responsive.
Tuesday's results, 8/26/2003 15 winners for $1330 12 losers for -$760 3 breakeven Net $570. Had one big hit in the morning after the "good" news on consumer confidence came out. A lot of churning today as the market swung from a fairly big decline to close up a little. Started the day with about 2/3 of my positions being shorts to finishing 2/3 longs. Took a lot of small losses getting from net short to net long plus the one big loss. Today revealed what I consider one of the methods weaknesses: It lets some big profits get cut when there is a intraday swing like today, I'm trying to come up with some adaptive exits that keep more of the gains when this happens. If you know of one, let me know. DS
Wednesday 8/27: 23 winners for $1045 12 losers for -$135 Net $910 The computer crashed about noon today as I got some variant of the lovesan virus that got through my firewall, updated XP, and updated Norton AV Pro. As Mr. Murphy would have it, my backup computer was with my wife on a roadtrip in Baltimore. I had to borrow a computer to close out all my positions. (With Tradestation, it is around 1c/share if you trade electronically and $20/trade if you do it by phone. For me the choice was, $700 by phone or around $50 by computer. Luckily a neighbor was home and let me use her computer.) I was amazed that I got a virus, but after talking to the guy at my local computer shop, who had to reformat my hard drive and reload Windows, some of the new virii get through even if you have an up to date firewall, antivirus software and operating system. The symptom was everything running on the computer ground to a halt and one of the "SVCHOST.exe" processes was taking all the cpu's. I tried to shutdown but could not, so I had to crash the computer, it would not restart; rather, it kept rebooting but never got past the XP starting screen. The computer guy says I need to add a hardware firewall, which I will do. I got the computer back today (Saturday, 8/30) as all the local computer shops I tried were backed up with computers infected with various virii. At least, I had good backups. DS