the plot of BV/share since 1965 shows two, distinct knees, one in 1981 and one in 1999

Yes, I agree. And we both also probably agree that the reason was 1970s Berkshire was structurally very different than 2000s Berkshire, the amount of money under management being a big part.

My reasoning for using trend book value is it will work better and be a more accurate measure until Berkshire changes structurally. Slow changes don't count (except slowly). So when Ted/Todd replaces Mr. Buffett and investment decisions decline from exceptional to merely great, the trend line will continue to be accurate.

Trend book value would lose its reliability if someone decided to break up the conglomerate and sell off the parts. Or change to buying and selling companies, a la Bruce Flatt. Or some other structural change. Until then, the trend line is the better choice because it is more accurate and less noisy than the quarterly stated book value.