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.

 
Steve Mayo

Legend
100100100100
Posts: 414

Joined: 10/11/2012
Location: Austin, TX

User Profile
 
Subject : RE: Effects of "Allow settings" & "Max Exposure"..
Posted : 1/21/2014 1:54 PM
Post #28936 - In reply to #28935

Nice work, as usual, Mark!

The thing I notice when comparing 4, 5, and 6 is that, yes, whereas normally the max allocation per trade setting does restrict strategy selection, OV does allow the strats with higher default allocations to be traded when the "allow settings" box is checked, but it certainly looks like it only allows the LAST ONE of the day to be taken. In other words, the impact of that setting -- and the reason the return goes up -- is mainly in letting the portfolio reach a higher level of %invested (plugging another trade into that last little bit of available margin); conversely, it does not appear to be so much the benefit of having added that particular high-allocation strategy, per se. In other words, you might well achieve the same increase in equity simply by adding more high-frequency, low-allocation strategies to the mix.

The max allocation setting (10%, 20%, etc.) is still likely to exclude most of the signals from the high-allocation strategies. At least, compared to an account that allows 30-40% per-stock allocation.

The result is that the process that OV uses to select/exclude signals day-to-day becomes a little more unpredictable. Don't get me wrong, there is value in testing and optimizing a portfolio. But, we have to keep in mind that some trades win/lose a dollar or two and some trades win/lose $1500 -- that's a huge standard deviation. Over a short period of time, a very small number of trades -- being included or being filtered out -- can make a big difference in the return for that period. We all have seen how you can get a whole string of small wins when all of a sudden a big lose comes along and sets you back several months. When backtesting such a high-variability system (as most leveraged stock trading systems are), it all depends on where you draw your boundries as to whether you call it a win or a lose for the period. That's why trying to optimize a portfolio using frequent (monthly, quarterly) re-optimization is so hard. You just don't know if the period you are evaluating is really reflexive of the portfolio's innate character or just the fluke of a few lucky/unlucky ball drops in the filtering lotto. It's why I keep recommending we use risk-adjusted and benchmark-relative metrics when scoring our portfolio candidates and not look too closely at a single month (or maybe even 2 years as Keith points out, but I agree, who wants to wait years to find out if your system is working well!)

For that reason also, I think the best approach is to optimize our selection of strategies in our portfolios, assemble and validate (using good metrics) the best portfolio we can using only minimal filters, and then once the best candidates are narrowed down, THEN look at the value of using the filters in the manner you describe when live-trading to control one-big-bad-trade risk, minimize commissions, avoid un-shortable stocks, etc.
Deleting message 28936 : RE: Effects of "Allow settings" & "Max Exposure"..


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