Thanks for the list but if you understood what I wrote about how I stress test, you might have deduced that I'd need to use historical data containing up, down and sideways market periods. Thanks for the resource but I have my own database. Also, if I didn't think this was worth testing, I wouldn't bother. Basically your methodology for buying "rockets" is buying stocks that are fundamentally (EPS) and technically strong (RS), and showing recent upside momentum (Stoch). Which sounds logical and is simple. Which is what I like. So just because I don't care for your style, it's not "a priori for me to have what you do fail." If I get decent results I'll be happy to give you the credit and add "rockets" to my collection of entries with an edge.
Thanks for the update on the settings. Not exactly on the exits. My first round of testing will be with time exits. Anything else would confound the results. I won't bother with other exits unless the entries show an edge on their own.
I'll save you some frustration which undoubtedly arises every time someone attempts to backtest Jack's Methods. The Universe of stocks must match the fundamental culling parameters during the time period being tested. Also, very few data providers possess the ability to sort equities by EPS and RS rank. In other words, Stock Symbol XYZ must fall into the top 80 - 90% of RS and EPS rankings, Price, Average Volume, Float and Positive Earnings for the entire period of the test. In addition, these same stocks should also cycle a minimum of five times over six months, and each time, increase by at least 20% over a period of six to eight days. As a result, backtesting these methods proves difficult at best. For example, MIND, a stock currently in my Final universe, would not have met the culling parameters 4 months ago. If one attempted to backtest my current Final Universe during the period of 2000 to 20001, many of the equities (including MIND) would provide invalid results. Due to the difficulties involved with backtesting this strategy, I find 'walk forward analysis' more useful. An easy set up to test using walk forward analysis involved Jack's 'Bruno R' Set-up. Many traders who follow my Journals have profited using the methods contained in the 'Bruno R' Usenet post. I have attached a copy of Jack's 'Bruno R' post here for your convenience. A rudimentary (read not yet completed) walk forward analysis of Jack's 'Bruno R' strategy can be found here. You can find a step by step list of the methods used in Journal One here. If you have additional questions, please do not hesitate to ask. - Spydertrader
When you get that work done you will see the connectiion between the univeres selected by using the IBD percentile arrangements recommended. It is the cremeing of the top percentiles that takes the "stress" function you have out of the consideration by having a universe that is "imune" to stress. It is normal for a person to approach the market from a perspective that is gained over time. Yours is different than mine. I wrote a long post to lilduck on a comparison of each of our first three years. They were very different. it got eaten by the computer when I submitted it. So you are going through taking apart the pieces and I am suggesting to you how the pieces cover the bases you have in your tool kit. Fortunately, you have the Spydertrader formulations as another reference and they parallel what you may come to find out. There are tons of people lurking around doing stuff along side you. Other local people are completing 2006 on a schedule that is similar to conclusions reached by some ET folks who make estimates on compounding based upon either the market's potential to offer capital of the historic unbelievable results obtained by representative usuers. I particularly like the gerry stuff. It is very brief and productive for him except for making money. If you want to google inandlong under the trading forum. He did an evaluation after coming, roughly, from the gerry viewpoint. He just added the NIKE part that gerry doesn't. What he said he found out was that, for him, it worked. The most interesting single thing you will be coming to is what happens to a stock before it does the fast Stochastic trip. It is the sequence of pieces that precede the "buy". This is why people named the process "Catchup, Getting Tomorrow's Paper Today. (43 chapters originally) I backed into it in the late 50's simply because of the way I inked in the master for the brownlines we used for pencilling in daily charts in the 50's. I go tired of plotting the volume in the ten boxes at the bootm of the P, V charts. So I just used the boxes as place holders. The coefficients of the powers of the base ten. This is where DU, FRV and peaking came from. lol..... The charts looked like hills and valleys and they followed the P, V relation of Granville that was the initial basis of making money for me. All I did for years was the simple task of plotting charts, annotating and timing trades according to the P,V relationship. Souping everything up to make tons more money just came about because the money is there and available. Being an engineer, I just extract it according to its availability. This is still "unbelievable" to you and many many others and all the optimization , etc.. is not anyway related to "stress" testing etc... It is more like going to where the money is and knowing when it is going to be available and taking it off the table. It would be like playing poker with the advantage of seeing all the cards dealt face up for yourself only. It is hard to get people to understand to play poker ahead of the hands that are being dealt while you have the cards shown to you as they are dealt. It just greatly compresses the process of taking money out of the market. It would be like trying to break you of the habit of taking things apart to find out that they do not work when they are taken apart. If I am lucky you will chuck the results you are getting and then turn to learning how to make a lot of money and get very rich.....fast. There is a discussion of distribution of profits and how, it turns out for many, they follow a specific distribution. the group is getting the feeling that this distribution is a satisfactory validation of a method of trading. lol..... Here is a stark contrast for you to consider: the market offers a potential to make money every day that is ,roughly, unchanging. Why should a trading method follow a distribution that does anything but follow that distribution. Why do you think a method will change in performance in a changing market? You believe this because you tie any method (pieced into parts) to the market in a way that is an artificial invention of yours. How did O'Neil circumvent all of this unfortunate kind of corrolating? He did it with percentiles; well almost, he only has 99 groups it turns out. I had to use percentiles before percentiles. My artificial percentiles were just a method of creming the market by universe selection. the alternative to peeling lilduck's onion is to break up the making of money into parts and do the parts. You are doing this sort of. Page 13 of the referenced handout has a schematic of the sequences to do the parts. Rolling it up into an onion is possible but that does not make the layers more and more complex. It just makes how you get to see them different. they all are going to be in an operational trading business plan. A perfected trading business plan has in it how making money doe not have to follow a distribution that has much if any variance in the individual profit components. To make more and more money per unit time, you do not go further away from the makret; you, instead get closer to it and get inside it. You get so close you get to see the faces of the cards that ar being dealt to everyone that is playing. There is not way getting a hot list can be automated if your are going to be pulling down 4 to 7% a day with your portfolio. Your "to do" is checking out when each hot list stock has it's fast stochastic go through 50% upwardly diverging. Thats because you were given the place to look (the indicator signal) the defaults to use and most of all, how to get the list to look at. You still do not know anything about how to sort the list and when to sort it and what you get by sorting it. How do you sort 125 stoks down to three groups of 10 and then get a hot list from these and your "recently owned" list? Read Spydertrader. Work hard this weekend. you will be passing hundreds of bystanders this weekend and you will have a great feeling of accomplishment. I have to connect up with a guy who wants to patent (for himself) the laser lighting "concept" of my black pool. He is focussed on getting hydraulic fluids straight for his mirage...lol..he always is. He says he lands better when things are working smoothly.
i would say to this poster that they should try looking at much longer time periods and paper trading them... sounds like that would help you.
There is an extremely great value in knowing and using the longer time periods as a context for making money. The two immediate benefits are: 1. The Columbus benefit of knowing that there is no edge of the earth to fall off of. (As determined by the slower fractal channels). 2. Of being able to discern how much money is being offered in those channels and how fast it mat be collected by timing the market. Spytrader is really giving the best and most succinct advice possible to trader666. It is a very difficult thing to be able to start to consider anything by cleaning the slate first. What is already on people's slates makes it impossible to clean them. I am not saying your recommendation is wrong at all. It is a slower trading method that gets a cedertain range of results. Trader666 knows the Turtles stuff. He knows 60K gets you a million bucks in 25 years. This is a gain that comes from 2500 to 5000 events total for all capital applied. Anyone can plug this into the compound interest formula and see what the average profit is per turn. The market simply offers more if one trades the faster "natural cycle" and takes what the market is offering. Spydertrader put up the results for a first year of operating with the natural cycle. This cannot be compared to the Turtles because the results are so different. 60K times 2 raised to the 25 power for each stream of capital as determined by money management. It is an "unbelievable" result for some reasons. Trading the natural cycle for stocks generates an annual exponent of 40 or so per year per stream of capital. It is 1,000 for 25 years per stream of capital. The cycle profit is a function of the character of the universe used. The highest quality universe which I recommend to minimize risk is absolutely different than the turtle set of tradeables. The turtle tradeable universe doesn't have components that make any money per cycle. This is a very definite handicap for a person who can only do so many trades per stream of capital in a lifetime. The three challenges are: 1. Get a quality universe (It will deliver profits at a rate of 4 to 7% a day) 2. Capitalize the trading to execute on the natural cycle. 3. Optimize the application of capital so that capital is continually moved from the lease profitable part of the portfolio to the stocks of the universe whose potential is increasing and is greater than those being sold off form the portfolio. It takes more than 100 pages to articulate this for a 25 year period (Trading business plans cover the foreseeable future). The two sydertrader jounals (over 600 ET pages long, perhaps) paint the picture in an excellent manner of how to begin and start doing this for the first year or so.