Re: version of Golub Chan LeVeque stable algorithm for sample variance applicable to weighted observations

In article <1152316445.668105.289070@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>,
koopman@xxxxxx says...
The unweighted code is wrong. Reverse the order of the updates,
change oldVar to var, and change (i-2) to (i-1) in the var update:
Sorry about the errors; thanks for the corrections. I wrote hurriedly
away from the office, so I was going from memory (I had a variable
oldVar in the code that was cluttering up my head, for reasons that
don't matter here...).

Anybody know the update for weighted observations?

Regards,
Will Grove
.