Richard,
Yeah, I don't track a portfolio metric which would count as "% winning trades". I guess I could measure in between rebalancing events and count a rising portfolio as a "win" but I'm not sure what I would do with such a number. A number I do care about is the percentage of days in which the portfolio has a higher value than it did 365 days prior. I want to keep this number above 95% and have succeeded so far.
I've read some of Howard's work. It seems to be representative of the tools and approaches used by the trading community broadly. Very good and interesting, but it's an entirely different research program than what I do. It is so much more focused on adapting casino statistics to build individual trading systems, I find a hard time connecting to it. I am more focused on developing new math for markets and securities as phenomena in their own right and using that to construct and adaptively manage hedged portfolios. I've no experience to claim which approach would produce better results. My approach came originally and continues to come out of natural intellectual curiosity about the questions which are at its center. I'm sure Howard feels the same way about the questions which drive him.
I don't have the portfolio system all built out into AFL yet, though I've got lots of pieces that I've not yet stitched together. But I do have a (very large, on the order of 8M cell) spreadsheet model that I manage the current portfolio from. To give a flavor of what it produces, I've attached a screenshot showing simulation of a long/short leveraged ETF portfolio for the 10.6 years between March 2004 and November 2014. These are generally considered very high risk securities individually, and the simulation takes it through the full crash, so this scenario poses a challenge for portfolio management.
You'll notice some things I imagine many traders couldn't accept, such as a 30% maxDD. I've seen some absurdly low maxDD numbers for some trading systems that make me quite jealous. My portfolios all tend to come in between 25% and 35% drawdown risk in historical simulation. Also, volatility metrics are kind of high, which many would not like. Balancing that, the portfolios also tend to have incredibly quick recovery times. You'll notice the %12-month gains are at 98%, meaning that 98% of all days in the test period showed a higher portfolio value than the portfolio had 365 days prior. To me, as I mentioned, that's an important metric. Also, tests such as the skew indicate the overall distribution of returns day-to-day is near normal, and the R-squared correlation of performance to stocks is fairly low at 0.15, both of which are also important to me.
I doubt this answers many of your questions but it's the most I can share to give a flavor. I don't know that this is the best approach but it gives nice returns for people who can accept some increased volatility.
__._,_.___
Posted by: rosenberggregg@yahoo.com
| Reply via web post | • | Reply to sender | • | Reply to group | • | Start a New Topic | • | Messages in this topic (15) |
**** IMPORTANT PLEASE READ ****
This group is for the discussion between users only.
This is *NOT* technical support channel.
TO GET TECHNICAL SUPPORT send an e-mail directly to
SUPPORT {at} amibroker.com
TO SUBMIT SUGGESTIONS please use FEEDBACK CENTER at
http://www.amibroker.com/feedback/
(submissions sent via other channels won't be considered)
For NEW RELEASE ANNOUNCEMENTS and other news always check DEVLOG:
http://www.amibroker.com/devlog/
This group is for the discussion between users only.
This is *NOT* technical support channel.
TO GET TECHNICAL SUPPORT send an e-mail directly to
SUPPORT {at} amibroker.com
TO SUBMIT SUGGESTIONS please use FEEDBACK CENTER at
http://www.amibroker.com/feedback/
(submissions sent via other channels won't be considered)
For NEW RELEASE ANNOUNCEMENTS and other news always check DEVLOG:
http://www.amibroker.com/devlog/
.
__,_._,___
EmoticonEmoticon