New Oddball variation - cannot explain results

[Beyond OddBall <beyond_oddball@xxxxxxxxx>]

Can anyone explain why the oddball code below gives
such good results for 1st quarter 2002 ? It appears
that the more slippage you input the less performance
degradation occurs with this new code. If you chart it
though it does not produce the sells I was hoping for,
only the buys (and it labels them buy2's).

For example, if you put "real world" slippage like say
$250, this new code performs in the same league as
OddBall Adv/Decl for the last 2.25 years, but
certainly for 1st quarter 2002 it blows the doors off
anything else. Interesting that it does not appear to
be improved by eliminating morning trades - it's not
time sensitive.

I am a terrible easy language coder, but I had this
theory to test that it is inherently more risky to buy
if price is already ABOVE the EMA20 and sell if price
is already below the EMA20. To test it I decided to
add
a "confirmation" whereby if conditions are met for a
buy, but the buy would be above the EMA20, the signal
would only be issued if price hit a stop placed above
the high of the bar which met buy conditions.

Likewise if conditions were met for a sell but that
sell would be below the EMA20, the signal would only
be issued if price hit a stop placed below the low of
the bar which met sell conditions.

"Normal" buys and sells would occur as always.

*****************************************

{Data1 = S&P 500.
  Data2 = NYSE Advancing Issues}
 
 
 


If time > 830 and time <= 1600 then begin

Inputs: RL(7), BZ(3), SZ(1), BU(1), BD(1),
Xaverage(20);

If Close data1 < XAverage and RateOfChange(Close
Data2, RL) > BZ Then Buy;
If Close data1 > XAverage and RateOfChange(Close
Data2, RL) < SZ Then Sell;
If Close data1 > XAverage and RateOfChange(Close
Data2, RL) > BZ Then Buy at high + BU Stop;
If Close data1 < XAverage and RateOfChange(Close
Data2, RL) < SZ Then Sell at low - BD Stop;

End;