Other than that, rolling time-window analysis can come in handy when com-
paring the performance of different strategies over several overlapping periods. It
also helps us keep things in perspective during exceptionally good or bad times.
For example, despite the fact that we currently are in a 36 percent drawdown and
have lost more than 24 percent of our equity over the last 12 months, over the last
five years (60 months) we’re still up 153 percent, which translates to a growth rate
of 20.44 percent per year. This isn’t bad at all by any standard and much better than
the 3.6 percent yearly produced by the same strategy applied to the Dow stocks.
Also, note that this version of the strategy has no losing 36-month period,
which is reassuring to know during the bad times. That is, even during the slowest
and dullest times, the NASDAQ version of the strategy has managed to produce a
new equity high at least every 36 months, which is not the case with the Dow ver-
sion. Naturally, the shorter the necessary time window to produce a profit, the bet-
ter off you are.
To be fair, however, no losing period longer than 36 months isn’t all that good
either. Ideally, there should be no losing periods longer than 24 months. The clos-
est we get to that in this book are with strategies 1 and 2. Table 28.5 shows that
Strategy 2 only loses (at most) 0.54 percent over two years. Table 28.3 shows that
Strategy 1 loses (at most) 15.84 percent over two years, but to compensate, it also
has made at least 10.40 percent a year for any rolling three-year period.
However, when looking at these numbers, we must keep in mind that we’re
looking at historical or back-tested results. To get a better feel for what we can expect
in the future, we must look at the standard deviations and compare them to the aver-
age period return. For example, Table 28.18 tells us that the average 12-month rolling
return is 20.51 percent and the standard deviation is 35.57 percent. Thus, 68 percent
of all 12-month returns should fall somewhere in the interval 15.05 to 56.08 per-
cent (20.51 35.57), which means that only 16 percent of all rolling returns should
come in below –15.05 percent [(1 0.68) / 2]. Because this is the case with the most
recent return of –24.38, we have reason to believe that this is not a typical result for
this strategy and that we therefore should do better in the upcoming periods.
For Strategy 11, we have reason to be even more optimistic about the future.
Because Table 28.16 tells us that the average period return is 10.22 and the stan-
dard deviation is 12.67, we can estimate that 95 percent of all period returns will
fall somewhere in the interval 15.12 to 35.56 percent (10.22 12.67 * 2).
Because the most recent return falls outside this interval, it is even less likely to
be a good estimate of what is to come. In this case, the most recent return belongs
to a group of only 2.5 percent [(1 0.95) / 2] of all returns that are as bad as this
one. However, there are no guarantees. The next period return can still be all over
the place, with a very high likelihood of being negative.
Looking at the rolling time-window analysis helps us get a feel for how reli-
able and stable the strategy or the fund manager is and how likely it is that either
one will continue to produce the same results in the future as it has in the past.CHAPTER 28 Combined Money Market Strategies 373