|
PureBytes Links
Trading Reference Links
|
Phsst,
Here are some of my peronal understanding about the report statistics
1) Total net profit = Profit of winners - Loss of losers + Interest
earned
There is $720.23 difference I could not explain from your report.
From your report, > 85% of the total net profit is from the
interest earned
when stocks sitting idle.
2) Bars (avg. days) in test: 14623657 (2407)
# of stocks in the portfolio = 14623657 / 2407 ~ 6075
This is only an estimate, as not all stock has the same length of
history.
In AB calculation, you need to have the initial cash of $20000 *
6000 to do
all your trading, hence the very low [Return on account]
3) 2407 / 250 ~ 10 years of history is used.
4) Considering there are 2339 trades, and there are 6000 stocks in the
portfolio,
apparently there are many stocks have never been traded in the
10-year history.
They contributed to the interested earned from $20000 principal for
each stock sitting
idle for 10 years.
Also because of this, the [Exposure] shows only 0.1%, and this very
low exposure
increases your [Risk adjusted ann. return] to 142.34%
Annual system % return: 1.14%
Exposure: 0.1%
Bars out of the market: 14606727
Bars (avg. days) in test: 14623657 (2407)
14606727 / 14623657 ~ 0.9988
1 - 0.9988 = 0.0012 ~ 0.1% [Exposure]
1.14% / 0.0012 ~ 950%
I thought I have figured out why RAR% = 143.34%, apparently
not.
3) Max. system % drawdown
As AB stands today, it is yet to support porfolio testing, this %
is a pessmistic figure
which is taken from the worst MxDD% from whichever single stock
traded.
I also have difficult time understand other [Max....drawdown]
statistics.
I think the best idea is for TJ to publish how he calculate those
numbers.
4) When you say there is no compounding:
Did you set positionsize so that you always buy 1 or 100 shares?
If you do not use positionsize at all, you might be using
100% of you equity all the time for each stock in each trade (I
believe this is
the default setting when positionsize is not used) which IS
compounding because the
more profit you earn, the more equity you have hence the more stock
shares you trade.
(I know there could be argument about the stock price fluctuates,
but let's ignore it now
because the price can go up or down and always affect the
positionsize result)
5) The Profit factor looks good, considering there are so many trades
cover a long history
on a vast portfolio.
The[Avg. trade (win & loss): 822.28] looks good too, if no
compounding is used.
My calculation from report:
average trade = (Profit of winners - Loss of losers) / Total
number of trades
= (3232044.21 - 1308737.32) / 2339 = 1923306.89 /
2339 = 822.277
matches AB calculation
6) I think this is a good system, taking other statistics into
consideration, but initial
cash might be an issue.
I think the selective entry might have contributed the very few
trades seen here,
which might be a good idea in entering a trade when all odds are in
favor.
Thomas
--- In amibroker@xxxxxxxxxxxxxxx, "Jayson" <jcasavant@xxxx> wrote:
> Psst,
>
> "I consider the draw down figures to be suspect along with the
Return on
> Account percentage, and the Buy and Hold data presented is useless
to me."
>
> I am curious, my goal is to out perform Buy and Hold, if I cannot
what is
> the point of putting in all the extra effort? Could you perhaps
explain why
> you feel this number is of no use?
>
>
>
> Jayson
> -----Original Message-----
> From: phsst [mailto:phsst@x...]
> Sent: Friday, March 28, 2003 1:22 PM
> To: amibroker@xxxxxxxxxxxxxxx
> Subject: [amibroker] Re: TRENDING vs. RANGING markets (phsst)
>
>
> Fred,
>
> They are indeed tough questions.
>
> When backtesting entire markets of stocks, the figures that I rely
> upon from the System Report are:
>
> 1) # Trades
> 2) # Winners
> 3) # Losers
> 4) Percent Profitable
> 5) Profit from winners MINUS Loss from Losers (wish Tomasz would
> display this figure as "Net Profit".
> 6) The avg data about # bars / winner or loser
> 7) Risk adj Annual Return
>
> I consider the drawdown figures to be suspect along with the Return
on
> Account percentage, and the Buy and Hold data presented is useless
to me.
>
> Have a good weeekend.
>
> Phsst
>
>
> --- In amibroker@xxxxxxxxxxxxxxx, "Fred" <fctonetti@xxxx> wrote:
> > Phsst,
> >
> > These are all tough questions and would be best answered by the
> > author. In regards to some of your other questions, I personally
> > find all measurements in terms of percents much more informative
than
> > measurements in terms of dollars for the reasons you state. I
don't
> > think the MaxDD% that the reports show when trading more than one
> > vehicle simultaneously are accurate as presented in terms of
account
> > balance but then I am not the author and I have not put much
effort
> > into this area as in general this is not something I do or at
least
> > not within a single account.
> >
> > --- In amibroker@xxxxxxxxxxxxxxx, "phsst" <phsst@xxxx> wrote:
> > > Fred,
> > >
> > > No offense taken. System report is below. I set up the backtest
> > using
> > > $20K per trade with no compounding.
> > >
> > > Let me ask some questions about Amibroker's System Test Report
and
> > the
> > > Equity function(s):
> > >
> > > If a system generates hundreds of trades over a 10 year
backtest,
> > does
> > > AB's sort all of the trades chronologically and then report dd's
and
> > > equity chronologically, or are the System Report figures
generated
> > > from the alphabetically sequenced trades (which would render the
> > > report meaningless)?
> > >
> > > And further...
> > >
> > > What real value are the reported drawdown figures? For example
> > > starting with $20,000 and then after a few years of using a
trading
> > > system that accumulates a portfolio worth over $1,000,000...
what
> > does
> > > a reported Max. system drawdown of $-67,000 and a Max. system %
> > > drawdown of -73.09% really represent? Are the dd% figures based
upon
> > > the original $20 portfolio or are they calculated on the value
of
> > the
> > > portfolio at the time the dd occurs? I know for example that
the
> > > system never experienced a -73% drawdown as measured by current
> > > portfolio value (single digit dd% are more like what I observed
> > > analyzing the trades in Excel).
> > >
> > > And the reported Max. system drawdown of -67,047 might be
> > devastating
> > > against a $100K portfolio, but it is just a minor cost of doing
> > > business when measured against a trading system that has
generated
> > > profit from winners of $3,232,044.
> > >
> > > And what is the difference between max system drawdown and max
trade
> > > drawdown since they are the same figure?
> > >
> > > One more question... I don't understand what the reported "Total
net
> > > profit: 13642497.00" figure is or how it is calculated.
> > >
> > > As you can see, since there are many things I don't understand
about
> > > the Amibroker System Test Report, I use my own methods of
rooting
> > out
> > > the viability of my trading systems.
> > >
> > > Anyway, here is the report: Let me know how you would interpret
this
> > > reported data. Sorry if I sound 'dense', but when I don't
understand
> > > exactly how a reporting tool works, then I tend to ignore it and
do
> > my
> > > own due diligence.
> > >
> > > Overall performance summary
> > >
> > > Total net profit: 13642497.00 Total commissions paid: 56136.00
> > > Return on account: 7.74 % Open position gain/loss 0.00
> > > Buy&Hold profit: 147573728.28 Bars (avg. days) in test:
14623657
> > (2407)
> > > Buy&Hold % return: 83.73% System to Buy&Hold index: -90.76%
> > >
> > > Annual system % return: 1.14% Annual B&H % return: 9.66%
> > >
> > > System drawdown: -13959.68 B&H drawdown: -19973.34
> > > Max. system drawdown: -67047.75 B&H max. drawdown: -76699124.
91
> > > Max. system % drawdown: -73.09% B&H max. % drawdown: -99.91%
> > > Max. trade drawdown: -67047.75
> > > Max. trade % drawdown: -65.76%
> > > Trade drawdown: -11338.25
> > >
> > > Total number of trades: 2339 Percent profitable: 49.8%
> > > Number winning trades: 1165 Number losing trades: 1174
> > > Profit of winners: 3232044.21 Loss of losers: -1308737.32
> > > Total # of bars in winners: 6062 Total # of bars in losers:
3072
> > > Commissions paid in winners: 27960.00 Commissions paid in
losers:
> > > 28176.00
> > >
> > > Largest winning trade: 53881.50 Largest losing trade: -11231.
63
> > > # of bars in largest winner: 7 # bars in largest loser: 2
> > > Commission paid in largest winner: 24.00 Commission paid in
> > largest
> > > loser: 24.00
> > >
> > > Average winning trade: 2774.29 Average losing trade: -1114.77
> > > Avg. # of bars in winners: 5.2 Avg. # bars in losers: 2.6
> > > Avg. commission paid in winner: 24.00 Avg. commission paid in
> > loser:
> > > 24.00
> > > Max consec. winners: 7 Max consec. losers: 5
> > >
> > > Bars out of the market: 14606727 Interest earned: 11718469.88
> > >
> > > Exposure: 0.1% Risk adjusted ann. return: 142.34%
> > > Ratio avg win/avg loss: 2.49 Avg. trade (win & loss): 822.28
> > > Profit factor: 2.47
> > >
> > > Regards,
> > >
> > > Phsst
> > >
> > >
> > > --- In amibroker@xxxxxxxxxxxxxxx, Fred Tonetti <ftonetti@xxxx>
> > wrote:
> > > > Phsst,
> > > >
> > > > This is not meant to offend, but posting a %Return compounded
or
> > not is
> > > > not what I'd call results.
> > > >
> > > > There isn't enough there to judge whether or not a particular
> > avenue is
> > > > even worth investigating. If the system has low dd's and a
fairly
> > > > straight equity curve then it's probably at least worth some
> > > > investigation.
> > > >
> > > > I'll go a little farther then a single number since pictures
are
> > usually
> > > > worth more than words . especially when they have statistics
> > associated
> > > > with them.
> > > >
> > > > PS I changed the equity line indicator recently to use a
> > multilane title
> > > > for the statistics and reposted it in the files section.
>
>
> 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: 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.
------------------------ Yahoo! Groups Sponsor ---------------------~-->
Your own Online Store Selling our Overstock.
http://us.click.yahoo.com/rZll0B/4ftFAA/46VHAA/GHeqlB/TM
---------------------------------------------------------------------~->
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 http://docs.yahoo.com/info/terms/
|