Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Strategies & Market Trends : MDA - Market Direction Analysis -- Ignore unavailable to you. Want to Upgrade?

To: Smooth Drive who wrote (23847)	8/26/1999 11:09:00 PM
From: Matthew L. Jones	Respond to of 99985

Eric, Thank you for your post. You know when I started compiling the statistical data, I allowed the computer to do a lot of the number crunching (6.5 meg Excel spreadsheet with column after column of long formulas). However, when I got the raw data (mean, median, mode, variance, standard deviation, etc.) and began to do a normal distribution, I based the break points in the array on information I took out a book I had recently read which dealt with risk management. At any rate, that is where I got the number for the percentage of a sample group in a standard deviation. Well, I thought I remembered (from my graduate statistics class) a different number, which was still different from your number. When you mentioned it, it made me decide to check it out in detail. Well to put it simply, after spending the whole day working on this problem, I finally figured out what was wrong. The percentage of a sample represented by a standard deviation varies. Incidently, I finally found the help I needed at the UCLA school of statistics website. After reading pages and pages of stuff, I finally found a link which connected directly to their statistics computer via a Java applet. It allowed me to take the mean, standard deviation, and value and gave me back the statistical probability in percentage. The advantage to being able to do that is that I am now able to predefine the increments (specific days or specific percentage levels of gain) and it would then give me back the probability. Then all I had to do was convert the probability to odds. (Probability / (1 - Probability) = Odds :1 This gave me a different set of odds. Some of the numbers were the same as what I had gotten before, however some were entirely different. The best part of all, however is that I am now able to predefine my desired data points and "back into" the probabilities. After spending the entire afternoon on line with UCLA stat dept. and crunching numbers, I put the updated data on my website. I am much happier with this new format. Because of the finer and even number increments I was now able to back into, I have eliminated the useless probability values from my tables. For example, once probabilities exceeded 100:1, I eliminated the remaining useless datapoints from the table. Thank you for calling attention to this oddity. Because of your comment (and the lingering thought from my stat class which agreed with your comment) I was able to refine my statistical data to a point where I am extremely confident in its accuracy. Thanks again for your help. BTW, the site where I posted the updated statistical data is: Table of Extent and Duration of $NDX Intermediate Trends home.earthlink.net Table of Absolute Movement Over Specific Time Periods home.earthlink.net Thanks for taking the time to let me know what you thought. I certainly don't claim to have cornered the market on statistical analysis. That is what is really cool about this thread. Matt