I know you have had several people try and suggest changes to this over the years. I thoroughly searched the archives before posting. I certainly won't try and tell you want to do or suggest some 'other' algorithm for measuring the entropy of random passwords, however I noticed a result that doesn't make any sense to me.
When choosing a random password of only 0 and 1 that is 512 in length the meter reads 14 Bits. This is obviously low.
Thank-you for considering a fix to this issue.