Second send of this problem

[Paul Robinson]

Hi,
My first send of this problem got tacked onto the end of a thread with
exactly the same subject header!

I have two named ranges of the same dimensions (both are 1 column
each) called Test1 and Exam. To fail Test1 is to get less than 65% and
to pass the Exam is to get an A, B+, B, B-, C+ or C.
I want the conditional probability of passing the exam given that you
have failed Test1. So I did

{Sum((Test1<65%)*((Exam = "A")+(Exam = "B+")+(Exam = "B")+(Exam =
"B-")+(Exam = "C+")+(Exam = "C"))/sum(Test1<65%)}

entered as an array formula. This gave me an answer of 1.
However, with the data I was looking at

{Sum((Test1<65%)*((Exam = "A")+(Exam = "B+")+(Exam = "B")+(Exam =
"B-")+(Exam = "C+")+(Exam = "C"))}

was 13 (which was correct)

and

{sum(Test1<65%)}

was 20 (also correct). So my original probability should be 13/20 =
0.65.

If the second and third sums above are in A1 and A2 respectively, then
the formula = A1/A2 also gave 1. However, the formula =
Value(A1)/Value(A2) gave the correct 0.65.

What is the problem with the first formula? Precedence??

regards
Paul