Rank: Advanced Member
Groups: Registered, Registered Users
Joined: 3/30/2009(UTC)
Posts: 46
Location: Hawaii, US
|
Hi, I tried applying trailing stops to a system created in the system tester in MS 10, and results seem unlikely. There was no drop off in performance with an excessively tight stop - performance increases steadily from using no stop, to 10% trailing stop, 5%, 1% and even .01%. Average wins do decrease, but avg losses go to zero for stops 5% or less, on a 500 period test. Adverse excursions listed are in excess of the stop value. The data used is yahoo EOD data and does include open, high and low in addition to the close, when calling up the chart it appears correct. Any help appreciated - thanks!
|
|
|
|
|
|
Rank: Advanced Member
Groups: Registered, Registered Users, Subscribers, Unverified Users
Joined: 10/28/2004(UTC)
Posts: 3,111
Location: Perth, Western Australia
Was thanked: 16 time(s) in 16 post(s)
|
Hi manatrader,
Welcome to the Forum.
What is the system test? If the system exit is exiting the trade before the trailing stop gets to kick in, then it doesnt matter what value you have for the trailing stop.
wabbit [:D]
|
|
|
|
|
|
Rank: Advanced Member
Groups: Registered, Registered Users
Joined: 3/30/2009(UTC)
Posts: 46
Location: Hawaii, US
|
Thanks wabbit, I'm running a bunch of different tests, and they all only get better with a trail stop, no matter how close it follows. With a really close stop, the number of wins starts to approach the periods in the test (485 in 500 periods, when it traded 100 with no stop), with losers near 0 (2, 50 with no stop), and again losers averaging $0. Not sure how to treat the data, but it makes me unsure to be sure..
|
|
|
|
|
|
Rank: Advanced Member
Groups: Registered, Registered Users, Subscribers, Unverified Users
Joined: 10/28/2004(UTC)
Posts: 3,111
Location: Perth, Western Australia
Was thanked: 16 time(s) in 16 post(s)
|
I will have a look as I am now curious about what could cause such EST events?
Can you please use the Organiser function in MS to export one of the problematic system tests and send to me via email? Or send me the system test (if it's a built in one) with all of the settings, by email.
wabbit [:D]
|
|
|
|
|
|
Rank: Advanced Member
Groups: Registered, Registered Users
Joined: 3/30/2009(UTC)
Posts: 46
Location: Hawaii, US
|
Thanks Wabbit, I tried to export using the organizer but it's not in the list of items to export. I think the error will show up comparing any MS test with no stop to one with a tight trailing stop - maybe not if using tick data, again I'm using daily with OHLCV. Here's one test:
B: Close<Mov(C,Opt1,E)
S: same but >
SS: >
BC: <
So, it is always in the market, and the system reenters the next day. So it makes sense the number of trades goes up with a fairly close trail stop, but the results can't be right..
I posted on another thread on making stop rules in the buy/sell section instead of using the stop tab - not sure how to reference the entry price, or how the system would calculate the close of a position if it doesn't have intraday data
|
|
|
|
|
|
Rank: Advanced Member
Groups: Registered, Registered Users
Joined: 3/30/2009(UTC)
Posts: 46
Location: Hawaii, US
|
Well, thanks to wabbit I got a good start on making a trailing stop rule in the buy/sell section of the strategy tester..
Sell to exit long position:
entryPrice:=ValueWhen(1,Simulation.CurrentPositionAge=0,OPEN);
stopPrice:=HHV(HIGH,XX) - (entryPrice*OPT4);
exitPrice:=if(ValueWhen(1, Ref(C,-1)<stopPrice,stopPrice)
or {winning exit} rsi(2)>80, open, c);
{return}
exitPrice;
The problem is the XX, I don't know how to specify the highest high while a position is open. Any help? Thanks!
|
|
|
|
|
|
Rank: Advanced Member
Groups: Registered, Registered Users, Subscribers, Unverified Users
Joined: 10/28/2004(UTC)
Posts: 3,111
Location: Perth, Western Australia
Was thanked: 16 time(s) in 16 post(s)
|
entry:=Simulation.CurrentPositionAge=0;
entryPrice:=ValueWhen(1,entry,OPEN);
stopPrice:=highestsince(1,entry,H) - (entryPrice*OPT4);
exitPrice:=if(ValueWhen(1, Ref(C,-1)<stopPrice,stopPrice)
or {winning exit} rsi(2)>80, open, c);
{return}
exitPrice;
wabbit [:D]
|
|
|
|
|
|
Rank: Advanced Member
Groups: Registered, Registered Users
Joined: 3/30/2009(UTC)
Posts: 46
Location: Hawaii, US
|
Awesome wabbit, that does the trick!
EDIT - oops, position info indicates positions are being closed at next open. Looks like max loss stop from the stops tab is the only thing giving good results when using daily OHLCV data. Regards :)
|
|
|
|
|
|
| Users browsing this topic |
|
Guest (Hidden)
|
Forum Jump
You cannot post new topics in this forum.
You cannot reply to topics in this forum.
You cannot delete your posts in this forum.
You cannot edit your posts in this forum.
You cannot create polls in this forum.
You cannot vote in polls in this forum.
