|
PureBytes Links
Trading Reference Links
|
Many of the benefits of Monte Carlo analysis occur in the design of a
method. There is some value in monitoring actual trading.
First there are a couple of ways to go about doing a Monte Carlo analysis.
Both start with a list of returns from past trades. Ideally these should
represent trades based on a constant dollar amount risked on each trade
after commissions.
The first way of doing a simulation is to take the group of historic trade
returns and randomly scramble the order while noting drawdown. There is no
value in recording final equity since all the trades are included once and
only once and final equity is the same for any order. This method only
gives insight into the range of variability of drawdown.
The second method is to randomly select returns with replacement. This goes
beyond scrambling the order since each trade may be included zero or more
times and in a random order. The advantage is that this gives an estimate
of both drawdown and final equity variations. This is the preferred method.
The results can be presented in tables, but it is more intuitive the see a
graphical presentation. (Check the links below for examples)
OK. What good is it? There are several good uses.
First a level of risk can be found that meets your drawdown limits. This is
much more realistic than using a formula like optimal-f. (See
http://www.tradelabstrategies.com/optimalRisk/trueOpt.htm )
Next, a number of risk modes can be applied to the returns during the
simulation. You can see what happens with a fixed dollar risk and adjust
the risk to various levels to see what the effect is. Or apply a %Risk
(fixed fractional) method where a percent of equity is risked on each trade.
Or using either of these add an additional fraction of profit to the risked
amount at each trade. (See
http://www.tradelabstrategies.com/mcSim/mcRisk.htm )
Next, you can check the lowest equity during the trials to get an estimate
of the minimum account size needed.
Simulation allows boundary conditions to be applied. The most useful are
starting equity and minimum equity. Minimum equity represents a margin
requirement. (See
http://www.tradelabstrategies.com/mcSim/pathDependance.htm )
Trading methods can be compared to see which is more robust.
Another use is to monitor a method during use. Grab the last 50 trades and
compare the curves with the previous 50 trades. This is quite useful for
daytrading where the feedback comes quickly. I have checked the results of a
daytrading method with weekly samples of returns and found the curves
virtually identical.
Larry Sanders
www.TradeLabStrategies.com
|