Current location | Thread information | |
![]() ![]() ![]() ![]() ![]() ![]() |
Last Activity 7/22/2015 8:11 AM 10 replies, 2986 viewings |
|
|
Printer friendly version |
^ Top | |||
Mark Holstius![]() Elite ![]() ![]() ![]() ![]() Posts: 744 Joined: 10/11/2012 Location: Sleepy Hollow, IL ![]() |
I know a number of you may be deciding whether to invest in the new Pro Tools, so I put together some of the results Steve & I have obtained using it during development. Admittedly, these results involved some extensive testing – but they do show what’s possible using the same capabilities now available to everyone. The first capture shows the familiar Portfolio tab in Omnivest with some of the 41 Portfolios I’ve created using the Portfolio Wizard. The # Str Column for the new “Custom Dynamic Portfolios” saved using Portfolio Wizard shows the number of Strategies I “fed” to the ECA (usually a list of Custom Strategies created using Strategy Lab). I use a naming system for those Portfolios to help our research, and I’ve put a box around the actual number of Strategies used each month in the Portfolios built. Most have 10 Strategies per month, but I built two with 17 Strategies per month to see how they might compare with Ed’s excellent ARM4 Margin Portfolio that also contains 17. Note the sizable difference in performance. Below is a capture of a download of all 41 of my Custom Portfolios along with the 16 Omnivest Portfolios so that I could average the results to see how they held up as a group. Last, here’s the performance of a Portfolio Steve & I may offer when Elite Trader is available (then again, we’ll probably have some better ones by then); As always, no one can guarantee these results will hold up going forward, but I wanted everyone to see why we’re so excited about the potential of this new tool… Best of luck in your trading, Mark [Edited by Mark Holstius on 8/3/2014 12:20 PM] ![]() ![]() ![]() | ||
^ Top | |||
kmcintyre![]() Elite ![]() ![]() ![]() ![]() ![]() ![]() ![]() Posts: 890 Joined: 10/11/2012 Location: Portland, OR ![]() |
Mark, Very nice! (But I'm not sure I can handle the MDD!) So when are you going to offer Elite Trader services? :-) It seems that you are pre-screening your strategies and only feeding ECA a handful (30) that you think (or know) will work well together. Have you tried ECA with all your strategies? Great posts! Keith [Edited by kmcintyre on 8/3/2014 12:51 PM] | ||
^ Top | |||
Mark Holstius![]() Elite ![]() ![]() ![]() ![]() Posts: 744 Joined: 10/11/2012 Location: Sleepy Hollow, IL ![]() |
Good morning Keith... Yes, I've tried using all the Strategies but the research Steve & I have done points toward better results using fewer. We suspect it has to do with choosing between Strategies having only infinitesimal differences between the Evaluation Functions and we're trying to limit it to those Strategies that've shown good performance over long periods (robust?) so that a "bad" EF choice won't hurt too much. We also don't like high MDD's (who does?), so that's another goal in our research. Here's an example of what can be done using these new Pro Tools along those lines. This chart covers the same time period we used for all those tests in my previous post above: 2007 thru 2013. (chosen so that it includes the large correction and various market types and not just the bull market the past couple of years) Note the MDD of only 10% during some pretty difficult years / reversals all while doing a "walk forward" every month - and still getting > 50% / year: Seeing as we're just getting started with these tools, they definitely look like a great addition to Omnivest! Mark [Edited by Mark Holstius on 8/4/2014 6:47 AM] ![]() | ||
^ Top | |||
Jim Dean![]() Elite ![]() ![]() ![]() Posts: 1059 Joined: 10/11/2012 Location: L'ville, GA ![]() |
Click here for a possible explanation for this, and a starting - point suggesting for an "intelligent design" approach as to determine what strats to use together in a given ECA run. That is: Is it better to dump a huge collection of strats into ECA and let it churn thru them, or is it better (as M+S seem to have done) to use a carefully selected group? My money's on the latter approach. The trick is, figuring out WHICH strats to dump into the "hopper" for ECA to use. My guess is that there will be a lot of discussion about this and many different successful approaches to it. | ||
^ Top | |||
Fred Gordon![]() Legend ![]() ![]() ![]() ![]() ![]() ![]() ![]() Posts: 481 Joined: 10/11/2012 Location: Fayetteville, Ga ![]() |
Mark, have you enabled either or both "closing and opening" conditions from Strat Lab for your lower risk Port Wiz Result? | ||
^ Top | |||
Steve Mayo![]() Legend ![]() ![]() ![]() ![]() Posts: 414 Joined: 10/11/2012 Location: Austin, TX ![]() |
The various covariates (factors) that impact system performance have differing relative impact. By that, I mean (and these are strictly fictional numbers), that the EF might account for 30% of the return while the strategy population might account for 30% and the account settings might explain the last 20%. It's a huge multiple-regression-statistics question, not unlike trying to figure out the relative impact of heart rate, kidney function, and arteriovenous intonation on changes in blood pressure. Ultimately, they all matter to the extent that even a small change in any one covariate is enough to give you problems over the longer term, whether it be a physical or financial stroke. [I do so love combining my two passions!] My point is that there may not be any one "best way" that works for everybody or in every situation. Sometimes, it's best to take a beta blocker to slow your heart rate, sometimes its better to take an ACE inhibitor to dilate your blood vessels; sometimes a diuretic is best. Trendline analysis might work better in certain markets while RTM approaches work better in others. Everybody will be different as to what side-effects (for investing, that would be level of risk!) they will endure. Repeat after me: there is no riskless investment; there is no drug without side-effects. :-) ECA is a game-changer because it gives us a totally new point of control, that being the EF and the periodic switching functionality. (It's sort of like we just discovered a new class of drugs for treating blood pressure!) And, unlike the confusing nuance of tweaking systems, lists and account settings, this new point of control is at a high level. It's like we've been trying to control blood pressure with only diet and exercise but have just now discovered how to directly dilate your blood vessels. In statistical terms, it is a major covariate with a high odds ratio, meaning it exerts a strong influence on performance (i.e., risk and return). But, it's still not the only covariate: system, list, settings, etc. still matter...a lot! There's a big fundamental difference switching and building. Our original concept was to switch BETWEEN PORTFOLIOS after building the optimum COLLECTION OF STRATEGIES. Although we developed an engine that can do both, simultaneously, Portfolio Wizard is primarily about strategy selection. The quantity and type of strategies used still exerts a strong influence on the results you get. Portfolio Wizard is giving us an incredible new way to directly and more-easily improve the performance of a portfolio. But, I think we still need the Portfolio Switcher to create a system that will truly react to material changes in the market. Switching is an even higher level of control than what is available with PW now. The EF that gives the best result when assembling a subset of strategies each month is likely not the same EF that will best detect a sudden change in market mode. The collection of strategies that work well in a choppy market is probably not the set that will work best in a rapidly falling market. Our latest efforts on using booleans in the EF and the move toward weekly intervals should get us much close to that goal, however. Meanwhile, PW is an incredible new tool...but I think the best way to use it is for improving the performance (and widening the scope/spectrum) of a DEFINED-PURPOSE portfolio, rather than trying to build a single BROAD SPECTRUM portfolio that is idea for every market condition. I think its good for deciding which drug in a specific class will work best for you, but perhaps not for choosing between different therapeutic categories. But, that alone is an incredibly powerful new capability we have never had before!!! The even better news is that we already have the engine in place to do true portfolio switching. It's coming soon! And, Ed just gave everybody a chance to get it for a bargain price! It will not replace human intelligence and perspiration just yet (which is why Elite Trader will be equally exciting), but ECA certainly moves the bar further in a giant step than anything I've seen since, well, the discovery of direct-acting vasodilators. [Edited by Steve Mayo on 8/4/2014 10:52 AM] | ||
^ Top | |||
Mark Holstius![]() Elite ![]() ![]() ![]() ![]() Posts: 744 Joined: 10/11/2012 Location: Sleepy Hollow, IL ![]() |
No, I haven't Fred... That latest low DD example is the first 'crude' result of what looks to be potentially good research using some balancing and allocation adjustments. I'm looking forward to those additional 'balancing' and Portfolio selection tools when they're developed - I think they'll add another outstanding dimension to this process. Mark | ||
^ Top | |||
Geoff![]() Veteran ![]() ![]() ![]() ![]() Posts: 180 Joined: 12/4/2012 Location: Byron Bay NSW Australia ![]() |
Thanks Mark, some great results there! I like the look of your low DD port, which is something that I have been exploring over the last 6 months and in particular since trying out the PW. After running many analyses I am getting some good simulation results for short term periods going back in the range 3 months to 2 years (CAR~ 25% to 50%, MDD~ 2% to 10% ~ approx). However, my simulation results for longer periods of 4 to 5 years are only producing results with CAR~<20% and MDD~>25% aprox. While these returns, if reliable, would be acceptable by many IRA account owners (I would like higher CAR of course!), the high DD's are more problematic for low risk retirement accounts, as you already appreciate. In your post you say "...first 'crude' result of what looks to be potentially good research using some balancing and allocation adjustments". Can you elaborate more on what you mean by this? Also, was this low DD result achieved using the tools currently available to Pro users or did you use the Portfolio Balancer or your own balancing tool? Thanks again Mark. | ||
^ Top | |||
Mark Holstius![]() Elite ![]() ![]() ![]() ![]() Posts: 744 Joined: 10/11/2012 Location: Sleepy Hollow, IL ![]() |
First, Geoff - I looked up Byron Bay and it looks like a gorgeous part of the world. I'm jealous... Second, everything I've posted concerning Portfolio Wizard was done using the same tools available now to everyone. Steve & I exported / downloaded some data to excel to facilitate our studies, but all the Custom Strategies were built using Strategy Lab, and the balancing and allocation adjustments I referred to are just "eyeball" allocation settings for various Portfolios saved from PW. It's not my (or Steve's) intention to "frustrate" people by posting great curves without giving all the details of how they were constructed. We've had the opportunity to work with the development (pre-beta) versions of PW and have invested many (hundreds?) of hours to get to this point. I'm sure that as more people "find their way", even better results will come out - and Elite Trader will be a means to share the fruit of all that labor. Ed is also on the verge of making some great "canned" Portfolios available too via enhancements to the Omnivest site, so the future is definitely looking interesting. BTW - just an opinion, but you might have better luck finding robust combinations by testing with a longer time frame in the past - that's what Steve & I found using 2007 thru 2013 in order to cover a variety of market states. Best of luck in your testing and trading! Mark | ||
^ Top | |||
Geoff![]() Veteran ![]() ![]() ![]() ![]() Posts: 180 Joined: 12/4/2012 Location: Byron Bay NSW Australia ![]() |
Thank you Mark for your response, really appreciate you taking the time to provide what I call, very constructive guidance in the use and potential of OV. That is how I view your posts. Your low DD example really opened my eyes to the possibilities for OV to trade unleveraged, long only, retirement accounts, so naturally I was very interested in the process used to arrive at that result. I see now that developing good strategies and portfolios is important, but I need to also look into tweaking other parts of OV like the account portfolio mix and allocation levels, and account settings. I would love to see more of your examples, as time permits! Yes, Australia is gorgeous! Can't think of a better place to be. We're not called 'the lucky country' for nothing...! Cheers | ||
^ Top | |||
tgrafa![]() Regular ![]() ![]() Posts: 63 Joined: 10/11/2012 Location: Midland, Texas ![]() |
Geoff: Mark generously offer the following to me the other day: Run Strategy Reports over the last 7 years and pick the top 25 (in my case only RTMs for my IRA account). I then used those strategies to plug into PW. I solved for 7 strategies using a 2 year run (probably should have been longer) and increased the CAR 38% with a MDD of 4.4% on an IB IRA 'margin' account. This was compared to a 23% CAR and I believe a 6% MDD prior to Mark's help (THANKS MARK!). I didn't find the 100% CAR but this will certainly work until I create something better or the Elite trader site is up and running :^) I did find some 80% CARs with a Calmar of 10 but that's using the last 2 years on a 175% margin account. |
|
|
Legend | Action | Notification | |||
Administrator
Forum Moderator |
Registered User
Unregistered User |
![]() |
Toggle e-mail notification |