Way of the turtle

The MAR Ratio
The MAR ratio is a measure that was devised by Managed
Accounts Reports, LLC, which reports on the performance of
hedge funds. The MAR ratio divides the annual return by the
largest drawdown, using month-end figures. This ratio serves as a
quick and dirty direct measure of risk/reward that I find very useful
for filtering out poorly performing strategies. It is very good for a
rough cut. The Donchian Trend system had a MAR ratio of 1.22
over the period tested above from January 1996 to June 2006, where
the CAGR% was 27.38 percent and the maximum drawdown using
month-end figures was 22.35 percent.
I find the use of month-end figures to be a bit arbitrary and have
discovered that it often understates the true drawdown figures; so,
in my personal testing I use the maximum drawdown from the peak
day to the trough day without regard to where those days fall dur-
ing the month. To give you some idea of how this might differ from
a measure that uses only month-end data, the actual maximum
104
•
Way of the Turtle

---

drawdown, including days other than the end of a month, was 27.58
percent rather than 22.35 percent using only month-end figures.
The resulting MAR ratio is 0.99 rather than the original 1.22 using
only month-end figures.
System Death Risk Revisited
One of the most interesting observations I have made about trad-
ing systems, strategies, and performance is that those strategies
which historically appear to offer extremely good risk/reward ratios
tend to be the ones that are the most heavily imitated by the
broader trading community. Soon you end up with billions of dol-
lars in trades chasing that strategy, and as a result, those strategies
can implode as they outgrow the liquidity of the markets in which
they are traded. They end up suffering from early system death.
Arbitrage strategies are perhaps the best example of this. An arbi-
trage in its purest form is an essentially risk-free trade. You buy
something in one place, sell it at another place, subtract the cost
of transportation or storage and pocket the difference. Most arbi-
trage strategies are not quite that risk-free, but many come close.
The problem is that these strategies make money only when there
is a spread between the prices at different locations or between the
price of one instrument and that of a similar instrument.
The more traders implement a particular strategy, the more the
spread will drop as those traders start to compete for essentially the
same trades. This effect kills off the strategy over time as it becomes
less and less profitable.
Conversely, systems and strategies that do not appeal to the typ-
ical investor tend to have much longer lifetimes. Trend following
By What Measure?
•
105

---

is a good example. Most large investors are uncomfortable with the
large drawdowns and equity fluctuations that are common to trend-
following strategies. For this reason, trend following continues to
work over a long period of time. 
Returns tend to be cyclical however. Every time there is a huge
new influx of capital after a period of relatively sedate returns, there
is generally a period of a few relatively tough years since the mar-
ket cannot easily digest the amount of new money from investors
who are using the same strategies in the same markets. This is gen-
erally followed by a period of good returns as investors withdraw
their money from trend following funds after periods of relatively
poor performance.
Be careful what you look for: If you get too greedy when exam-
ining a strategy, you are going to increase the chances that you will
not get the results you seek. Strategies that seem to be the best in
retrospect are also those most likely to attract new investors and as
a consequence often start performing poorly soon after the new
investments take place.