| Index | Recent Threads | Who's Online | User List | Search |
|
|
 |
Portfolio123 » List all forums » Forum: Ranking Systems and Factors/Functions » Thread: Best Ranking Systems |
|
Total posts in this thread: 6 |
[Request a Feature]
[Post new Thread] [Add To My Favorites] [Watch this Thread] |
| Author |
|
|
moonrunreport1
Advanced Member
|
Today I went through and did performance tests on factors, compiled the best of them in one system and it underperformed. Apparently packing them together doesn't do well. I would like to hear from users as to their favorite ranking system and/or how they went about creating ranking systems. I am way behind on theory of how rankers are formed to give best results. Responses? Bryan |
||
|
|
dwpeters
Advanced Member
|
Hi, My experience is the opposite. Somewhere on these forum's is a file posted by Dan Parquette, back in 2005-2006, listing performance for a variety of factors. In 2007 I manually retested the top factors, combined a diverse set of these into a ranking system .. not just the top factors, but the top factors for value, earnings growth, efficiency, momentum, etc. in an attempt to make a balanced ranking. At the beginning of 2008 I selected my top 5 favorite ranking systems and tracked them via generic 20 stock ports. The one I created is up over 140%, outperforming all others that I am tracking - it is an aggressive ranking and has benefited greatly from the micro cap runup started in the spring. I have recently repeated this exercise to create new long and short ranking systems. I believe that the best factors are to some extent timeless, but that it also pays to go with what's working lately. I also have found that with ranking systems, more factors are generally better, while with buy/sell rules, less are better. And while it may pay to overweight the best factors, I don't think optimizing a ranking system is effective going forward. I do test the factors individually within the ranking system (by weighting each to 0 to see how much value it adds to the ranking). Some factors seem to add more or less value then their independent tests would show. Don |
||
|
|
grokkalot
Advanced Member
|
You can post links to your ranking system and simulation so people can see precisely what you did. One thing to be aware of is that taking the top ranked stocks on many different factors can sometimes lead to a portfolio that is concentrated with respect to some type of market sector - e.g. all small cap, or mostly emerging market, or lots of payday lender companies (where pehaps the market is anticipating negative legislative changes) or something of that nature. How to think about that or what to do varies by case and individual preference - e.g. you may be okay with small cap concentration but just need to judge performance against an appropriate benchmark. For the case of frequent rebalancing with a small cap portfolio, I look at the Value Line Arithmetic index as a kind of idealized (though not directly realizable) benchmark. |
||
|
|
dwpeters
Advanced Member
|
Hi, The ranking is private but the port is public: http://www.portfolio123.com/port_summary.jsp?portid=345154 It is definitely concentrated to smaller caps but it does include min liquidity. Don |
||
|
|
martinfierro
Advanced Member
|
Grokkalot: You are exactly right, some ranking systems have a tendency to select very, very small stocks. Almost invariably, these stocks display spectacular results on backtesting. Alas, in real life these stocks are rarely tradeable, and real results are very far from backtests. My personal rule is: regardless of the ranking system I use, I add several liquidity filters to prevent these phantom stocks from being selected. Amazing how these filters bring the results back to earth. Martin |
||
|
|
grokkalot
Advanced Member
|
I use liquidity constraints also. But the size of the liquidity constraint one needs is a function mainly of portfolio size, and to a lesser extent trading platform. The typical private investor who screens for value using appropriate liquidity constraints will still wind up with a small cap portfolio skewed toward particular sectors (actually, one of my beefs with conventional small cap indices is that they skew toward financials).
|
|||
|
| [Show Thread Printable Version] [Post new Thread] |
