Jim, and others who may have thoughts on this,

I believe in the past you've suggested you prefer Price/PeakBook to Price/Book as a metric, to the extent you still track the ratios. I can't recall your entire justification, but I believe it rested in part on the observation that, for BRK, dips in Book tend to be fleeting, and that Price/PeakBook has historically given a more accurate projection of future expected returns and/or the historical growth rate in book. Apologies if I have that wrong.

In any case, my question is how your preference for PeakBook over last-reported-Book squares with your critique re: trendline analysis, in the very specific case of it being applied to BRK (the proverbial tortoise)?

Intuitively I would expect Price to PeakB to have a "lag" in identifying a step down in the average returns for BRK (i.e., off the trendline), whereas Price to Book would pick up the drop more quickly. In other words, I would expect PeakBook to display a historical trendline/past-looking bias.

I think you personally mitigate this difference by applying haircuts to overvalued stocks (e.g., haircutting Apple when it was at 180) in your calculation of Book, and by only using a history that is in tune with the current reality of BRK (e.g., post GEN Re or post 2008), but from where I'm sitting the general question still stands. Do you view using Peak rather than last reported Book as importing in a "trendline" bias, perhaps to be mitigated as you do, or as something different?