[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Scared of "Catscan"



PureBytes Links

Trading Reference Links

Roy Johanson writes
>
> Found a system called catscan (could be number 9 on your
> list) on the net somewhere. I sincerely hope nobody paid
> money for this. 
>

The code Roy posted included seven parameters
   1.  10 = days in avg true range
   2.  0.182 = exponential smoothing factor
   3.  1.25 = *XMA1 entry location
   4.  0.5 = *XMA2 entry location
   5.  40 = days in channel breakout
   6.  $2000 = disaster avoidance stop
   7.  ChopLimits threshold parameter to decide choppy/trendy


One frightening thing is that Roy's code contained a big
section which sets parameter number 7. to a different value
for each different market.  Since there were 39 different
markets listed, it's natural to conclude that the Catscan
code Roy posted has **45** parameters (39+6).  Wow, that's
a big scary number.  It seem like it might be "overfitting".

On the other hand, good old Chuck LeBeau advocates developing
a different system for each different market.  In that case,
Catscan is 39 different systems with an average of 45/39 = 1.15
parameters per system.  I bet Chuck LeBeau loves it.

But I don't.  I am nervous that it is excessively over-fitted.
I'll stick to systems that use the same code and the same
parameter values on all markets.