Bergen notes a "more accurate" point count based on computer analysis. The more accurate count is
A = 4.5
K = 3
Q = 1.5
J = 3/4
10 = 1/4
The argument I have been making is that queens, jacks, and tens are more valuable when they are supported than when they are unsupported. Bergen's revised point evaluation does not solve this problem.
Otherwise, this system in some sense reflects the corrections I make for protected and unprotected honors. In my system, the queen sometimes counts as 2 points but sometimes counts as 1.5, so the average is less than 2. Same for the jack. In my system, the ace and king are valuable not only for themselves but also for providing protection to lower honors. Note that in both my system and Bergen's computer count, AQ is worth 6 points.
But my system does not completely match the computer count. My system is close on kings and probably jacks, but it still underestimates the power of an ace and overestimates the power of a queen.
Of course, my half-point system is a better fit to quick tricks than the normal point count. AQ in the same suit is 6 PHCP and 1 1/2 quick tricks; A in one suit and Q in another is 5.5 PHCP and 1 quick trick. Similarly, KQ in one suit, as opposed to being in two different suits, is more PHCP and more quick tricks.