Alphascanner Blog Update

[Billy]

FYI, I have updated my Alphascanner blog with today's comment from the Algo&Multipivot forum.
(Nothing new for robot subscribers).
http://blog.alphascanner.com/?p=819

Billy

[pete]

hi billy,

thanks very much for your interesting post.

how do you calculate the correlation score?

Is the Accumulation/Distribution % score calculated as 2*%A + %B -%D -2*%E from the MS values excluding ETFs?


pete

[Billy]

hi billy,

thanks very much for your interesting post.

how do you calculate the correlation score?

Is the Accumulation/Distribution % score calculated as 2*%A + %B -%D -2*%E from the MS values excluding ETFs?


pete

For the IBD score, I compute the correlation coefficient with IWM for each % X IBD rating. Then the IBD correlation score is the sum of (correlation coefficient times % X rating ).
Same process with the daily stages where the % X ratings are replaced by the % of stocks in each substage.
Then I compute the correlation coefficient with IWM of the IBD and Stages correlation scores.
The final composite score is weighted by the correlation coefficients times correlation scores.
Billy

[pete]

How many days do you use for calculation of correlation coefficient?
a running number like the last 20 days or all available from start of your data til calculation day?

pete

[Billy]

How many days do you use for calculation of correlation coefficient?
a running number like the last 20 days or all available from start of your data til calculation day?

pete

I use all available past data and update the coefficients at least once a week.
Billy