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

[amibroker] Compounding, optimizing techniques, etc. (for phsst)



PureBytes Links

Trading Reference Links




Thanks for your 
comments.
<FONT face=Arial 
size=2> 
I wasn't 
very articulate when I said that "since my 
largest loss can only be $10,000 and the profit can be in the millions, I keep 
in all losing trades when evaluating system performance".    On 
it's own merit, that sentence is silly.   Of course, the maximum loss 
could be anything depending on how much is invested.   The $10,000 
figure came from an earlier posting, but it is the total invested in each 
position.   Every buy or short had a PositionSize of 
$10,000.   So, the most I might lose would be 
$10,000.    Out of the 45,000 trades, hundreds generated profits 
of more than $750,000 on the $10,000 investment.   I eliminated the 
stocks where trades generated more than $200,000 profit (about 100 stocks) and 
re-optimized.  I didn't eliminate the stocks with the largest losing 
trades because there were few of them and the losses, albeit high in percentage 
terms, were negligible compared to the portfolio size.   

<FONT face=Arial 
size=2> 
I went on to 
say that in the backtesting and in real life, I have lost more than the original 
investment.   You then said "I have 
NEVER experienced losses in a short position which even approached the original 
investment".   I don't doubt you.   But in that particular 
example I posted earlier, there were 45,000 trades and there were eight examples 
where overnight gaps caused losses on shorts that exceeded the initial 
investment.    It's fairly easy to imagine.   I use 40% 
stops, only to keep investors happy.   The systems I tend to develop 
simply don't work well with stops, so I don't use the normal 10-20% stop 
arrangement..   A short position could be hovering around the 30% 
loss point and then gap up enough to create more than a 100% loss in one 
tick.   I was only stating that it could happen to prevent 
someone from correcting me.   Eight trades out of 45,000 is just about 
"none".
<BLOCKQUOTE 
>
  <FONT face="Times New Roman" 
  size=2>-----Original Message-----From: phsst 
  [mailto:phsst@xxxxxxxxx]Sent: Friday, April 18, 2003 1:08 
  AMTo: amibroker@xxxxxxxxxxxxxxxSubject: [amibroker] Re: 
  Chuck---To compound or not to compound... that is the 
  questionChuck,Just spent a lot of time trying 
  to catch up on more than 24 hoursworth of posts and could not pass up the 
  opportunity to comment uponyour posts during that period.Wish I 
  was as articulate about these matters as 
  you.Regarding:Compounding... I agree. Fixed position sizing is 
  mandatory.Positionsize compounding only skewx trading system results to 
  thepoint of losing all relevence in real world trading."individual 
  set of parameters for each stock."   Talk about 
  'curvefitting'Unlike you, I don't use 'extinct' stocks, only 
  because my dataprovider eliminates extinct stocks from my 
  database.Concerning the statement "Do I set aside any losing 
  trades/stocks? No.   Since my largest loss can only be $10,000 
  and the profit can bein the millions, I keep in all losing trades."  
  While unusual, I haveexperienced adverse 'gap' conditions that exceed 
  $10K, it is morenormal for Stop Loss orders get me out of positions at 
  fractions of apercent below my entry level."As I explained in 
  previous email, the losses CAN BE morethan the investment on a few short 
  trades."  We differ here... I haveNEVER experience losses in a short 
  position which even approached theoriginal investment.It is late 
  and I need to call it quits for the night. But I hope you accept my 
  compliments for being able to articulate arational trading 
  approach.Phsst--- In amibroker@xxxxxxxxxxxxxxx, "Chuck 
  Rademacher"<chuck_rademacher@x> wrote:> MessageAnswers to 
  Dingo's questions below:> > 1.  Which column(s) to use when 
  deciding which parameters are"best"?   I> actually take 
  the trade results to another program, but given thecolumns> that 
  are available in AB, I would use total net profit and/oraverage 
  trade.> Hopefully, some of the parameters will be same at the top of 
  both ofthose> lists.> > 2.  Which date ranges do 
  I use?   I pretty much always use 1992 tocurrent> 
  date.> > 3.  I use all 13,500 active and extinct stocks 
  that have ever tradedsince> January, 1992 that were over $1 on at 
  least one day and had an average> 50-day volume of 75,000 shares on at 
  least one day.> > 4.  Do I set aside any losing 
  trades/stocks?  No.   Since my largestloss> can only 
  be $10,000 and the profit can be in the millions, I keep in all> losing 
  trades.  As I explained in previous email, the losses CAN BE more> 
  than the investment on a few short trades.> > 5.  How do I 
  determine that the best parameters (if I include huge> profitable 
  trades) could be on the edge of the parameter space?   
  Ican't> really determine that information, nor do I really 
  care.   I removethose> stocks from my watchlist and 
  re-optimise.   If I end up with a newlist of> huge 
  winners, I might repeat the process.   I just don't want 
  tochoose a> parameter that squeezes in a couple of huge trades but 
  is inferioron the> rest of the market.> > 6.  
  How do I decide which signals to take every day (if I had 
  moresignals> than cash)?   Using AB, I would place 
  something that I could sort the> signals by that I have proven to 
  myself increases the likelihood of the> signal being a good 
  one.   For simplicity, let's say that I was anadvocate> 
  of using low P/E for buying stocks (I'm not).  I would add a columnto 
  my> Explore that showed the current P/E.> > > I'm 
  quite happy to discuss any of this further.   Obviously, 
  theseare only> my views and they are no better than those you or 
  others might haveon the> subject.   I have the benefit of 
  trading for (too) many years and the> possible disadvantage of "being 
  past it".>   -----Original Message----->   
  From: dingo [mailto:dingo@xxxx]>   Sent: Thursday, April 17, 
  2003 11:10 AM>   To: 
  amibroker@xxxxxxxxxxxxxxx>   Subject: RE: [amibroker] To 
  compound or not to compound... that is the> question> > 
  >   I can understand and appreciate why you use fixed trade 
  sizes inorder to> get the best parameters. But how do you get a 
  reasonable measure of> drawdowns that way?  Do you use some other 
  technique to evaluatedrawdowns?> >   Re your param 
  selection method: Do I understand the steps correctly:> 
  >   1. You optimize for the best 
  params>           a. 
  Based on what column or 
  calculation?>           
  b. What date ranges would you be using 
  currently?>           
  c. What subset of stocks would you be optmizing on?> 
  >   2. You set aside the the top 
  100.>           a. Do 
  you set aside any at the 
  bottom?>           b. 
  How did you determine that the first set of paramswould be at> the 
  edge of the parameter space?> >   3. You reoptimize 
  the resultant set from step 2 and those are theones you> 
  use.> >   Given the size of your trading capital how 
  do you decide whatstocks to> trade on a particular day?> 
  >   I'm not trying to pick a fight here I'm intensely curious 
  as I've been> struggling with these questions for quite some time 
  now.> >   Thanks for any comments you choose to 
  make.> >   d>     
  -----Original Message----->     From: Chuck 
  Rademacher [mailto:chuck_rademacher@xxxx]>     
  Sent: Thursday, April 17, 2003 6:58 AM>     To: 
  amibroker@xxxxxxxxxxxxxxx>     Subject: [amibroker] 
  To compound or not to compound... that is the> question> 
  > >     Reply to Fred:> 
  >     Yes... and no.> 
  >     Absolutely, in real time trading I am 
  compounding.> >     To determine parameters 
  via optimization.... not if my lifedepended on> it!   
  And, I guess my life does depend on it, as I make my 
  livingmanaging> funds for others.> 
  >     I mentioned one trade (AOL) where my system 
  made $1.5 million on a> $10,000 investment.  That's not 
  bragging... I'm sure you could comeup with> a system that could 
  achieve similar performance.   Since the averagetrade> 
  generated a profit of $2,700 for every $10,000 invested, the AOLtrade 
  could> cover up lots of bad trades made using one parameter set.  
  Compounding that> trade would exacerbate the problem.   A 
  minor tweak to theparameters could> cut out the AOL trade, yet that 
  very tweak could improve performancegoing> forward.> 
  >     When choosing parameters, I want plain 
  vanilla trades, eachstanding on> their own merit, with no 
  compounding.> >     We may have to agree to 
  disagree.   It's like absolute gospel tome and> I cannot 
  see clear to do it any other way.>       
  -----Original Message----->       From: 
  Fred [mailto:fctonetti@xxxx]>       Sent: 
  Thursday, April 17, 2003 3:16 AM>       
  To: amibroker@xxxxxxxxxxxxxxx>       
  Subject: [amibroker] FW: [aaft_ta] Re: TradingRecipes> > 
  >       Chuck,> 
  >       I'm sure you'd agree, wouldn't 
  you ?, that one way or another you>       
  compound.  If you are not compounding by increasing bet sizethen 
  you>       are compounding by increasing 
  the number of stocks 
  you'llpotentially>       take 
  simultaneous positions in as equity grows, right ?> 
  >       --- In amibroker@xxxxxxxxxxxxxxx, 
  "Chuck Rademacher">       
  <chuck_rademacher@x> wrote:>       
  > For what it is worth, I use fixed bet size for all 
  backtesting>       purposes.   
  I>       > coudn't imagine 
  backtesting/optimizing using any otherapproach.  
  I>       even 
  go>       > a step further if I'm 
  doing any optimizing.   I recentlyposted 
  an>       
  equity>       > curve showing 
  something like $80 million in profit.   
  Withinthat>       
  $80>       > million, the top 100 
  stocks (out of 13,500) generated $20million 
  in>       > profits.  AOL, by 
  itself, generated $1.5 million in profits.  
  In>       each 
  case,>       > the original trade was 
  only $10,000.>       
  >>       > As I said, I go a step 
  further than just using a fixed bet 
  size.>       After 
  my>       > first pass at optimizing, 
  I remove the top performing 
  100stocks.>       I 
  then>       > re-optimize without 
  those stocks.  Granted, I could end up 
  with>       some 
  new>       > "top" stocks.  
  However, my objective is to remove the 
  extremely>       
  large>       > winners so that the 
  profits from those stocks don't cause me 
  to>       
  select>       > parameters on the edge 
  of the parameter space.>       
  >>       > I don't bother removing 
  the worst performers as the largest 
  loss>       might 
  be>       > something like $16,000 
  (even though the original trade was 
  only>       
  $10,000).>       > This can happen if 
  a short trade goes against you.>       
  >>       > As I said... for what 
  it's worth...>       >   
  -----Original Message----->       
  >   From: Bob Jagow 
  [mailto:bjagow@xxxx]>       
  >   Sent: Thursday, April 17, 2003 2:21 
  AM>       >   To: 
  Amibroker>       >   
  Subject: [amibroker] FW: [aaft_ta] Re: 
  TradingRecipes>       
  >>       
  >>       >   Re the 
  "portfolio level testing" magic 
  bullet.>       
  >>       >   
  Bob>       >   -----Original 
  Message----->       >   
  From: Palmer Wright 
  [mailto:palmerw@xxxx]>       
  >   Sent: Wednesday, April 16, 2003 8:27 
  PM>       >   To: 
  aaft_ta@xxxxxxxxxxxxxxx>       
  >   Subject: Re: [aaft_ta] Fwd: Re: Available Portfolio 
  testing>       programs 
  for>       > 
  TS2000i>       
  >>       
  >>       >   Since 
  Michael forwarded the two messages (see below), he 
  added>       
  four>       > additional ones. The 
  issue about whether a "basket system" 
  like>       
  Aberration>       > is worth trading I 
  will not discuss here (I still trade it). 
  The>       other 
  main>       > issue is about the 
  effect of compounding when testing with 
  TR>       
  (Trading>       > Recipes), and I 
  comment here on that.>       
  >>       >   Traders buy 
  TR because it can test portfolios of systems 
  and>       markets 
  using>       > position sizing. A 
  position-sizing strategy such as 
  fixed->       fractional 
  money>       > management brings two 
  advantages: it normalizes markets 
  (eg.,>       
  calculating>       > many contracts 
  for corn, but few for natural gas), andlimits 
  entry>       risk 
  for>       > each position to a fixed- 
  fraction of current equity--thus>       
  preventing>       > overtrading. If 
  you do not use TR, I do not know how you can 
  get>       the 
  large>       > returns that 
  compounding multiple markets can 
  bring.>       
  >>       >   Leslie Walko 
  points to the potential danger of curve 
  fitting>       caused 
  by>       > compounding. I agree, and 
  have been concerned for yearsabout 
  how>       one 
  market>       > in a portfolio 
  (commodity X) by being dramaticallyprofitable in 
  a>       
  single>       > year can misleadingly 
  bias the results of the whole 
  portfolio.>       
  >>       >   During a 
  multi-year test in TR, starting equity is 
  low,perhaps>       
  $100,000,>       > but compounding 
  raises equity to many million in lateryears. 
  The>       
  one-year>       > outperformance of 
  commodity X cand produce two kinds of 
  curve->       fitting 
  bias:>       > early-years bias and 
  end-years bias. Mark Johnson's 
  message>       describes 
  the>       > first, where X gives "a 
  big turbocharged boost" to 
  theportfolio's>       
  equity,>       > which then gives a 
  head-start boost to the number of tradesin 
  all>       
  the>       > commodities traded. The 
  second occurs when X's monstertrades 
  occur>       in 
  the>       > final years of the 
  simulated time period when the largenumber 
  of>       
  contracts>       > makes X's profit 
  far larger than if its big year came 
  early.Here>       
  the>       > profits contributed by X 
  dwarf what they were in the first 
  case.>       
  >>       >   As the 
  message from M points out, we can avoid such biases 
  by>       
  normalizing>       > with a 
  fixed-dollar bet size in testing to remove the 
  galloping>       
  equity>       > effect. I proposed 
  this method in 1999, and still use it 
  tocompare>       with 
  the>       > compounded performance. I 
  confess, however, that my testing 
  has>       failed 
  to>       > find as much performance 
  bias as I suspected I would find. 
  The>       method 
  is>       > most important when 
  selecting markets for a portfolio.>       
  >>       >   Palmer 
  Wright>       
  >     ----- Original Message 
  ----->       >     
  From: Michael Guess>       
  >     To: 
  aaft_ta@xxxxxxxxxxxxxxx>       
  >     Sent: Sunday, April 13, 2003 9:14 
  AM>       >     
  Subject: [aaft_ta] Fwd: Re: Available Portfolio 
  testing>       programs 
  for>       > 
  TS2000i>       
  >>       
  >>       >     
  This is for Pat Mazur & Palmer Wright. Others are invited 
  to>       comment. 
  I>       > forwarded these two 
  messages from another list because we 
  have>       
  discussed>       > these issues in the 
  past. It appears one of the posts is 
  saying>       
  Trading>       > Recipes is in error 
  in the way it calculates. In fact, that 
  it>       curve 
  fits>       > data in a particular 
  case. Comments are invited.>       
  >>       >     
  Michael>       
  >>       
  >>       
  >>       >   Your use of 
  Yahoo! Groups is subject to the Yahoo! Terms 
  of>       
  Service.>       
  >>       
  >         Yahoo! Groups 
  Sponsor>       
  >               
  ADVERTISEMENT>       
  >>       
  >>       
  >>       
  >>       >   Send BUG 
  REPORTS to bugs@xxxx>       
  >   Send SUGGESTIONS to 
  suggest@xxxx>       >   
  ----------------------------------------->       
  >   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 Yahoo! Terms 
  of>       Service.> > 
  > >       Send BUG REPORTS to 
  bugs@xxxx>       Send SUGGESTIONS to 
  suggest@xxxx>       
  ----------------------------------------->       
  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 Yahoo! Terms ofService.> > > 
  >     Send BUG REPORTS to 
  bugs@xxxx>     Send SUGGESTIONS to 
  suggest@xxxx>     
  ----------------------------------------->     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 
  Yahoo! Terms of Service.> > 
  >         Yahoo! Groups 
  Sponsor> > > >   Send BUG REPORTS to 
  bugs@xxxx>   Send SUGGESTIONS to 
  suggest@xxxx>   
  ----------------------------------------->   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 Yahoo! Terms 
  of Service.Send 
  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












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.