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

Re: [amibroker] Optimize/OverOptimize



PureBytes Links

Trading Reference Links




Dave,
 

<SPAN 
class=626242917-29092003>- sorry to ask, but can you explain the scatter plots 
on slides 3 and 4? what exactly is plotted on x and y? the punch line, which I'm 
too ignorant to see, is that the system fails with out of sample data. the one 
part I understand, I think, is that the correlation coefficient, presumably 
between in and out of sample results, is poor. right? how does the plot show 
this?
 
Please ask Dave directly.  I don't want 
"interpret" his material.  A few group members have had this presentation 
for a day or so and some or all of the presentation is beyond most of our math 
capabilities.  I took notes during the presentation... after I quaffed a 
couple of pints at the local pub.
 

<SPAN 
class=626242917-29092003>- slide 24 mentions "Trend Following on Commodities", 
as "100 day lookback, trade 34% before breakout". I don't understand what this 
means. something about MA or EMA(100), maybe, but what's the 34% 
piece?
<SPAN 
class=626242917-29092003> 
I believe Dave trades nine 
commodities with a modified "Turtles/Aberration system".  Since, the system 
only works on commodities and doesn't begin to fit my style or preference, I 
have never sought to understand the rules.  I have enough trouble 
remembering my own boundaries.  Sorry.
 

<SPAN 
class=626242917-29092003>- how would I compute the daily standard deviation of 
the S&P500 in AB, in a way that gives the same .95%/day figure he mentions? 
I ask so I can generate similar figures for other markets.
<SPAN 
class=626242917-29092003> 
Another Dave 
question.
 
<SPAN 
class=626242917-29092003>- the parameters I get 
optimizing today compensate for transient market behaviors that will eventually 
end, and eventually it will do very poorly. but if those behaviors persist, at 
least somewhat, for a little while, might the system to do better than 
average in the short term? if so, is constant re-optimization worth 
exploring?
<SPAN 
class=626242917-29092003><SPAN 
class=626242917-29092003> 
<SPAN 
class=626242917-29092003>Lot's of things to 
consider here.  This is probably the real "meat" of the optimizing 
issue.  One of the biggies is the number of variables when you conduct your 
optimizations.  The fewer the variables, the less chance that 
eventually an approach will do poorly.  

<SPAN 
class=626242917-29092003><SPAN 
class=626242917-29092003> 
<SPAN 
class=626242917-29092003>The equity and commodity 
markets are not horse races:  these "bets" are not handicapped with 
assigned odds.  Seems to me that since all the horses in the race are 
"even-money", one should chart the "form" and only bet on the ponies that have 
performed the best (over the most recent period of time).  

<SPAN 
class=626242917-29092003><SPAN 
class=626242917-29092003> 
<SPAN 
class=626242917-29092003>Expecting a particular 
stock to test well with data from 1997 is foolish.  Expecting a particular 
system to trade well in 1997 is not foolish.  Issues fall into and out of 
patterns of accumulation and distribution.  Sometimes, supply and 
demand takes a trip down "random walk street".  I think it makes 
sense, if you have a sound approach to the markets, to concentrate on "ranking" 
issues based on return.  If you concentrate your efforts on selecting 
issues that have consistently performed with your systems parameters, the odds 
favor continued consistency.  If supply and demand changes it's character 
for a particular issue...you must stop trading it.  

<SPAN 
class=626242917-29092003><SPAN 
class=626242917-29092003> 
<SPAN 
class=626242917-29092003>I believe, constant 
reoptimization is the way to go.  What is the definition 
of constant?  For me, I reoptimize every quarter.  Most of the 
time, reoptimization does not mean a reassignment of triggers or filters.  
Most of the time, it means adjusting WHAT I'm trading (adding and subtracting 
issues in my universe)...not HOW I'm trading.  

<SPAN 
class=626242917-29092003><SPAN 
class=626242917-29092003> 
<SPAN 
class=626242917-29092003>Please share if Dave 
responds to your questions,
<SPAN 
class=626242917-29092003><SPAN 
class=626242917-29092003> 
<SPAN 
class=626242917-29092003>Take 
care,
<SPAN 
class=626242917-29092003><SPAN 
class=626242917-29092003> 
<SPAN 
class=626242917-29092003><SPAN 
class=626242917-29092003>Steve
<BLOCKQUOTE 
>
  ----- Original Message ----- 
  <DIV 
  >From: 
  Dave Merrill 
  
  To: <A title=amibroker@xxxxxxxxxxxxxxx 
  href="">amibroker@xxxxxxxxxxxxxxx 
  Sent: Monday, September 29, 2003 12:07 
  PM
  Subject: RE: [amibroker] 
  Optimize/OverOptimize
  
  <SPAN 
  class=626242917-29092003>thanks very much for this steve. very interesting, 
  though I don't get all of it. a topic I'm thinking about pretty much 
  constantly these days.
  <SPAN 
  class=626242917-29092003> 
  <SPAN 
  class=626242917-29092003>a couple of questions, if I 
  might...
  <SPAN 
  class=626242917-29092003> 
  <SPAN 
  class=626242917-29092003>- sorry to ask, but can you explain the scatter plots 
  on slides 3 and 4? what exactly is plotted on x and y? the punch line, which 
  I'm too ignorant to see, is that the system fails with out of sample data. the 
  one part I understand, I think, is that the correlation coefficient, 
  presumably between in and out of sample results, is poor. right? how does the 
  plot show this?
  <SPAN 
  class=626242917-29092003> 
  <SPAN 
  class=626242917-29092003>- slide 24 mentions "Trend Following on Commodities", 
  as "100 day lookback, trade 34% before breakout". I don't understand what this 
  means. something about MA or EMA(100), maybe, but what's the 34% 
  piece?
  <SPAN 
  class=626242917-29092003> 
  <SPAN 
  class=626242917-29092003>- how would I compute the daily standard deviation of 
  the S&P500 in AB, in a way that gives the same .95%/day figure he 
  mentions? I ask so I can generate similar figures for other 
  markets.
  <SPAN 
  class=626242917-29092003> 
  <SPAN 
  class=626242917-29092003>- the parameters I get optimizing today compensate 
  for transient market behaviors that will eventually end, and eventually it 
  will do very poorly. but if those behaviors persist, at least somewhat, 
  for a little while, might the system to do better than average in the short 
  term? if so, is constant re-optimization worth exploring?
  <SPAN 
  class=626242917-29092003> 
  <SPAN 
  class=626242917-29092003>thanks again for passing this on. makes me wish I 
  lived somewhere nearby...
  <SPAN 
  class=626242917-29092003> 
  <SPAN 
  class=626242917-29092003>dave
  <BLOCKQUOTE 
  >
    <FONT face=Arial 
    size=2>Dave is an Agilent, triple-degreed, engineer.  Two weeks ago, he 
    presented this work to our Denver Trading Group's weekly meeting (actually, 
    this group meets every Thursday and most Saturday's).  Once a 
    month, I moderate a SIG on mechanical trading (and I haven't 
    seen less than eighty people in the room since I've been 
    attending).
    
     
    Although, I don't agree with certain aspects of 
    his presentation and I somewhat object to his assigning my name to the 
    "Karnish System" (it has become a 
    bastardized off-shot of my work), I still believe that there is a lot of 
    merit to aspects of his work.  The "Karnish System" has become the 
    moniker for systems (along the front range of Colorado) that 
    stochastically smoothes a momentum oscillator that initiates buy and sell 
    signals using symmetrical triggers.  
     
    I neither want to endorse, defend or criticize 
    Dave's work...but, offer this for group members to stimulate 
    thought.
     
    Take care,
     
    SteveSend 
  BUG REPORTS to bugs@xxxxxxxxxxxxxSend SUGGESTIONS to 
  suggest@xxxxxxxxxxxxx-----------------------------------------Post 
  AmiQuote-related messages ONLY to: amiquote@xxxxxxxxxxxxxxx (Web page: <A 
  href="">http://groups.yahoo.com/group/amiquote/messages/)--------------------------------------------Check 
  group FAQ at: <A 
  href="">http://groups.yahoo.com/group/amibroker/files/groupfaq.html 
  Your use of Yahoo! Groups is subject to the <A 
  href="">Yahoo! Terms of Service. 







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.