[Biopython] comparision of alignment scores

[George Devaniranjan]

I have several statistical comparison of alignment scores for a list of
proteins--generated using biopython with the use of BLOSUM and other matrix
generated by me.

All matrix methods (inc BLOSUM) correctly identies it's own sequence in a
collection of seq (high score set apart from the other scores) but I want to
see if my own matrix is better performing than the BLOSUM.
i.e --is my matrix more sensitive than BLOSUM.

Is there a way using statistics to find this out?
I know that this might not be the most appropriate forum to ask this
question but since many of you work in this area  I thought I will try.

Thank you,
George