|
PureBytes Links
Trading Reference Links
|
Thanks, Dave. My problem is that I am going to test swing trade setups and
they don't have a lot of trades on an individual stock sometimes. But when
you take them in context of 100's of stocks, they work fine. So I'm
interested in getting the data correct to build a valid data base of maybe
thousands of trades.
So, I'd like to get the number of trades, winners, equity, etc accurate.
Now it appears that "Periods" defines not only defines the Indicator
duration (a MA in this case) but could also be used as a valid starting
point for the data. For example, if you had a 50 day MA then the indicator
is not valid until you have 50 days of data. Thus you would have up to the
the number of "Periods" to do the intitalization on the POSN.
So what if you added an if statement to the appropriate formulas to
disregard all the trades (specifically the setup trade) prior to the time
the data was valid? Examples below.
EQTY:=If(Cum(1)<Periods,0,Cum(If(Ref(POSN,-1)=1,EntryPoint-Ref(EntryPoint,-1
),0)));
{keeps equity at zero through the # of periods. You would then force your
dummy trade to happen prior to the end of periods}
TotTrades:=If(Cum(1)<Periods,0,Cum(Cross(POSN,0.5)));
WinTrade:=If(Cum(1)<Periods,0,(ValueWhen(1,Ref(Cross(POSN,0.5),-1),EntryPoin
t)<ValueWhen(1,Ref(Cross(0.5,POSN),-1),EntryPoint)));
Winners:=If(Cum(1)<Periods,0,Cum(Cross(WinTrade,0.5)));
This would say that you would use the "Periods" even if your indicator
didn't require it for the trade. You might set it at 3 for a minimum and
set the EnterLong dummy at 1 and ExitLong dummy at 2.
Thoughts?
Tom
----- Original Message -----
From: "Dave Nadeau" <dave_nadeau@xxxxxxxxx>
To: <metastock@xxxxxxxxxxxxx>
Sent: Thursday, May 24, 2001 11:29 PM
Subject: RE: System Tester Equity Line
> Tom,
>
> That's exactly the reason. I'm forcing a quick one day trade as soon as
the
> moving averages 'come alive' (or rather become defined). Because I have
to
> use the BarsSince() function to determine if the system is long or short,
> this quick trade starts the comparison. Otherwise, the POSN formula
misses
> the first trade. It's not a big deal if you have a system that generates
> many trades, and the one trade for one bar has a negligible effect,
> especially if you have a security that's split many times, and the early
> data has ridiculously small values. On the other hand, if you have only a
> few trades over a fewer number of bars, like a daily 90 day S&P contract,
> then it's a "lesser of two evils" solution.
>
> Dave
>
> -----Original Message-----
> From: owner-metastock@xxxxxxxxxxxxx
[mailto:owner-metastock@xxxxxxxxxxxxx]On
> Behalf Of Tom Sprunger
> Sent: Thursday, May 24, 2001 4:09 PM
> To: metastock@xxxxxxxxxxxxx
> Subject: Re: System Tester Equity Line
>
>
> Dave, I have a question on the code you provided.
>
> On the two lines for the EnterLong and ExitLong:
>
> EnterLong:=If(Cum(1)=Periods+3,1,Cross(C,Mov(C,Periods,S)) AND
> (V>1.1*Ref(V,-1) AND Ref(V,-1)>1.1*Ref(V,-2)));
>
> ExitLong:=If(Cum(1)=Periods+4,1,Cross(Mov(C,Periods,S),C));
>
> Why do you have the "Cum(1)=Periods +3,1" in the EnterLong and the
> "Cum(1)=Periods+4,1" in the ExitLong?
>
> It would appear that this gives you a forced entry 13 days after the start
> of the chart and a forced exit 14 days after the start of the chart. Is
> this to initialize? If so, why?
>
> By the way this is pretty slick!
>
> Thanks for you answer.
>
> Tom
>
>
> _________________________________________________________
> Do You Yahoo!?
> Get your free @yahoo.com address at http://mail.yahoo.com
>
|