[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: [amibroker] performance with periodic optimization vs optimizing every day



PureBytes Links

Trading Reference Links





<SPAN 
class=909581303-20102003>sure, I plan to. I need to spend some time making it a 
bit more user friendly<SPAN 
class=744532303-20102003> though. it's a framework you can plug your 
own buy/sell functions into, set equity feedback and other parameters, and 
press Go. because you need to modify the code to use your own indicators etc, it 
really does have to be explained somewhat.
<SPAN 
class=909581303-20102003> 
<SPAN 
class=909581303-20102003>and instead of documenting it, I keep trying to get it 
to turn a healthy profit on some indicator, any indicator, which mostly<SPAN 
class=744532303-20102003> it stubbornly refuses to do. that in turn leads 
me to spend time posting here on how this can be so...
<SPAN 
class=909581303-20102003> 
<SPAN 
class=909581303-20102003>but once it's posted, I'd love your (and everyone's) 
feedback, bug reports, ideas, etc. maybe the lack of performance is the result 
of coding issues.
<SPAN 
class=909581303-20102003> 
<FONT 
color=#0000ff>I'm curious though<SPAN 
class=744532303-20102003>, what about my posting<SPAN 
class=744532303-20102003> sounded promising to you? performance with this simple 
EMAxEMA system bites, so it can't be that, and you and others seem to have 
issues with the whole idea behind it. I'm not trying to chase you away or 
anything, just wondering (:-).
<SPAN 
class=909581303-20102003> 
<SPAN 
class=909581303-20102003>dave
<BLOCKQUOTE 
>
  This looks promising.  Is your auto-optimization code something 
  you canshare with us?Thanks,Howard<SPAN 
  class=744532303-20102003><FONT color=#0000ff 
  size=2> 
   > I modified my 
  auto-optimization code so it can skip a selected number of> bars 
  between optimizations. the idea was to contrast re-optimizing every> 
  day> (what I've been calling auto-optimizing), with the more common 
  method of> optimizing up to some date, using those settings for a 
  while, then> optimizing again at some later date.> > 
  using a simple rule of switching long to short when one EMA crosses> 
  another> EMA, 1/1/95 to present, trading next day at avg price, $7 
  commission,> here's> what I found:> > trading the 
  NASDAQ Composite index:> 
  ----------------------------------------------------> bars 
  skipped      annual 
  return      sharpe ratio> 
  ----------------------------------------------------> 0 (every 
  day)      18.19%      
        .52> 21 (1 
  month)      16.63%      
        .47> 63 (3 
  months)      15.78%      
        .45> 252 (1 
  year)      17.23%      
        .49> 
  ----------------------------------------------------> > > 
  trading the Russell 200 index:> 
  ----------------------------------------------------> bars 
  skipped      annual 
  return      sharpe ratio> 
  ----------------------------------------------------> 0 (every 
  day)      6.82%      
        .16> 21 (1 
  month)      3.82%      
        .04> 63 (3 
  months)      5.64%      
        .11> 252 (1 
  year)      -1.23%      
        -.21> 
  ----------------------------------------------------> > > 
  I'd interpret this as no radical difference. if anything, it favors> 
  optimizing constantly.> > CAVEATS:> this is too small a 
  sample to be anything like conclusive. there are other> other 
  meaningful performance measures besides annual return and sharpe> 
  ratio. this trading rule doesn't perform well enough in any of these> 
  circumstances to be very interesting.> > 
dave






Yahoo! Groups Sponsor


  ADVERTISEMENT











Send BUG REPORTS to bugs@xxxxxxxxxxxxx
Send SUGGESTIONS to suggest@xxxxxxxxxxxxx
-----------------------------------------
Post AmiQuote-related messages ONLY to: amiquote@xxxxxxxxxxxxxxx 
(Web page: http://groups.yahoo.com/group/amiquote/messages/)
--------------------------------------------
Check group FAQ at: http://groups.yahoo.com/group/amibroker/files/groupfaq.html



Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.