The 1,3,5,10,15,& 20 yr IRR returns, relative to the S&P, are all negative varying from;

-1.2%/yr...(20 yrs)..... ($785 in Aqua for every $1000 in S&P)
-5.9%/yr...(10 yrs)..... ($554 in Aqua for every $1000 in S&P)


FWIW, my own comment was not about the Aquamarine Fund, but about what metrics are sensible ones. This isn't a sensible one. These metrics are not sufficient to tell you much so should be ignored unless you have much better data to supplement it. The reason is that all those figures end at the same spot, so it's really only one data point.

For example, consider Berkshire's performance relative to market (S&P 500) ending in early 2000:

1 year - underperformance by -50.5%/year
2 years - underperformance by -26.9%/year
3 years - underperformance by -14.3%/year
4 years - underperformance by -17.4%/year
5 years - underperformance by -10.3%/year
6 years - underperformance by -2.7%/year
7 years - tie
10 years - outperformance by only 1.9%/year

The reason wasn't bad performance and consistent lagging for a decade, it was that the end period of ~18 months to Feb 2000 was unusually poor. That poor stretch was anomalous and transient. Poor relative-to-market trailing 1/2/3/5/10 year performance stats by themselves can't distinguish between a recent hiccup and a long slow torture of money loss.

Again, this isn't saying anything about the Aquamarine Fund being useful or not, merely that you can't tell one way or another from that statistic. Money managers love to quote if when they have just had a good stretch specifically BECAUSE it's a terrible metric of their performance. One good short stretch unduly flatters a whole lot of longer intervals ending at the same point.

Jim