Opinons Please...
I have been modifying the system I am trying to develop and it is starting to look pretty good. Using Richard's Q+ scan and modifying to fit stocks I am interested in, modifying Dave's DNS system (Binary) to something I like, and then using some basic preliminary requirements of my own, the system is starting to take some real shape.
Here is something I noticed and I really need opinions no this subject. As some may know, I have only been trying to learn some TA over the past month or so and don't have some long term knowledge of how this all plays out.
When I did system tests (backtesting) on using, Load All Data & Load Only 1 year in WOW, both of these showed ok results. When I changed the amount of data loaded into WOW to only 9 months, the win ratio is much better. I think more importantly that when the win ratio is good, money management needs to be there also. Of the losing trades, they were out with nominal losses; IE: Acceptable when compared to profits on the winning trades.
As an exampe: using the scan 28 good stocks out of 39 stocks show profit which leaves 11 bad stocks. By this, 28 stocks show profits. Per trade percentage looks good overall. Nothing in these 11 bad stocks show any reason to eliminate them from the scan, so it is an acceptable limit, I think. These are using a backtest of no stop limits, but in actual real trades, I would probably want something, but can't be sure yet without further testing into the future on paper to see what happens either way.
I then have another group in WOW with 54 stocks in the list. Of those 44 were profitable and 10 are not. Again, the losing trades were out with minimal losses.
Now the opinions on this subject, now that I have bored you with some preliminaries. <g>
1) If I only load in 9 months of data to work with, I am heading into trouble? Are there any Pros or Cons to only loading a portion of the data when I have 2 years of data to play with? I came to the conclusion of 9 months since I am looking at some underlying fundamentals of stocks and I was thinking these stocks could be out of favor with bad earnings and loses. Yet if we are trying to catch a stock rebounding and starting to show potential, the past 2 quarters will help identify this. I should also mention, that when I display my binary system onto the screen, it only shows a 3 month graph of the system, but the full 9 months of the data. Could this be because my system needs 3 months of data in the scan to calculate and if it trys to calculate something before some point in the past, there is no data to calculate? That is my assumption.
2) I have found no coorelation betwen separating the Nasdaq stocks with the NYSE stocks when I am dealing with stocks under $5. My original premise was that the bid/ask spread would not do as well in testing that the NYSE stocks would do. Closing price on a NYSE stock is the current real price. Closing price on a Nadaq stock is unidentifyable? Am I wrong on that? I have assumed at this point that the closing price would be the last trade of the day. This could mean that a bid or ask price. I fnot, do I have it wrong and the closing price takes the High price of the day?
That's it for now. I would appreciate any comments on the 2 above questions/concerns I have. I guess what I need to now is although I am loading the data for 9 months, will buy signals still come in daily as they hit. I think so, but since I am new, I don't want to proceed without knowing for sure what it is I am doing. <g>
Tim |
