mahkoh
1,065 posts
msg #100824
- Ignore mahkoh |
5/20/2011 7:00:15 PM
I would like to add something from one of Kevin's other posts:
Check this link daily:
http://stockcharts.com/h-sc/ui?s=$VIX:$SPX&p=D&yr=1&mn=0&dy=0&id=p72249291784
When ROC (15) is above 0 and VIX-SPX correlation is above SMA close your position (and/or go long volatility)
|
Kevin_in_GA
4,599 posts
msg #100829
- Ignore Kevin_in_GA
modified |
5/20/2011 10:50:03 PM
Note that as of today the filter says to be in bonds (actually had the call made a few days ago but the system is backtested and based on weekly close data).
The $VIX:$SPX chart mahkoh referenced had you out of equities last Friday, but tends to act more frequently and can get whipsawed.
If you substitute TLT for AGG, it had you out of equities on 5/6. Probably not a bad thing to do, since TLT is a "purer" measure of bonds, and is even less correlated to equities than AGG.
|
pgomez
1 posts
msg #100841
- Ignore pgomez |
5/21/2011 9:19:59 PM
Kevin,
I've been following this thread off and on for some time. Are you using this portfolio approach for your overall investment portfolio? If not, what do you use (if you don't mind sharing)?
I'm intrigued by it's potential, but I'd like to hear a little about how this strategy performs in real life.
--Phil
|
Kevin_in_GA
4,599 posts
msg #100842
- Ignore Kevin_in_GA |
5/21/2011 9:58:46 PM
I use a slight variation of this on all of my investment accounts (TSP, 401k, my wife's TIAA-CREF account). The variation I use is a 14 week ROC rather than a 13 week. I suggested the 13 week here because 3 month performance data is almost universally available in most investment accounts, so it doesn't even require a filter. 14 weeks has historically back tested to be the most profitable settings.
In 2010 the14 week ROC approach returned 24% with no trades unprofitable. Compare that to the market return of 12.4%.
As for trading, I use a set of filters developed using another piece of software. Not easily replicated on SF.
|
gmg733
788 posts
msg #100845
- Ignore gmg733 |
5/22/2011 7:52:32 AM
Kevin,
Have you every considered back testing the contrarian view. Buy the laggard and not the runner.
Just curious.
Thanks.
|
Kevin_in_GA
4,599 posts
msg #100847
- Ignore Kevin_in_GA |
5/22/2011 10:28:27 AM
Buy the laggard versus buy the leader:
Using StrataSearch, I just set the buy condition to the worst performer, then let it look at all possible weekly ROCs from 1 to 52 for the period 5/2/2004 until last week. Symlist was (agg,iwm,spy,eem).
27 week ROC for buying the laggard worked out best -
CAGR: 9.99% (versus 2.63% for the ^SPX)
win%: 77.36%
total number of trades: 53
average trade return: 1.78%
average days held: 33
Total Percent Gain: 95.60%
current ETF selected: EEM as of last Friday (was in AGG prior to this)
Compare this to buying the top performer on the 14 week ROC:
CAGR: 21.80% (versus 2.63% for the ^SPX)
win%: 66.67%
total number of trades: 57
average trade return: 2.79%
average days held: 30
Total Percent Gain: 301.45%
current ETF selected: AGG
|
gmg733
788 posts
msg #100849
- Ignore gmg733 |
5/22/2011 11:37:22 AM
Interesting. Good work btw. I was just curious. Being a contrarian usually keeps me on the right side of the trade. But there are a thousand ways to make money in the market and you seem to have found a good simple system.
Thanks.
|
wkloss
231 posts
msg #100851
- Ignore wkloss |
5/22/2011 12:31:47 PM
mankoh & kevin,
Kevin's observation was that mankoh's work could produce more frequent trading and possible whipsaws. I assume that has to do with the values in the formula.
Would either of you suggest a way to change the values to trade less frequently?
Bill
|
mahkoh
1,065 posts
msg #100858
- Ignore mahkoh |
5/22/2011 3:20:41 PM
Bill,
The work is Kevin's. All I did was combine two of his threads. My reasoning was that if you buy the top pick from the list on Monday the market could crash that same day and sell off all week. Rather than sitting it out until Friday I would look for other opportunities.
To avoid getting whipsawed too much I would not use the ROC cross above 0 in the stockcharts link as a hard stop. The crosses above 0 in early Oct 2010 and January 2011 were merely tests of this line, also the MA was still in a downtrend at that point.
The MA starting to move up and ROC retests 0 from above like in August 2010 and this May are more solid indications to leave the market. By the way notice the divergence between the correlation line and the ROC from early April until May.
|
papagatorga
124 posts
msg #100860
- Ignore papagatorga |
5/22/2011 8:32:35 PM
Kevin, Have you considered using the options on these ETF's? And why or why not?
|