Before we Begin

       These are the steps we will go through.

       o  Get the correct setup

       o  Find the best indicators

       o  Build 50,000 systems

       o  Pick the top 250 of the 50,000 systems

       o  Look at out-of-sample statistics and use this as a benchmark

               - Optional.  Change one thing and repeat the ENTIRE process above

               - Optional.  To be precise repeat up to 4 times

       o  Pick the top few families in the top 250 systems to trade

       o  Validate the systems to check if out-of-sample results are good.

We need to note a few very important point before we begin the process.

       o  Look at out-of-sample statistics and use this as a benchmark

               - Optional.  Change one thing and repeat the ENTIRE process above

               - Optional.  To be precise repeat up to 4 times

1)  When we try to improve things we must change only one thing at a time.  If you change more than one thing at a time you will have no idea which one accounts for any improvement.

2)  You can run the identical setup multiple times and get different results.  This is because of some of the random seeding that used in building the systems.  If you second identical test is very similar that may be okay but if not you can repeat up to 4 times.  This is not critical to do but it is an optional extra.

3)  At first we are not looking at a system but groups of systems that we know are robust.  Then we will choose the top families and then choose one or two systems to use for live trading.

4)  After that we will validate the systems with a significantly large out-of-sample period.

In More Detail we will look at

o  Peter Zwag's out-of-sample results

o  A good Crude Oil System (this is one of the free systems in the GSB private forum)

o  How to choose the correct indicators using the two pass method

o  Exactly how we setup GSB for Crude Oil

o  Understanding the GSB architecture