[R-SIG-Finance] [R-sig-finance] Any suitable backtest functions?

R@Nabble vlanschot at yahoo.com
Wed Jan 21 18:08:44 CET 2009

You may want to have a look at Amibroker (www.amibroker.com). It is cheap but
much more powerful and versatile than Wealth Lab, Trade Station, etc. More
importantly, it has an R plug-in, freely available for registred AB-users,
which allows you to backtest any of your R-models within AB.


After learning from replies of the list, I think maybe I shoud not use
'backtest' here.
I am new to R so my future job would consist of  two phase.

Phase One:  Analysis of Signals.
Step 1.  Signal generating:   All trading systems should have signals for
trade.  Even the trading system of the great Simons, I think.    Signal
generating would be the first step, using technical indicators, fundamental
changes, news, statistical  indicators or whatever you like.   The
flexibility of R makes it the best candidate for signal generating in
prototyping process.  But it would be hard to define some univesal functions
for signal generating.  
Step 2.  Signal analysis:   If we generate trading signals according to some
specific ideas, then, the 2nd step would be analyzing the statistical
pattern of such signals and the following yeild/risk pattern.   The
frequence of the signal, the distribution of the signals.  If the signal is
some threshold of an indicator, then the distribution of the indicator. and
so on.  And the maximum yeild/risk during the following days/months.   There
would be different yeild/risk distribution pattern in different time frames.  
The most useful output would be histgram and charts and so on since I don't
think such distributions would obey specific classical distributions in
textbook.    I wonder whether there are some packages help me in the step 2.   
Indeed, there would be difference between studies on single instruments such
as futures of euro or  index and studies on stocks in a stock market.    The
latter should consider the sorting of thousands of stocks. 

If there are some signal robust and significant in alpha generating, then
comes the phase two.
Phase Two: Backtest/Simulation.
There maybe some packages focusing in this phase.  Maybe not.  Backtest for
stock markets is different from backtest for single instrument.    

It's the step 2 of phase 1 that I am most interested in currently.  So I
think maybe I should not use 'backtest' in the topic.   


Brian G. Peterson wrote:
> Wind wrote:
>> I have checked backtest package.  it seems there is no such functions I
>> needed.
>> I will check it again though.
>> Thanks Gabor.
> Can you please be specific about what features you are looking for? 
> There has been some discussion of creating a more generally applicable 
> backtest framework, so I'm curious what your requirements might be, and 
> how much effort you'd be willing to put into a shared effort.
> Regards,
>      - Brian

View this message in context: http://www.nabble.com/Any-suitable-backtest-functions--tp20789888p21587613.html
Sent from the Rmetrics mailing list archive at Nabble.com.

More information about the R-SIG-Finance mailing list