Search
× Search
Tuesday, November 19, 2024

Archived Discussions

Recent member discussions

The Algorithmic Traders' Association prides itself on providing a forum for the publication and dissemination of its members' white papers, research, reflections, works in progress, and other contributions. Please Note that archive searches and some of our members' publications are reserved for members only, so please log in or sign up to gain the most from our members' contributions.

My backtested trading model achieved a Sharpe ratio of 2.05, a Maximum drawdown of 6% and a CAGR of 30.6%.

photo

 Hock Tong Koh, Quantitative Analysis, Algorithmic Trading, VaR Risk Models, Matlab, Develop Backtesting model, Trading model validation

 Friday, September 19, 2014

Hi everyone, I had recently improve my mid term trading model and manage to improve overall performance to a Sharpe ratio of 2.05, a Maximum drawdown of 6% and a CAGR of 30.6%. I would like to know if your professional point of view, is this a good model. All comments/views are welcome.


Print

5 comments on article "My backtested trading model achieved a Sharpe ratio of 2.05, a Maximum drawdown of 6% and a CAGR of 30.6%."

photo

 Ulrich Horst Benzing, Freelance Consultant

 Tuesday, September 30, 2014



I suggest you do forward testing while taking trading costs, market liquidity and slippage into account...


photo

 Hock Tong Koh, Quantitative Analysis, Algorithmic Trading, VaR Risk Models, Matlab, Develop Backtesting model, Trading model validation

 Tuesday, September 30, 2014



Hi Ulrich, I trade using Market on close order. And my account size is less than 1 mil. With that, will I be affected by the slippage?


photo

 Azouz Gmach, Owner at QuantShare

 Wednesday, October 1, 2014



You shouldn't worry too much about slippage if you trade liquid ETFs.


photo

 Volker Knapp, Consultant bei WealthLab

 Thursday, October 2, 2014



Hock, I am not going through all the posts again, but did you mention you only used the past four years? Was that on EOD data? Which market?

If you only tested on the past four years I would be highly worried regardless of the account size.


photo

 Hock Tong Koh, Quantitative Analysis, Algorithmic Trading, VaR Risk Models, Matlab, Develop Backtesting model, Trading model validation

 Sunday, October 5, 2014



Hi Volker,

Yes, I used only the past 4 years data to optimize the parameters. The reason is that the strategies trade on ETFs which only gain popularity in the past 8 years. With that, I think you can say 2009 to 2014 data is the train set. 2007 to 2009 is the test set. (Out of sample testing) Let me summary the returns here. From July 2007 to Dec 2007 = 2.74%, 2008 return = 36.63%, 2009 return =40.60%, 2010 return = 33.42%, 2011 return = 27.82%, 2012 return = 38.33%, 2013 return 50.11%, 2014 =2.5% as of 25th Sept. ETFs only start to get popular after 2006. As such, I do not have many years of ETFs data to simulate. I had recently generated about 39 years of data using Monte Carlo Simulation. I kept the relationship of every ETFs I used in the model the same. I had so far generated for only one strategy and I found that the randomly rearranging of events reducing the Sharpe Ratio by half and double the Maximum draw down. I see that it is possible that even with 3 parameters, I may still be over fitting the curve. However, even double the Maximum drawdown, and half the sharpe ratio, I still see that it is quite a profitable strategy. I am in the progress of tidying up the details so I do not have the exact CAGR with me. But looking at the chart, it seems good.

Please login or register to post comments.

TRADING FUTURES AND OPTIONS INVOLVES SUBSTANTIAL RISK OF LOSS AND IS NOT SUITABLE FOR ALL INVESTORS
Terms Of UsePrivacy StatementCopyright 2018 Algorithmic Traders Association