Switching Horses

Last post I got stuck on the 20% drawdown and haven't solved the problem yet.  So I have decided to switch horses and put that simulation aside for a while.  Today I have an exciting new simulation to show off.

My plan for the blog is as follows:

• This post I will present the basic performance and statistics
• Next post I will walk through the trading rules and logic behind the rules
• After that I will demonstrate how I test for system robustness.  This will take many posts and a lot of time.
• Then it goes live and I will report my real life results

So I hope readers enjoy and continue to follow this blog!

Trading System Description

This trading system is called "Industry Breakout 19DEC2008" and trades approximately 20 stocks at any given time.  Each stock purchase is 5% of total trading capital with an allowed variance of 30%.   It is based on detecting industry breakouts using an industry Relative Strength Indicator and then choosing stocks based on a Portfolio123 stock ranking system.  The simulation starts with $100K trading capital.  Here is the summary page:

As can be seen here the annualized return is 36% while the drawdown was held to 12.6%.  The realized winners was 62%.  This is probably the lowest drawdown I've ever had with a simulation.

Trading Statistics

The next screen shot provides the trading statistics. 

On this page I like to view the average return (14%), winner / loser return ratio (28.15% / 9.54% or approximately 3:1) and the total number of trades (434).  The 14% return is reasonable - with future degradation it still provides some headroom for commissions/slippage and maintain a healthy profit.  I always like to see a large number of trades, 434 is good.  And the 3:1 profit/loss gives me a nice warm feeling in my stomach.

Stock Trading System Risk

Next screen shot shows the risk measurements since inception and the trailing three years.

From this page I like to look at the trailing 3 year risk measurements.  The system standard deviation is reduced significantly from the S&P 500 index, from 30.89% down to 17.48%.  The Sortino Ratio is a healthy 1.61.  Given the terrible time the stock markets have been having recently, this number is not too bad.  The three year annualized alpha is 30%.  With my research I am really just starting to look at alpha and beta so I don't have a lot of experience with these greeks.  But the idea is to obtain the highest alpha with lowest beta.

Stock Trading System Annualized Performance

The next screen shot provides the annualized performance by calendar year.  I use this page more than any other for evaluating my systems.

As yo can see, the excess return versus the S&P 500 is quite healthy for each year of the backtest.  I can't remember seeing this level of year-by-year consistency before in any of my stock simulation backtests.

Stock Trading System Drawdown

Next is a screen shot of the system equity over time, drawdown and cash invested.

Of note in these graphs is that the drawdown is quite well clipped in spite of the falling S&P 500 index.  What I like about the cash invested graph is the fast ramp up form 0% to 100% invested.  On some of my previous sims, I have seen the trading system take long periods of time to become fully invested i.e. a slow reaction when the markets turn around.  I really want to be there when they turn so this system is good in that respect.

That's all for today.  Next time I will go over the trading rules and simulation setup.

Comments

[gfagerlin]

Steve:

Nice sim! I know the feeling - "This is probably the lowest drawdown I've ever had with a simulation." In September I had a similar experience with a ranking and sim I was developing where the drawdown was -9.8% since 3/1/01(including slippage and commissions) and the annualized returns were very healthy. Alas, it did not like October& November 2008.

Are you building the suspense or did you overlook a link to the sim and ranking?

Glenn

[StockMarketStudent]

Backtests always have to take with a grain of salt since it is 20-20 hindsight (sorry for the cliche). No one can predict what the future holds. Take hedge funds for example. Most of them did well through several bear markets only to be brought down in October 2008.

But I believe that simulations do provide some positive benefit. But the development process is more important than the end result. And a portfolio123 port may not be the entire answer. Hedging may a necessary part of the overall strategy.

I`ll post links to the sim and ranking system next post.

Steve

[sc]

Hi Steve,

Nice looking system. Echoes some work I've been doing on ETF rotation. Why do you go to 0% invested at times?

Shaun

[StockMarketStudent]

Hi Shaun - You are observant to note that the % invested drops to zero at times. This has to do with some market timing that I put in. I will be discussing that aspect more in the next post. And in fact this is one area that I am not satisfied with for this new system. When I design a new system I like to stand back and examine it for awhile. Then tear it to shreds i.e. consider the system with a very critical eye. There is no point to having my head in the clouds when I'm about to risk my hard earned money.

I hope everyone has some patience. My whole family, myself included, came down with a stomach flu at Christmas. I'm just starting to get over it. I also have a second blog called Blog On Smog that I alternate between.

http://www.blogonsmog.com/environmental-friendly/biodiesel-how-did-we-go-so-wrong.html

So I hope to be back at it very soon.

Steve