This article is about Monte Carlo Analysis (often abbreviated to
MC or MCA), of course, not about the country of luxury yachts and super
rich people. Monte Carlo Analysis (MCA) has been one of my favorite
tools since the very beginning, perfectly fitting my rather conservative
trading style.
 |
| Simulated Application of Monte Carlo Analysis |
So, what is it about? It is a very simple tool
that can help you manage your risk in a better way, prepare you for
realistic trading scenarios (I wrote about this in another important
article) and give you a realistic estimation about how big your trading
account should be. Remember, the only thing you can really control in
your trading is risk. Therefore, you should be damn sure you do it
really well! I was very fortunate to come across MCA in a very early
stage of my trading career. I have already been using MCA in my past;
when I had been using a quite simple discretional day trading system
(more than 10 years ago). Since then, I've developed very important
habits to use MCA for every new system, my overall portfolio, and for
making important risk-related decisions.
How exactly does MCA work and how do I use it?
The
MCA principle is very simple: Its goal is to take all the trades from
your backtest (backtesting is the very first and highly important step
in a successful trading. I've always been backtesting extensively, even
when trading discretionary and using only Excel spreadsheet, putting all
the trades into it manually) and randomly shuffle the order of the
backtested trades many, many times (even up to 10.000 times) to see
different outcomes of my trading results and trading equities when the
order of the trades is different from backtest results. Why? Simply
because the future order of your trades will very probably be highly
independent of the past order! In other words, get ready for your future
trading equity to look much different than the backtested equity -
because the order of your trades will be different as well! And although
your trades will be similar in its nature, the different order of wins
and losses will produce significantly different characteristics than the
"static" backtest - especially characteristics like Maximum Drawdown
(among others).
Many beginners are surprised when their live trading
equity looks different than the backtested one and they often think that
their system might be broken, but it often looks this way just because
they haven't performed the MCA and aren't ready to face the reality!
So,
let's be more practical here and see some examples. In general, you can
perform an MCA in an Excel Spreadsheet (I'm sure there is MCA in Excel
on the internet), or use some more advanced, commercial tools, like
Market System Analyzer, which I often use.
In its essence, it usually works like this:
You
take all your backtest results, the more, the better. If you have just
your P/L values of all your trades, that is sufficient.
You
randomly change the order of all your backtest trades (you "shuffle"
your results). With the trade order changed, there will be a different
worst case drawdown than the one you have from your backtest results.
You'll note down the value of the new maximum drawdown obtained from the
randomization of the trades order.
You repeat this process many
times. The required minimum is 500 times. However, my heavy testing of
MC concepts shows that you should use many more iterations - preferably
10,000.
Now you'll have 10,000 different variations of the worst
drawdown (each iteration will produce a different worst max drawdown
value). What you need to do, is getting a so called "95% confidence
level".
It simply means that you'll sort all those 10,000
different values of the worst possible drawdown from the smallest to the
biggest. Now you'll find the line number 9.500 and that drawdown value
you find on this line is the one that represents the 95% confidence
level.
And this is "your number". This is the real worst possible
maximum drawdown, which you'll probably experience in the future (in
live trading). This is the number you need to be prepared for and to use
for your capitalization. When this maximum drawdown is exceeded by even
a bigger one in real-time trading, you should seriously start
questioning your strategy - its lifetime might be at its very end.
As
I already said, MCA is an absolutely essential and crucial concept for
my trading. I use it, in many ways (new and creative) every day. It's
the most important part of a workflow for any serious trader. It's the
best (and perhaps the only) way to build a bulletproof defense for your
trading. Don't fight this concept. It's one of the most important
concepts in trading. I've been extremely lucky to use it since almost
the beginning of my trading career.
For more trading related articles, free eBooks or to find out how am I building my own hedge fund, go to http://www.systemsontheroad.com/blog
Article Source:
http://EzineArticles.com/expert/Tomas_Nesnidal/2231820
Tag :
business
0 Komentar untuk "Application of Monte Carlo Analysis"