To help you understand the program operation, we're going to use RaceCom’s M7 Chaos Matrix™ to develop a simple test to see if AOL, GE, IBM, and Yahoo can predict the closing price of the QQQ 2 days in advance.

We’ve supplied a pre-populated Excel file, QQQ.xls, so that you can run the test yourself. Our objective is not to accept that AOL, GE, IBM, and Yahoo can predict the QQQ 2 days in the future, but, rather, to determine if that small group has predictive qualities. In other words, can the closing price of those stocks actual predict where the QQQ will be in 2 days with reasonable accuracy? Further, if there were some predictive value, what would be the error? How far off can you expect to be, stated differently?

About the M7 Chaos Matrix™:
• Uses the high-performance RaceCom matrix engine
• The M7 Chaos Matrix™ Excel Add-In is capable of analyzing 32,000 rows by 255 columns


Terminology:
Target: A target is what we want the M7 Chaos Matrix™ to learn to predict.
Inputs: These are historical data from which we can draw some inference and, thus, make a prediction as to what value the target will be. To explain, we want to have a formula we can apply to the input data in order to predict what the target will be.

We’ve downloaded data from Http://table.finance.yahoo.com as spreadsheet information. This comes down as a *.csv file. Once you collect all the data you want, you simply import the data into Excel using the data option. Look at the ‘Raw Financial Data’ sheet of the QQQ.xls file for guidance. We copied all the data onto the next worksheet named ‘Selected Data’. Here we manipulated the data as follows:

• We removed the columns we weren’t going to use.
• Added header names to all the columns we were going to use
• We shifted the target of QQQ column down 2 rows (2 days). This corresponds to looking 2 days ahead.

Many people have trouble with the concept of shifting financial data. The reasoning is simple, however. Let’s take July 29th, 2003, for example. We can see the closing prices for AOL, GE, IBM, and Yahoo. We want to be able to take those prices, apply some formula, and predict where the QQQ will close on July 31st, 2003. To accomplish this, we have to shift QQQ down 2 rows so that the close for July 31st, 2003 is on the same row as the input data corresponding to 2 days before July 31st, 2003. This means we’ll use the inputs for July 29th, 2003 and the QQQ close that occurred on July 31st, 2003. The M7 Chaos Matrix™ will associate the July 29th, 2003 data with the close of the QQQ on July 31st, 2003 if we put all the data on the same row.