|
PureBytes Links
Trading Reference Links
|
At 9:31 AM -0600 1/6/02, Ernie Bonugli wrote:
>
>So should the scale be read as follows:
>
>< 2 - start on another idea, discard this one
>2-3 - maybe or "shows potential", if the idea is its primitive form, ie no tweaking.
>4-5 - ok, may be production worthy, but keep an eye I on it,
>5-10 - good, you can rest now!
>
>This scale probably shows my bias and state of affairs.
You seem to be confusing Sharpe Ratio and what I call "tradeability".
The Sharpe Ratio is an absolute measure of the worth of any
investment. It is basically a reward-to-risk ratio. Go to the
MorningStar web site and you will see that there are only about 25
funds with a Sharpe Ratio of 1.5 or higher. Still lots of people
invest in other funds. MorningStar says a Sharpe Ratio of 1.0 is
"pretty good" for a passive investment.
A trader who trades full time needs to factor in the worth of his
time. It would be pretty dumb to spend all your time trading and
still make less than you could get in a passive investment.
Tradeability is altogether different. I have seen lots of systems
that have high Sharpe Ratios that I could never trade. The famous
OddBall is one of them. It has a pretty good Sharpe Ratio but gets
into some pretty bad trades. Sitting in front of a screen watching a
system lose a thousand dollars a minute would drive me batty. Others
may not mind it at all. I suppose if I could have a money manager
trade it and just see monthly statements it would be OK. Mark Brown
said his wife trades it for him (but he was probably joking...)
I also do not like systems that require me to sit in front of a tube
all day doing nothing until the point when you have to do something
very quickly. Someone said it is like sailing - hours of boredom
punctuated by moments of shear terror...
Daily bar systems tend to be easy to trade. You run the code at
night, enter orders for the next day, then the next night see what
happened.
Really short-term scalping systems tend to be easy to trade. You are
in the trade a few minutes on high alert then out. You can break for
lunch or run an errand without worrying.
Daytrading system make it easier to sleep at night. You can never
tell when some nut is going to create havoc somewhere in the world
in the middle of the night, especially now days.
Sitting through many weeks of no-profit is also very hard. You start
wondering if the system has stopped working and start fiddling with
it.
I once designed automatic test systems that took several hours to run
but required an operator to flip a switch or insert a plug once in a
while. We had lots of trouble with the tests until we discovered that
it was a very boring job, requiring a skilled operator. So the
operator would start doing something like adjust something or look at
printouts and try to "improve " the test. We solved this by including
messages to the operator from time to time to flip this switch, turn
that dial, ask him what color some panel light was, etc., and wait
until the operator did it, even though the actions did nothing. What
it did do was to keep the operator occupied so that he didn't screw
things up! After we did that the problems went away...
So tradeability can be as important as an absolute measure of
performance in the real world but don't confuse it with a good or
bad Sharpe Ratio.
Bob Fulks
|