OmniVest Forum OmniVest Forum
forums calendars search
today this week
 
register logon control panel Forum Rules
You are currently browsing as a guest.
You should logon to access more features
A Self-Moderated Community - ALL MEMBERS, PLEASE READ!
Vote for Members who contribute the most to your trading, and help us moderate content within the Forums.


[Random Quote] -


Only Forum Moderators, Administrators, and the owner of this message may delete it.

 
Jim Dean

Elite
10002525
Posts: 1059

Joined: 10/11/2012
Location: L'ville, GA

User Profile
 
Subject : RE: Conditional Experiments
Posted : 5/8/2014 2:23 PM
Post #30444 - In reply to #30442

Yes, it is amazing ... and I'm sure that as Steve mentioned, if you follow whatever approach you are using that the % invested could go even higher. I'd be interested in seeing a blow by blow writeup about how you got to that group, btw.

JUST IN CASE there are some readers that aren't really understanding the "curve fitting" effect ... it's probably useful to make a few comments about it to put this into perspective.

First, as Mark mentioned, he is testing during a clearly bullish year. So, things are gonna look rosy, to some degree or another, just throwing darts blindfolded. (Well, maybe without the blindfold, and only a couple of feet away ;~)

Second, he is presenting (afaik) the chart that covers the SAME year that he used for developing the combo. That is, the combo is specifically tuned to that year. The important question is, of course ... is that tuning "representative" of "all bullish years" ... and, as a companion note ... how does one know about the bullish year, BEFORE the year begins?

If in fact the combo that Mark has come up with provides good generalized performance, and is not heavily dependent on "specialized" conditions that would not normally repeat in other bullish periods, then that portfolio can be quite valuable.

The trick is the second question above ... how do we know a bullish year, before (and during) that year ... that is, how do we know to USE this wonder-bull porfolio, and to KEEP using it, via some kind of conditional test.

One possible way is to develop "Market State" formulae, but I've rambled on enough about that in other places.

Another exciting approach is the upcoming Portfolio Balancer. That is, IF that portfolio is sufficiently "generic", then since it's performance seems to be a pretty straight line with consistent small proportionate wiggles, it's reasonable to assume that the "Calmar RATE" would show up for that portfolio early-on in a 2013-bullish-type year ... and PortBal would switch us over to jump onto the rocketship.

=====

The "classic" way to test/verify the general usefulness of this combo of strategies is to identify several independent bullish periods of time, and develop *separate* portfolios, each "sort of tuned" to the different periods. Then, STOP developing them (no more changes allowed!), and test each of them on all the OTHER bullish periods, besides the ones that they were developed on. The set of strats that holds up the best for the other segments is the one to use!

========

So ... Mark's work IS really exciting to see (he's got the magic touch at picking out strat's). What will be even neater to find out, is if he could take the port as it stands, then find a few other "bullish years", and rerun it in those time windows, to see how well it works.

Thanks, Mark!


[Edited by Jim Dean on 5/8/2014 2:27 PM]

Deleting message 30444 : RE: Conditional Experiments


Nirvana Systems
For any problems or issues please contact our Webmaster at webmaster@nirvsys.com.