tennisplayer2
210 posts
msg #107385
- Ignore tennisplayer2 |
8/6/2012 12:31:17 PM
Please correct me if I'm wrong, but I thought that when %D5 crosses above %D10, we should be long the market (in SSO) and when %D5 crosses below %D10 we should be out of the market (or in SDS). Thanks.
|
Kevin_in_GA
4,599 posts
msg #107386
- Ignore Kevin_in_GA |
8/6/2012 12:49:31 PM
Consider yourself corrected - it is the other way. Seems counterintuitive but that's the way the filter works.
|
VenturaTrader
43 posts
msg #107405
- Ignore VenturaTrader |
8/6/2012 10:40:28 PM
Also, to make it simpler just ignore the "go_short" column, as this is simply to confirm that the mirror image ETF is sending the opposite signal at the same time.
***********************************************************************************************************************************
Way to go Kevin!!
To avoid confusion I just took the "go_short" column out of your filter so the short signals are not seen. For those of us who are color blind it makes it easier to see the signals. Nice job!
Mike
|
Eman93
4,750 posts
msg #107407
- Ignore Eman93
modified |
8/6/2012 11:17:42 PM
It looks like on the back test you did not utilize the 3% trail..
It says max draw down is 25%
Average position draw down is 5.08%
worst year is 6% and best year is 280%
That is also kind of a flag to me also that you could have that big a year. I think SDS did a reverse spit that could skew the results.
I just want people to understand the results need to be full analyzed and don't just throw your total account into it....
as you can see from this chart it is a good system...
http://stockcharts.com/public/1107832/chartbook/250425937;
Nice work and always enjoy your contributions.
|
Kevin_in_GA
4,599 posts
msg #107414
- Ignore Kevin_in_GA
modified |
8/7/2012 8:59:44 AM
The data used in the backtest is from Yahoo. I think (but will check to be sure) that they have corrected for the splits and those crazy distributions they used to do every quarter.
The larger losses are the result of gaps down/up at the open.
UPDATE: Yahoo historical prices for both of these ETFs are corrected for splits and dividends.
|
tombrown1
61 posts
msg #107420
- Ignore tombrown1 |
8/7/2012 10:50:25 AM
I did get results from both etfs. I think I'm going to do a full manual backtest over the next couple of days from the practical inception of these funds. I've never manually tested a trailing stop though. Is the trailing stop only calculated by the closing price?
TB
|
Kevin_in_GA
4,599 posts
msg #107421
- Ignore Kevin_in_GA |
8/7/2012 10:59:44 AM
Trailing stops are usually from the high.
I would suggest back testing without a stop loss first - it is easier and from what I have seen also quite profitable.
|
drew9
171 posts
msg #107426
- Ignore drew9 |
8/7/2012 12:20:14 PM
Hi Kevin,
After reviewing this, it is again some really nice work! If you really look at the signals, they tend to be a leading indicator of what is to come. If you moved the entries one or two days after the signal, the results appear to be much better. Perhaps you may want to run the backtest and try it? I think this is really impressive to use as a heads up with other indicators to confirm possible entries and exits. It seems to be dead on with a two day lag in choppy markets. I love leading indicators and this is one to add to the collection. Thanks!
|
tombrown1
61 posts
msg #107428
- Ignore tombrown1 |
8/7/2012 1:30:09 PM
I understand they are from the high, but intraday high or end of day high? Also, do you sell intraday when stop is hit immediately? I can take some guesses, but I just wanted to make sure we are on the same page.
Best,
Tom
|
Kevin_in_GA
4,599 posts
msg #107429
- Ignore Kevin_in_GA |
8/7/2012 1:58:09 PM
In backtesting, the daily low will be used to see if the trailing stop is triggered. Looking this up, the trailing stop is based off the close, not the high. Here is the Metastock formula for calculating it.
SVE_Stop_Trail%
perc:=Input("Trailing Loss % :",0,100,14);
loss:=C*perc/100;
trail:=
If(C>PREV AND Ref(C,-1)>PREV,
Max(PREV,C-loss),
If(CMin(PREV,C+loss),
If(C>PREV,C-loss,C+loss)));
Trail
|