Dr Smith Predicts NASDAQ High and Lows

Last Update: June 11, 2002 - Next Low Pedicted 06/18/02

DR. SMITH PREDICTS AND TIMES THE NASDAQ for BUY LOW and SELL HIGH by optimized multivariable polynomial type predicting algorithms up to sixth order with variable interactions considered. Upon urging of my students, I have posted this information for their use in our little competitions.

PredictMove ValuePerfectDateBstVal--Diff--%Corr---PPP
04/14/0 buy 3321.29 05/23 3164.55 156.74 95.05 97.67
07/17/0 sell 4274.76 07/17 4274.76 0.000 100.00 98.04
08/03/0 buy 3759.88 08/02 3658.46 101.42 97.23 97.95
08/28/0 sell 4042.68 09/01 4234.33 191.65 95.47 97.88
10/18/0 buy 3171.56 10/12 3074.68 96.88 96.84 97.61
11/01/0 sell 3333.39 11/03 3451.58 118.19 96.58 97.18
11/21/0 buy 2871.45 11/30 2597.93 273.52 89.47 95.87
12/11/0 sell 3015.10 12/11 3015.10 0.000 100.00 96.33
12/21/0 buy 2340.12 12/20 2332.78 007.34 99.68 96.91
12/28/0 sell 2557.76 01/03 2616.69 058.93 97.75 96.63
01/10/1 buy 2524.18 01/08 2395.92 128.26 94.65 96.31
01/30/1 sell 2838.35 01/23 2840.39 002.04 99.93 96.86
03/01/1 buy 2183.37 04/04 1638.80 544.57 66.77 93.10
05/18/1 sell 2198.88 05/22 2313.85 114.97 95.03 92.91
06/22/1 buy 2034.8 06/18/1 1988.6 46.21 97.68 93.94
08/06/1 sell 2034.3 08/02/1 2983.4 49.11 97.64 93.64
09/21/1 buy 1423.1 09/21/01 1423.1 00.0 100.0 93.68
12/10/1 sell 1992.1 12/06/1 2054.3 62.15 96.97 93.58
02/07/2 buy 1782.1 02/02/2 1716.2 65.87 96.16 93.77
03/19/2 sell 1880.9 03/08/2 1929.7 48.8 97.5 93.46
06/18/2 buy . . Verified

Who is Dr. Smith?
Dr. Smith is a University professor who has studied economics and the past fifteen years has evaluated thousands of mathematical models for predicting the turns of the stock market. In the past several years the concentration has been on the NASDAQ data. Doctoral and Masters students are challenged to develop their own predicting algorithms and try to BEAT THE PROFESSOR when using the prediction performance percentage (PPP) as defined below.

Prediction Performance Percentage (PPP)
The prediction performance percentage is a measure of how good the prediction method has been in the past eight moves at making a buy or sell recommendation. The buy or sell close value of the date predicted for the inflection point (high or low) is in retrospect, subtracted from the best actual highest close, or best actual lowest close that occurred, whether it preceded, was on, or came after the prediction date. The buy or sell close value is divided into the difference (made positive) in determining that move (in or out) percentage error. The move performance percentage is one hundred minus the percentage error for that move. The PPP requires to average the move performance percentage for the last eight moves (four buys and four sells). The best lowest close (buy) is found anywhere between its bordering two sell recommendation dates. Conversely, the best highest close (sell) is found anywhere between the low prediction dates (buys) that precede and follow it. Thus, the last recommendation (buy or sell) is not used in the calculation of PPP, because the optimum move is not yet determined.

Dr. Smith's Prediction Method
This prediction method first removes the linear long term (30 year) economic growth line from the NASDAQ composite closing data values. Then the economic cycle oscillation period (found by auto-correlation) is removed from the data (4 to 8 years). This leaves intermediate timing peaks and valleys inherent in the NASDAQ composite data. The optimized predicting algorithm is regressed from this data using a proprietary group of leading indicators as variables for the algorithm. The minimum risk predictor equation form is identified. Risk is defined as the square root of the average square of the difference residuals between the predictor equation and the past 200 or so data values. This equation is differentiated, and the next predicted inflection point (high or low) is identified. If the optimal prediction equation remains the same as new daily closing data values are entered as the inflection point is approached (or more often, the equation changes slightly, but the inflection point date remains the same), and the inflection point is validated the day preceding the day of inflection, this information is published the morning of the inflection day as predicted by the equation set. Once an inflection point (date) is published, IT CANNOT BE CHANGED. Students: I must receive your inflection day by e-mail before or on the day of your predicted inflection before 4:00 PM (market close) or it will be assigned to the next trading day.
e-mail to: DrSmithNASDAQ@hotmail.com

Dr Smith,Dr Smith's students or associates do not (and will not) suggest which stocks to buy or sell. The predicted inflection points are for the NASDAQ composite close values and are provided as information to demonstrate the power of multi-variable regression algorithms in defining general trends. This site is provided for educational purposes only and as a baseline against which students can test their skills. If you buy or sell based on this information, you do so at your own risk.